3
3
3
115
115
109
109
111
111
... | ... |
@@ -44,100 +44,100 @@ |
44 | 44 |
if HAVE_SOPLEX |
45 | 45 |
lemon_libemon_la_SOURCES += lemon/soplex.cc |
46 | 46 |
endif |
47 | 47 |
|
48 | 48 |
if HAVE_CLP |
49 | 49 |
lemon_libemon_la_SOURCES += lemon/clp.cc |
50 | 50 |
endif |
51 | 51 |
|
52 | 52 |
if HAVE_CBC |
53 | 53 |
lemon_libemon_la_SOURCES += lemon/cbc.cc |
54 | 54 |
endif |
55 | 55 |
|
56 | 56 |
lemon_HEADERS += \ |
57 | 57 |
lemon/adaptors.h \ |
58 | 58 |
lemon/arg_parser.h \ |
59 | 59 |
lemon/assert.h \ |
60 | 60 |
lemon/bellman_ford.h \ |
61 | 61 |
lemon/bfs.h \ |
62 | 62 |
lemon/bin_heap.h \ |
63 | 63 |
lemon/binomial_heap.h \ |
64 | 64 |
lemon/bucket_heap.h \ |
65 | 65 |
lemon/capacity_scaling.h \ |
66 | 66 |
lemon/cbc.h \ |
67 | 67 |
lemon/circulation.h \ |
68 | 68 |
lemon/clp.h \ |
69 | 69 |
lemon/color.h \ |
70 | 70 |
lemon/concept_check.h \ |
71 | 71 |
lemon/connectivity.h \ |
72 | 72 |
lemon/core.h \ |
73 | 73 |
lemon/cost_scaling.h \ |
74 | 74 |
lemon/counter.h \ |
75 | 75 |
lemon/cplex.h \ |
76 | 76 |
lemon/cycle_canceling.h \ |
77 | 77 |
lemon/dfs.h \ |
78 | 78 |
lemon/dheap.h \ |
79 | 79 |
lemon/dijkstra.h \ |
80 | 80 |
lemon/dim2.h \ |
81 | 81 |
lemon/dimacs.h \ |
82 | 82 |
lemon/edge_set.h \ |
83 | 83 |
lemon/elevator.h \ |
84 | 84 |
lemon/error.h \ |
85 | 85 |
lemon/euler.h \ |
86 | 86 |
lemon/fib_heap.h \ |
87 | 87 |
lemon/full_graph.h \ |
88 | 88 |
lemon/glpk.h \ |
89 | 89 |
lemon/gomory_hu.h \ |
90 | 90 |
lemon/graph_to_eps.h \ |
91 | 91 |
lemon/grid_graph.h \ |
92 |
lemon/hartmann_orlin.h \ |
|
93 |
lemon/howard.h \ |
|
92 |
lemon/hartmann_orlin_mmc.h \ |
|
93 |
lemon/howard_mmc.h \ |
|
94 | 94 |
lemon/hypercube_graph.h \ |
95 |
lemon/ |
|
95 |
lemon/karp_mmc.h \ |
|
96 | 96 |
lemon/kruskal.h \ |
97 | 97 |
lemon/hao_orlin.h \ |
98 | 98 |
lemon/lgf_reader.h \ |
99 | 99 |
lemon/lgf_writer.h \ |
100 | 100 |
lemon/list_graph.h \ |
101 | 101 |
lemon/lp.h \ |
102 | 102 |
lemon/lp_base.h \ |
103 | 103 |
lemon/lp_skeleton.h \ |
104 | 104 |
lemon/maps.h \ |
105 | 105 |
lemon/matching.h \ |
106 | 106 |
lemon/math.h \ |
107 | 107 |
lemon/min_cost_arborescence.h \ |
108 | 108 |
lemon/nauty_reader.h \ |
109 | 109 |
lemon/network_simplex.h \ |
110 | 110 |
lemon/pairing_heap.h \ |
111 | 111 |
lemon/path.h \ |
112 | 112 |
lemon/planarity.h \ |
113 | 113 |
lemon/preflow.h \ |
114 | 114 |
lemon/quad_heap.h \ |
115 | 115 |
lemon/radix_heap.h \ |
116 | 116 |
lemon/radix_sort.h \ |
117 | 117 |
lemon/random.h \ |
118 | 118 |
lemon/smart_graph.h \ |
119 | 119 |
lemon/soplex.h \ |
120 | 120 |
lemon/static_graph.h \ |
121 | 121 |
lemon/suurballe.h \ |
122 | 122 |
lemon/time_measure.h \ |
123 | 123 |
lemon/tolerance.h \ |
124 | 124 |
lemon/unionfind.h \ |
125 | 125 |
lemon/bits/windows.h |
126 | 126 |
|
127 | 127 |
bits_HEADERS += \ |
128 | 128 |
lemon/bits/alteration_notifier.h \ |
129 | 129 |
lemon/bits/array_map.h \ |
130 | 130 |
lemon/bits/bezier.h \ |
131 | 131 |
lemon/bits/default_map.h \ |
132 | 132 |
lemon/bits/edge_set_extender.h \ |
133 | 133 |
lemon/bits/enable_if.h \ |
134 | 134 |
lemon/bits/graph_adaptor_extender.h \ |
135 | 135 |
lemon/bits/graph_extender.h \ |
136 | 136 |
lemon/bits/map_extender.h \ |
137 | 137 |
lemon/bits/path_dump.h \ |
138 | 138 |
lemon/bits/solver_bits.h \ |
139 | 139 |
lemon/bits/traits.h \ |
140 | 140 |
lemon/bits/variant.h \ |
141 | 141 |
lemon/bits/vector_map.h |
142 | 142 |
|
143 | 143 |
concept_HEADERS += \ |
1 | 1 |
/* -*- C++ -*- |
2 | 2 |
* |
3 | 3 |
* This file is a part of LEMON, a generic C++ optimization library |
4 | 4 |
* |
5 | 5 |
* Copyright (C) 2003-2008 |
6 | 6 |
* Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport |
7 | 7 |
* (Egervary Research Group on Combinatorial Optimization, EGRES). |
8 | 8 |
* |
9 | 9 |
* Permission to use, modify and distribute this software is granted |
10 | 10 |
* provided that this copyright notice appears in all copies. For |
11 | 11 |
* precise terms see the accompanying LICENSE file. |
12 | 12 |
* |
13 | 13 |
* This software is provided "AS IS" with no warranty of any kind, |
14 | 14 |
* express or implied, and with no claim as to its suitability for any |
15 | 15 |
* purpose. |
16 | 16 |
* |
17 | 17 |
*/ |
18 | 18 |
|
19 | 19 |
#ifndef LEMON_CYCLE_CANCELING_H |
20 | 20 |
#define LEMON_CYCLE_CANCELING_H |
21 | 21 |
|
22 | 22 |
/// \ingroup min_cost_flow_algs |
23 | 23 |
/// \file |
24 | 24 |
/// \brief Cycle-canceling algorithms for finding a minimum cost flow. |
25 | 25 |
|
26 | 26 |
#include <vector> |
27 | 27 |
#include <limits> |
28 | 28 |
|
29 | 29 |
#include <lemon/core.h> |
30 | 30 |
#include <lemon/maps.h> |
31 | 31 |
#include <lemon/path.h> |
32 | 32 |
#include <lemon/math.h> |
33 | 33 |
#include <lemon/static_graph.h> |
34 | 34 |
#include <lemon/adaptors.h> |
35 | 35 |
#include <lemon/circulation.h> |
36 | 36 |
#include <lemon/bellman_ford.h> |
37 |
#include <lemon/ |
|
37 |
#include <lemon/howard_mmc.h> |
|
38 | 38 |
|
39 | 39 |
namespace lemon { |
40 | 40 |
|
41 | 41 |
/// \addtogroup min_cost_flow_algs |
42 | 42 |
/// @{ |
43 | 43 |
|
44 | 44 |
/// \brief Implementation of cycle-canceling algorithms for |
45 | 45 |
/// finding a \ref min_cost_flow "minimum cost flow". |
46 | 46 |
/// |
47 | 47 |
/// \ref CycleCanceling implements three different cycle-canceling |
48 | 48 |
/// algorithms for finding a \ref min_cost_flow "minimum cost flow" |
49 | 49 |
/// \ref amo93networkflows, \ref klein67primal, |
50 | 50 |
/// \ref goldberg89cyclecanceling. |
51 | 51 |
/// The most efficent one (both theoretically and practically) |
52 | 52 |
/// is the \ref CANCEL_AND_TIGHTEN "Cancel and Tighten" algorithm, |
53 | 53 |
/// thus it is the default method. |
54 | 54 |
/// It is strongly polynomial, but in practice, it is typically much |
55 | 55 |
/// slower than the scaling algorithms and NetworkSimplex. |
56 | 56 |
/// |
57 | 57 |
/// Most of the parameters of the problem (except for the digraph) |
58 | 58 |
/// can be given using separate functions, and the algorithm can be |
59 | 59 |
/// executed using the \ref run() function. If some parameters are not |
60 | 60 |
/// specified, then default values will be used. |
61 | 61 |
/// |
62 | 62 |
/// \tparam GR The digraph type the algorithm runs on. |
63 | 63 |
/// \tparam V The number type used for flow amounts, capacity bounds |
64 | 64 |
/// and supply values in the algorithm. By default, it is \c int. |
65 | 65 |
/// \tparam C The number type used for costs and potentials in the |
66 | 66 |
/// algorithm. By default, it is the same as \c V. |
67 | 67 |
/// |
68 | 68 |
/// \warning Both number types must be signed and all input data must |
69 | 69 |
/// be integer. |
70 | 70 |
/// \warning This algorithm does not support negative costs for such |
71 | 71 |
/// arcs that have infinite upper bound. |
72 | 72 |
/// |
73 | 73 |
/// \note For more information about the three available methods, |
74 | 74 |
/// see \ref Method. |
75 | 75 |
#ifdef DOXYGEN |
76 | 76 |
template <typename GR, typename V, typename C> |
77 | 77 |
#else |
78 | 78 |
template <typename GR, typename V = int, typename C = V> |
79 | 79 |
#endif |
80 | 80 |
class CycleCanceling |
81 | 81 |
{ |
82 | 82 |
public: |
83 | 83 |
|
84 | 84 |
/// The type of the digraph |
85 | 85 |
typedef GR Digraph; |
... | ... |
@@ -879,104 +879,104 @@ |
879 | 879 |
} else { |
880 | 880 |
// Search for node disjoint negative cycles |
881 | 881 |
std::vector<int> state(_res_node_num, 0); |
882 | 882 |
int id = 0; |
883 | 883 |
for (int u = 0; u != _res_node_num; ++u) { |
884 | 884 |
if (state[u] != 0) continue; |
885 | 885 |
++id; |
886 | 886 |
int v = u; |
887 | 887 |
for (; v != -1 && state[v] == 0; v = pred[v] == INVALID ? |
888 | 888 |
-1 : rgr.id(rgr.source(pred[v]))) { |
889 | 889 |
state[v] = id; |
890 | 890 |
} |
891 | 891 |
if (v != -1 && state[v] == id) { |
892 | 892 |
// A negative cycle is found |
893 | 893 |
cycle_found = true; |
894 | 894 |
cycle.clear(); |
895 | 895 |
StaticDigraph::Arc a = pred[v]; |
896 | 896 |
Value d, delta = _res_cap[rgr.id(a)]; |
897 | 897 |
cycle.push_back(rgr.id(a)); |
898 | 898 |
while (rgr.id(rgr.source(a)) != v) { |
899 | 899 |
a = pred_map[rgr.source(a)]; |
900 | 900 |
d = _res_cap[rgr.id(a)]; |
901 | 901 |
if (d < delta) delta = d; |
902 | 902 |
cycle.push_back(rgr.id(a)); |
903 | 903 |
} |
904 | 904 |
|
905 | 905 |
// Augment along the cycle |
906 | 906 |
for (int i = 0; i < int(cycle.size()); ++i) { |
907 | 907 |
int j = cycle[i]; |
908 | 908 |
_res_cap[j] -= delta; |
909 | 909 |
_res_cap[_reverse[j]] += delta; |
910 | 910 |
} |
911 | 911 |
} |
912 | 912 |
} |
913 | 913 |
} |
914 | 914 |
|
915 | 915 |
// Increase iteration limit if no cycle is found |
916 | 916 |
if (!cycle_found) { |
917 | 917 |
length_bound = static_cast<int>(length_bound * BF_LIMIT_FACTOR); |
918 | 918 |
} |
919 | 919 |
} |
920 | 920 |
} |
921 | 921 |
} |
922 | 922 |
|
923 | 923 |
// Execute the "Minimum Mean Cycle Canceling" method |
924 | 924 |
void startMinMeanCycleCanceling() { |
925 | 925 |
typedef SimplePath<StaticDigraph> SPath; |
926 | 926 |
typedef typename SPath::ArcIt SPathArcIt; |
927 |
typedef typename |
|
927 |
typedef typename HowardMmc<StaticDigraph, CostArcMap> |
|
928 | 928 |
::template SetPath<SPath>::Create MMC; |
929 | 929 |
|
930 | 930 |
SPath cycle; |
931 | 931 |
MMC mmc(_sgr, _cost_map); |
932 | 932 |
mmc.cycle(cycle); |
933 | 933 |
buildResidualNetwork(); |
934 |
while (mmc. |
|
934 |
while (mmc.findCycleMean() && mmc.cycleCost() < 0) { |
|
935 | 935 |
// Find the cycle |
936 | 936 |
mmc.findCycle(); |
937 | 937 |
|
938 | 938 |
// Compute delta value |
939 | 939 |
Value delta = INF; |
940 | 940 |
for (SPathArcIt a(cycle); a != INVALID; ++a) { |
941 | 941 |
Value d = _res_cap[_id_vec[_sgr.id(a)]]; |
942 | 942 |
if (d < delta) delta = d; |
943 | 943 |
} |
944 | 944 |
|
945 | 945 |
// Augment along the cycle |
946 | 946 |
for (SPathArcIt a(cycle); a != INVALID; ++a) { |
947 | 947 |
int j = _id_vec[_sgr.id(a)]; |
948 | 948 |
_res_cap[j] -= delta; |
949 | 949 |
_res_cap[_reverse[j]] += delta; |
950 | 950 |
} |
951 | 951 |
|
952 | 952 |
// Rebuild the residual network |
953 | 953 |
buildResidualNetwork(); |
954 | 954 |
} |
955 | 955 |
} |
956 | 956 |
|
957 | 957 |
// Execute the "Cancel And Tighten" method |
958 | 958 |
void startCancelAndTighten() { |
959 | 959 |
// Constants for the min mean cycle computations |
960 | 960 |
const double LIMIT_FACTOR = 1.0; |
961 | 961 |
const int MIN_LIMIT = 5; |
962 | 962 |
|
963 | 963 |
// Contruct auxiliary data vectors |
964 | 964 |
DoubleVector pi(_res_node_num, 0.0); |
965 | 965 |
IntVector level(_res_node_num); |
966 | 966 |
BoolVector reached(_res_node_num); |
967 | 967 |
BoolVector processed(_res_node_num); |
968 | 968 |
IntVector pred_node(_res_node_num); |
969 | 969 |
IntVector pred_arc(_res_node_num); |
970 | 970 |
std::vector<int> stack(_res_node_num); |
971 | 971 |
std::vector<int> proc_vector(_res_node_num); |
972 | 972 |
|
973 | 973 |
// Initialize epsilon |
974 | 974 |
double epsilon = 0; |
975 | 975 |
for (int a = 0; a != _res_arc_num; ++a) { |
976 | 976 |
if (_res_cap[a] > 0 && -_cost[a] > epsilon) |
977 | 977 |
epsilon = -_cost[a]; |
978 | 978 |
} |
979 | 979 |
|
980 | 980 |
// Start phases |
981 | 981 |
Tolerance<double> tol; |
982 | 982 |
tol.epsilon(1e-6); |
... | ... |
@@ -1087,84 +1087,84 @@ |
1087 | 1087 |
if (--iter > 0) { |
1088 | 1088 |
for (int u = 0; u != _res_node_num; ++u) { |
1089 | 1089 |
level[u] = 0; |
1090 | 1090 |
} |
1091 | 1091 |
for (int i = proc_head; i > 0; --i) { |
1092 | 1092 |
int u = proc_vector[i]; |
1093 | 1093 |
double p = pi[u]; |
1094 | 1094 |
int l = level[u] + 1; |
1095 | 1095 |
int last_out = _first_out[u+1]; |
1096 | 1096 |
for (int a = _first_out[u]; a != last_out; ++a) { |
1097 | 1097 |
int v = _target[a]; |
1098 | 1098 |
if (_res_cap[a] > 0 && tol.negative(_cost[a] + p - pi[v]) && |
1099 | 1099 |
l > level[v]) level[v] = l; |
1100 | 1100 |
} |
1101 | 1101 |
} |
1102 | 1102 |
|
1103 | 1103 |
// Modify potentials |
1104 | 1104 |
double q = std::numeric_limits<double>::max(); |
1105 | 1105 |
for (int u = 0; u != _res_node_num; ++u) { |
1106 | 1106 |
int lu = level[u]; |
1107 | 1107 |
double p, pu = pi[u]; |
1108 | 1108 |
int last_out = _first_out[u+1]; |
1109 | 1109 |
for (int a = _first_out[u]; a != last_out; ++a) { |
1110 | 1110 |
if (_res_cap[a] == 0) continue; |
1111 | 1111 |
int v = _target[a]; |
1112 | 1112 |
int ld = lu - level[v]; |
1113 | 1113 |
if (ld > 0) { |
1114 | 1114 |
p = (_cost[a] + pu - pi[v] + epsilon) / (ld + 1); |
1115 | 1115 |
if (p < q) q = p; |
1116 | 1116 |
} |
1117 | 1117 |
} |
1118 | 1118 |
} |
1119 | 1119 |
for (int u = 0; u != _res_node_num; ++u) { |
1120 | 1120 |
pi[u] -= q * level[u]; |
1121 | 1121 |
} |
1122 | 1122 |
|
1123 | 1123 |
// Modify epsilon |
1124 | 1124 |
epsilon = 0; |
1125 | 1125 |
for (int u = 0; u != _res_node_num; ++u) { |
1126 | 1126 |
double curr, pu = pi[u]; |
1127 | 1127 |
int last_out = _first_out[u+1]; |
1128 | 1128 |
for (int a = _first_out[u]; a != last_out; ++a) { |
1129 | 1129 |
if (_res_cap[a] == 0) continue; |
1130 | 1130 |
curr = _cost[a] + pu - pi[_target[a]]; |
1131 | 1131 |
if (-curr > epsilon) epsilon = -curr; |
1132 | 1132 |
} |
1133 | 1133 |
} |
1134 | 1134 |
} else { |
1135 |
typedef |
|
1135 |
typedef HowardMmc<StaticDigraph, CostArcMap> MMC; |
|
1136 | 1136 |
typedef typename BellmanFord<StaticDigraph, CostArcMap> |
1137 | 1137 |
::template SetDistMap<CostNodeMap>::Create BF; |
1138 | 1138 |
|
1139 | 1139 |
// Set epsilon to the minimum cycle mean |
1140 | 1140 |
buildResidualNetwork(); |
1141 | 1141 |
MMC mmc(_sgr, _cost_map); |
1142 |
mmc. |
|
1142 |
mmc.findCycleMean(); |
|
1143 | 1143 |
epsilon = -mmc.cycleMean(); |
1144 |
Cost cycle_cost = mmc.cycleLength(); |
|
1145 |
int cycle_size = mmc.cycleArcNum(); |
|
1144 |
Cost cycle_cost = mmc.cycleCost(); |
|
1145 |
int cycle_size = mmc.cycleSize(); |
|
1146 | 1146 |
|
1147 | 1147 |
// Compute feasible potentials for the current epsilon |
1148 | 1148 |
for (int i = 0; i != int(_cost_vec.size()); ++i) { |
1149 | 1149 |
_cost_vec[i] = cycle_size * _cost_vec[i] - cycle_cost; |
1150 | 1150 |
} |
1151 | 1151 |
BF bf(_sgr, _cost_map); |
1152 | 1152 |
bf.distMap(_pi_map); |
1153 | 1153 |
bf.init(0); |
1154 | 1154 |
bf.start(); |
1155 | 1155 |
for (int u = 0; u != _res_node_num; ++u) { |
1156 | 1156 |
pi[u] = static_cast<double>(_pi[u]) / cycle_size; |
1157 | 1157 |
} |
1158 | 1158 |
|
1159 | 1159 |
iter = limit; |
1160 | 1160 |
} |
1161 | 1161 |
} |
1162 | 1162 |
} |
1163 | 1163 |
|
1164 | 1164 |
}; //class CycleCanceling |
1165 | 1165 |
|
1166 | 1166 |
///@} |
1167 | 1167 |
|
1168 | 1168 |
} //namespace lemon |
1169 | 1169 |
|
1170 | 1170 |
#endif //LEMON_CYCLE_CANCELING_H |
1 | 1 |
/* -*- C++ -*- |
2 | 2 |
* |
3 | 3 |
* This file is a part of LEMON, a generic C++ optimization library |
4 | 4 |
* |
5 | 5 |
* Copyright (C) 2003-2008 |
6 | 6 |
* Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport |
7 | 7 |
* (Egervary Research Group on Combinatorial Optimization, EGRES). |
8 | 8 |
* |
9 | 9 |
* Permission to use, modify and distribute this software is granted |
10 | 10 |
* provided that this copyright notice appears in all copies. For |
11 | 11 |
* precise terms see the accompanying LICENSE file. |
12 | 12 |
* |
13 | 13 |
* This software is provided "AS IS" with no warranty of any kind, |
14 | 14 |
* express or implied, and with no claim as to its suitability for any |
15 | 15 |
* purpose. |
16 | 16 |
* |
17 | 17 |
*/ |
18 | 18 |
|
19 |
#ifndef LEMON_HARTMANN_ORLIN_H |
|
20 |
#define LEMON_HARTMANN_ORLIN_H |
|
19 |
#ifndef LEMON_HARTMANN_ORLIN_MMC_H |
|
20 |
#define LEMON_HARTMANN_ORLIN_MMC_H |
|
21 | 21 |
|
22 | 22 |
/// \ingroup min_mean_cycle |
23 | 23 |
/// |
24 | 24 |
/// \file |
25 | 25 |
/// \brief Hartmann-Orlin's algorithm for finding a minimum mean cycle. |
26 | 26 |
|
27 | 27 |
#include <vector> |
28 | 28 |
#include <limits> |
29 | 29 |
#include <lemon/core.h> |
30 | 30 |
#include <lemon/path.h> |
31 | 31 |
#include <lemon/tolerance.h> |
32 | 32 |
#include <lemon/connectivity.h> |
33 | 33 |
|
34 | 34 |
namespace lemon { |
35 | 35 |
|
36 |
/// \brief Default traits class of |
|
36 |
/// \brief Default traits class of HartmannOrlinMmc class. |
|
37 | 37 |
/// |
38 |
/// Default traits class of |
|
38 |
/// Default traits class of HartmannOrlinMmc class. |
|
39 | 39 |
/// \tparam GR The type of the digraph. |
40 |
/// \tparam |
|
40 |
/// \tparam CM The type of the cost map. |
|
41 | 41 |
/// It must conform to the \ref concepts::Rea_data "Rea_data" concept. |
42 | 42 |
#ifdef DOXYGEN |
43 |
template <typename GR, typename |
|
43 |
template <typename GR, typename CM> |
|
44 | 44 |
#else |
45 |
template <typename GR, typename LEN, |
|
46 |
bool integer = std::numeric_limits<typename LEN::Value>::is_integer> |
|
45 |
template <typename GR, typename CM, |
|
46 |
bool integer = std::numeric_limits<typename CM::Value>::is_integer> |
|
47 | 47 |
#endif |
48 |
struct |
|
48 |
struct HartmannOrlinMmcDefaultTraits |
|
49 | 49 |
{ |
50 | 50 |
/// The type of the digraph |
51 | 51 |
typedef GR Digraph; |
52 |
/// The type of the length map |
|
53 |
typedef LEN LengthMap; |
|
54 |
/// The type of the arc lengths |
|
55 |
typedef typename LengthMap::Value Value; |
|
52 |
/// The type of the cost map |
|
53 |
typedef CM CostMap; |
|
54 |
/// The type of the arc costs |
|
55 |
typedef typename CostMap::Value Cost; |
|
56 | 56 |
|
57 |
/// \brief The large |
|
57 |
/// \brief The large cost type used for internal computations |
|
58 | 58 |
/// |
59 |
/// The large value type used for internal computations. |
|
60 |
/// It is \c long \c long if the \c Value type is integer, |
|
59 |
/// The large cost type used for internal computations. |
|
60 |
/// It is \c long \c long if the \c Cost type is integer, |
|
61 | 61 |
/// otherwise it is \c double. |
62 |
/// \c Value must be convertible to \c LargeValue. |
|
63 |
typedef double LargeValue; |
|
62 |
/// \c Cost must be convertible to \c LargeCost. |
|
63 |
typedef double LargeCost; |
|
64 | 64 |
|
65 | 65 |
/// The tolerance type used for internal computations |
66 |
typedef lemon::Tolerance< |
|
66 |
typedef lemon::Tolerance<LargeCost> Tolerance; |
|
67 | 67 |
|
68 | 68 |
/// \brief The path type of the found cycles |
69 | 69 |
/// |
70 | 70 |
/// The path type of the found cycles. |
71 | 71 |
/// It must conform to the \ref lemon::concepts::Path "Path" concept |
72 | 72 |
/// and it must have an \c addFront() function. |
73 | 73 |
typedef lemon::Path<Digraph> Path; |
74 | 74 |
}; |
75 | 75 |
|
76 |
// Default traits class for integer value types |
|
77 |
template <typename GR, typename LEN> |
|
78 |
|
|
76 |
// Default traits class for integer cost types |
|
77 |
template <typename GR, typename CM> |
|
78 |
struct HartmannOrlinMmcDefaultTraits<GR, CM, true> |
|
79 | 79 |
{ |
80 | 80 |
typedef GR Digraph; |
81 |
typedef LEN LengthMap; |
|
82 |
typedef typename LengthMap::Value Value; |
|
81 |
typedef CM CostMap; |
|
82 |
typedef typename CostMap::Value Cost; |
|
83 | 83 |
#ifdef LEMON_HAVE_LONG_LONG |
84 |
typedef long long |
|
84 |
typedef long long LargeCost; |
|
85 | 85 |
#else |
86 |
typedef long |
|
86 |
typedef long LargeCost; |
|
87 | 87 |
#endif |
88 |
typedef lemon::Tolerance< |
|
88 |
typedef lemon::Tolerance<LargeCost> Tolerance; |
|
89 | 89 |
typedef lemon::Path<Digraph> Path; |
90 | 90 |
}; |
91 | 91 |
|
92 | 92 |
|
93 | 93 |
/// \addtogroup min_mean_cycle |
94 | 94 |
/// @{ |
95 | 95 |
|
96 | 96 |
/// \brief Implementation of the Hartmann-Orlin algorithm for finding |
97 | 97 |
/// a minimum mean cycle. |
98 | 98 |
/// |
99 | 99 |
/// This class implements the Hartmann-Orlin algorithm for finding |
100 |
/// a directed cycle of minimum mean |
|
100 |
/// a directed cycle of minimum mean cost in a digraph |
|
101 | 101 |
/// \ref amo93networkflows, \ref dasdan98minmeancycle. |
102 | 102 |
/// It is an improved version of \ref Karp "Karp"'s original algorithm, |
103 | 103 |
/// it applies an efficient early termination scheme. |
104 | 104 |
/// It runs in time O(ne) and uses space O(n<sup>2</sup>+e). |
105 | 105 |
/// |
106 | 106 |
/// \tparam GR The type of the digraph the algorithm runs on. |
107 |
/// \tparam |
|
107 |
/// \tparam CM The type of the cost map. The default |
|
108 | 108 |
/// map type is \ref concepts::Digraph::ArcMap "GR::ArcMap<int>". |
109 | 109 |
/// \tparam TR The traits class that defines various types used by the |
110 |
/// algorithm. By default, it is \ref HartmannOrlinDefaultTraits |
|
111 |
/// "HartmannOrlinDefaultTraits<GR, LEN>". |
|
110 |
/// algorithm. By default, it is \ref HartmannOrlinMmcDefaultTraits |
|
111 |
/// "HartmannOrlinMmcDefaultTraits<GR, CM>". |
|
112 | 112 |
/// In most cases, this parameter should not be set directly, |
113 | 113 |
/// consider to use the named template parameters instead. |
114 | 114 |
#ifdef DOXYGEN |
115 |
template <typename GR, typename |
|
115 |
template <typename GR, typename CM, typename TR> |
|
116 | 116 |
#else |
117 | 117 |
template < typename GR, |
118 |
typename LEN = typename GR::template ArcMap<int>, |
|
119 |
typename TR = HartmannOrlinDefaultTraits<GR, LEN> > |
|
118 |
typename CM = typename GR::template ArcMap<int>, |
|
119 |
typename TR = HartmannOrlinMmcDefaultTraits<GR, CM> > |
|
120 | 120 |
#endif |
121 |
class |
|
121 |
class HartmannOrlinMmc |
|
122 | 122 |
{ |
123 | 123 |
public: |
124 | 124 |
|
125 | 125 |
/// The type of the digraph |
126 | 126 |
typedef typename TR::Digraph Digraph; |
127 |
/// The type of the length map |
|
128 |
typedef typename TR::LengthMap LengthMap; |
|
129 |
/// The type of the arc lengths |
|
130 |
typedef typename TR::Value Value; |
|
127 |
/// The type of the cost map |
|
128 |
typedef typename TR::CostMap CostMap; |
|
129 |
/// The type of the arc costs |
|
130 |
typedef typename TR::Cost Cost; |
|
131 | 131 |
|
132 |
/// \brief The large |
|
132 |
/// \brief The large cost type |
|
133 | 133 |
/// |
134 |
/// The large value type used for internal computations. |
|
135 |
/// By default, it is \c long \c long if the \c Value type is integer, |
|
134 |
/// The large cost type used for internal computations. |
|
135 |
/// By default, it is \c long \c long if the \c Cost type is integer, |
|
136 | 136 |
/// otherwise it is \c double. |
137 |
typedef typename TR:: |
|
137 |
typedef typename TR::LargeCost LargeCost; |
|
138 | 138 |
|
139 | 139 |
/// The tolerance type |
140 | 140 |
typedef typename TR::Tolerance Tolerance; |
141 | 141 |
|
142 | 142 |
/// \brief The path type of the found cycles |
143 | 143 |
/// |
144 | 144 |
/// The path type of the found cycles. |
145 |
/// Using the \ref |
|
145 |
/// Using the \ref HartmannOrlinMmcDefaultTraits "default traits class", |
|
146 | 146 |
/// it is \ref lemon::Path "Path<Digraph>". |
147 | 147 |
typedef typename TR::Path Path; |
148 | 148 |
|
149 |
/// The \ref |
|
149 |
/// The \ref HartmannOrlinMmcDefaultTraits "traits class" of the algorithm |
|
150 | 150 |
typedef TR Traits; |
151 | 151 |
|
152 | 152 |
private: |
153 | 153 |
|
154 | 154 |
TEMPLATE_DIGRAPH_TYPEDEFS(Digraph); |
155 | 155 |
|
156 | 156 |
// Data sturcture for path data |
157 | 157 |
struct PathData |
158 | 158 |
{ |
159 |
|
|
159 |
LargeCost dist; |
|
160 | 160 |
Arc pred; |
161 |
PathData( |
|
161 |
PathData(LargeCost d, Arc p = INVALID) : |
|
162 | 162 |
dist(d), pred(p) {} |
163 | 163 |
}; |
164 | 164 |
|
165 | 165 |
typedef typename Digraph::template NodeMap<std::vector<PathData> > |
166 | 166 |
PathDataNodeMap; |
167 | 167 |
|
168 | 168 |
private: |
169 | 169 |
|
170 | 170 |
// The digraph the algorithm runs on |
171 | 171 |
const Digraph &_gr; |
172 |
// The length of the arcs |
|
173 |
const LengthMap &_length; |
|
172 |
// The cost of the arcs |
|
173 |
const CostMap &_cost; |
|
174 | 174 |
|
175 | 175 |
// Data for storing the strongly connected components |
176 | 176 |
int _comp_num; |
177 | 177 |
typename Digraph::template NodeMap<int> _comp; |
178 | 178 |
std::vector<std::vector<Node> > _comp_nodes; |
179 | 179 |
std::vector<Node>* _nodes; |
180 | 180 |
typename Digraph::template NodeMap<std::vector<Arc> > _out_arcs; |
181 | 181 |
|
182 | 182 |
// Data for the found cycles |
183 | 183 |
bool _curr_found, _best_found; |
184 |
|
|
184 |
LargeCost _curr_cost, _best_cost; |
|
185 | 185 |
int _curr_size, _best_size; |
186 | 186 |
Node _curr_node, _best_node; |
187 | 187 |
int _curr_level, _best_level; |
188 | 188 |
|
189 | 189 |
Path *_cycle_path; |
190 | 190 |
bool _local_path; |
191 | 191 |
|
192 | 192 |
// Node map for storing path data |
193 | 193 |
PathDataNodeMap _data; |
194 | 194 |
// The processed nodes in the last round |
195 | 195 |
std::vector<Node> _process; |
196 | 196 |
|
197 | 197 |
Tolerance _tolerance; |
198 | 198 |
|
199 | 199 |
// Infinite constant |
200 |
const |
|
200 |
const LargeCost INF; |
|
201 | 201 |
|
202 | 202 |
public: |
203 | 203 |
|
204 | 204 |
/// \name Named Template Parameters |
205 | 205 |
/// @{ |
206 | 206 |
|
207 | 207 |
template <typename T> |
208 |
struct SetLargeValueTraits : public Traits { |
|
209 |
typedef T LargeValue; |
|
208 |
struct SetLargeCostTraits : public Traits { |
|
209 |
typedef T LargeCost; |
|
210 | 210 |
typedef lemon::Tolerance<T> Tolerance; |
211 | 211 |
}; |
212 | 212 |
|
213 | 213 |
/// \brief \ref named-templ-param "Named parameter" for setting |
214 |
/// \c |
|
214 |
/// \c LargeCost type. |
|
215 | 215 |
/// |
216 |
/// \ref named-templ-param "Named parameter" for setting \c |
|
216 |
/// \ref named-templ-param "Named parameter" for setting \c LargeCost |
|
217 | 217 |
/// type. It is used for internal computations in the algorithm. |
218 | 218 |
template <typename T> |
219 |
struct SetLargeValue |
|
220 |
: public HartmannOrlin<GR, LEN, SetLargeValueTraits<T> > { |
|
221 |
|
|
219 |
struct SetLargeCost |
|
220 |
: public HartmannOrlinMmc<GR, CM, SetLargeCostTraits<T> > { |
|
221 |
typedef HartmannOrlinMmc<GR, CM, SetLargeCostTraits<T> > Create; |
|
222 | 222 |
}; |
223 | 223 |
|
224 | 224 |
template <typename T> |
225 | 225 |
struct SetPathTraits : public Traits { |
226 | 226 |
typedef T Path; |
227 | 227 |
}; |
228 | 228 |
|
229 | 229 |
/// \brief \ref named-templ-param "Named parameter" for setting |
230 | 230 |
/// \c %Path type. |
231 | 231 |
/// |
232 | 232 |
/// \ref named-templ-param "Named parameter" for setting the \c %Path |
233 | 233 |
/// type of the found cycles. |
234 | 234 |
/// It must conform to the \ref lemon::concepts::Path "Path" concept |
235 | 235 |
/// and it must have an \c addFront() function. |
236 | 236 |
template <typename T> |
237 | 237 |
struct SetPath |
238 |
: public HartmannOrlin<GR, LEN, SetPathTraits<T> > { |
|
239 |
typedef HartmannOrlin<GR, LEN, SetPathTraits<T> > Create; |
|
238 |
: public HartmannOrlinMmc<GR, CM, SetPathTraits<T> > { |
|
239 |
typedef HartmannOrlinMmc<GR, CM, SetPathTraits<T> > Create; |
|
240 | 240 |
}; |
241 | 241 |
|
242 | 242 |
/// @} |
243 | 243 |
|
244 | 244 |
protected: |
245 | 245 |
|
246 |
|
|
246 |
HartmannOrlinMmc() {} |
|
247 | 247 |
|
248 | 248 |
public: |
249 | 249 |
|
250 | 250 |
/// \brief Constructor. |
251 | 251 |
/// |
252 | 252 |
/// The constructor of the class. |
253 | 253 |
/// |
254 | 254 |
/// \param digraph The digraph the algorithm runs on. |
255 |
/// \param length The lengths (costs) of the arcs. |
|
256 |
HartmannOrlin( const Digraph &digraph, |
|
257 |
const LengthMap &length ) : |
|
258 |
_gr(digraph), _length(length), _comp(digraph), _out_arcs(digraph), |
|
259 |
|
|
255 |
/// \param cost The costs of the arcs. |
|
256 |
HartmannOrlinMmc( const Digraph &digraph, |
|
257 |
const CostMap &cost ) : |
|
258 |
_gr(digraph), _cost(cost), _comp(digraph), _out_arcs(digraph), |
|
259 |
_best_found(false), _best_cost(0), _best_size(1), |
|
260 | 260 |
_cycle_path(NULL), _local_path(false), _data(digraph), |
261 |
INF(std::numeric_limits<LargeValue>::has_infinity ? |
|
262 |
std::numeric_limits<LargeValue>::infinity() : |
|
263 |
|
|
261 |
INF(std::numeric_limits<LargeCost>::has_infinity ? |
|
262 |
std::numeric_limits<LargeCost>::infinity() : |
|
263 |
std::numeric_limits<LargeCost>::max()) |
|
264 | 264 |
{} |
265 | 265 |
|
266 | 266 |
/// Destructor. |
267 |
~ |
|
267 |
~HartmannOrlinMmc() { |
|
268 | 268 |
if (_local_path) delete _cycle_path; |
269 | 269 |
} |
270 | 270 |
|
271 | 271 |
/// \brief Set the path structure for storing the found cycle. |
272 | 272 |
/// |
273 | 273 |
/// This function sets an external path structure for storing the |
274 | 274 |
/// found cycle. |
275 | 275 |
/// |
276 | 276 |
/// If you don't call this function before calling \ref run() or |
277 |
/// \ref |
|
277 |
/// \ref findCycleMean(), it will allocate a local \ref Path "path" |
|
278 | 278 |
/// structure. The destuctor deallocates this automatically |
279 | 279 |
/// allocated object, of course. |
280 | 280 |
/// |
281 | 281 |
/// \note The algorithm calls only the \ref lemon::Path::addFront() |
282 | 282 |
/// "addFront()" function of the given path structure. |
283 | 283 |
/// |
284 | 284 |
/// \return <tt>(*this)</tt> |
285 |
|
|
285 |
HartmannOrlinMmc& cycle(Path &path) { |
|
286 | 286 |
if (_local_path) { |
287 | 287 |
delete _cycle_path; |
288 | 288 |
_local_path = false; |
289 | 289 |
} |
290 | 290 |
_cycle_path = &path; |
291 | 291 |
return *this; |
292 | 292 |
} |
293 | 293 |
|
294 | 294 |
/// \brief Set the tolerance used by the algorithm. |
295 | 295 |
/// |
296 | 296 |
/// This function sets the tolerance object used by the algorithm. |
297 | 297 |
/// |
298 | 298 |
/// \return <tt>(*this)</tt> |
299 |
|
|
299 |
HartmannOrlinMmc& tolerance(const Tolerance& tolerance) { |
|
300 | 300 |
_tolerance = tolerance; |
301 | 301 |
return *this; |
302 | 302 |
} |
303 | 303 |
|
304 | 304 |
/// \brief Return a const reference to the tolerance. |
305 | 305 |
/// |
306 | 306 |
/// This function returns a const reference to the tolerance object |
307 | 307 |
/// used by the algorithm. |
308 | 308 |
const Tolerance& tolerance() const { |
309 | 309 |
return _tolerance; |
310 | 310 |
} |
311 | 311 |
|
312 | 312 |
/// \name Execution control |
313 | 313 |
/// The simplest way to execute the algorithm is to call the \ref run() |
314 | 314 |
/// function.\n |
315 |
/// If you only need the minimum mean length, you may call |
|
316 |
/// \ref findMinMean(). |
|
315 |
/// If you only need the minimum mean cost, you may call |
|
316 |
/// \ref findCycleMean(). |
|
317 | 317 |
|
318 | 318 |
/// @{ |
319 | 319 |
|
320 | 320 |
/// \brief Run the algorithm. |
321 | 321 |
/// |
322 | 322 |
/// This function runs the algorithm. |
323 | 323 |
/// It can be called more than once (e.g. if the underlying digraph |
324 |
/// and/or the arc |
|
324 |
/// and/or the arc costs have been modified). |
|
325 | 325 |
/// |
326 | 326 |
/// \return \c true if a directed cycle exists in the digraph. |
327 | 327 |
/// |
328 | 328 |
/// \note <tt>mmc.run()</tt> is just a shortcut of the following code. |
329 | 329 |
/// \code |
330 |
/// return mmc. |
|
330 |
/// return mmc.findCycleMean() && mmc.findCycle(); |
|
331 | 331 |
/// \endcode |
332 | 332 |
bool run() { |
333 |
return |
|
333 |
return findCycleMean() && findCycle(); |
|
334 | 334 |
} |
335 | 335 |
|
336 | 336 |
/// \brief Find the minimum cycle mean. |
337 | 337 |
/// |
338 |
/// This function finds the minimum mean |
|
338 |
/// This function finds the minimum mean cost of the directed |
|
339 | 339 |
/// cycles in the digraph. |
340 | 340 |
/// |
341 | 341 |
/// \return \c true if a directed cycle exists in the digraph. |
342 |
bool |
|
342 |
bool findCycleMean() { |
|
343 | 343 |
// Initialization and find strongly connected components |
344 | 344 |
init(); |
345 | 345 |
findComponents(); |
346 | 346 |
|
347 | 347 |
// Find the minimum cycle mean in the components |
348 | 348 |
for (int comp = 0; comp < _comp_num; ++comp) { |
349 | 349 |
if (!initComponent(comp)) continue; |
350 | 350 |
processRounds(); |
351 | 351 |
|
352 | 352 |
// Update the best cycle (global minimum mean cycle) |
353 | 353 |
if ( _curr_found && (!_best_found || |
354 |
|
|
354 |
_curr_cost * _best_size < _best_cost * _curr_size) ) { |
|
355 | 355 |
_best_found = true; |
356 |
|
|
356 |
_best_cost = _curr_cost; |
|
357 | 357 |
_best_size = _curr_size; |
358 | 358 |
_best_node = _curr_node; |
359 | 359 |
_best_level = _curr_level; |
360 | 360 |
} |
361 | 361 |
} |
362 | 362 |
return _best_found; |
363 | 363 |
} |
364 | 364 |
|
365 | 365 |
/// \brief Find a minimum mean directed cycle. |
366 | 366 |
/// |
367 |
/// This function finds a directed cycle of minimum mean length |
|
368 |
/// in the digraph using the data computed by findMinMean(). |
|
367 |
/// This function finds a directed cycle of minimum mean cost |
|
368 |
/// in the digraph using the data computed by findCycleMean(). |
|
369 | 369 |
/// |
370 | 370 |
/// \return \c true if a directed cycle exists in the digraph. |
371 | 371 |
/// |
372 |
/// \pre \ref |
|
372 |
/// \pre \ref findCycleMean() must be called before using this function. |
|
373 | 373 |
bool findCycle() { |
374 | 374 |
if (!_best_found) return false; |
375 | 375 |
IntNodeMap reached(_gr, -1); |
376 | 376 |
int r = _best_level + 1; |
377 | 377 |
Node u = _best_node; |
378 | 378 |
while (reached[u] < 0) { |
379 | 379 |
reached[u] = --r; |
380 | 380 |
u = _gr.source(_data[u][r].pred); |
381 | 381 |
} |
382 | 382 |
r = reached[u]; |
383 | 383 |
Arc e = _data[u][r].pred; |
384 | 384 |
_cycle_path->addFront(e); |
385 |
|
|
385 |
_best_cost = _cost[e]; |
|
386 | 386 |
_best_size = 1; |
387 | 387 |
Node v; |
388 | 388 |
while ((v = _gr.source(e)) != u) { |
389 | 389 |
e = _data[v][--r].pred; |
390 | 390 |
_cycle_path->addFront(e); |
391 |
|
|
391 |
_best_cost += _cost[e]; |
|
392 | 392 |
++_best_size; |
393 | 393 |
} |
394 | 394 |
return true; |
395 | 395 |
} |
396 | 396 |
|
397 | 397 |
/// @} |
398 | 398 |
|
399 | 399 |
/// \name Query Functions |
400 | 400 |
/// The results of the algorithm can be obtained using these |
401 | 401 |
/// functions.\n |
402 | 402 |
/// The algorithm should be executed before using them. |
403 | 403 |
|
404 | 404 |
/// @{ |
405 | 405 |
|
406 |
/// \brief Return the total |
|
406 |
/// \brief Return the total cost of the found cycle. |
|
407 | 407 |
/// |
408 |
/// This function returns the total |
|
408 |
/// This function returns the total cost of the found cycle. |
|
409 | 409 |
/// |
410 |
/// \pre \ref run() or \ref |
|
410 |
/// \pre \ref run() or \ref findCycleMean() must be called before |
|
411 | 411 |
/// using this function. |
412 |
Value cycleLength() const { |
|
413 |
return static_cast<Value>(_best_length); |
|
412 |
Cost cycleCost() const { |
|
413 |
return static_cast<Cost>(_best_cost); |
|
414 | 414 |
} |
415 | 415 |
|
416 | 416 |
/// \brief Return the number of arcs on the found cycle. |
417 | 417 |
/// |
418 | 418 |
/// This function returns the number of arcs on the found cycle. |
419 | 419 |
/// |
420 |
/// \pre \ref run() or \ref |
|
420 |
/// \pre \ref run() or \ref findCycleMean() must be called before |
|
421 | 421 |
/// using this function. |
422 |
int |
|
422 |
int cycleSize() const { |
|
423 | 423 |
return _best_size; |
424 | 424 |
} |
425 | 425 |
|
426 |
/// \brief Return the mean |
|
426 |
/// \brief Return the mean cost of the found cycle. |
|
427 | 427 |
/// |
428 |
/// This function returns the mean |
|
428 |
/// This function returns the mean cost of the found cycle. |
|
429 | 429 |
/// |
430 | 430 |
/// \note <tt>alg.cycleMean()</tt> is just a shortcut of the |
431 | 431 |
/// following code. |
432 | 432 |
/// \code |
433 |
/// return static_cast<double>(alg. |
|
433 |
/// return static_cast<double>(alg.cycleCost()) / alg.cycleSize(); |
|
434 | 434 |
/// \endcode |
435 | 435 |
/// |
436 |
/// \pre \ref run() or \ref |
|
436 |
/// \pre \ref run() or \ref findCycleMean() must be called before |
|
437 | 437 |
/// using this function. |
438 | 438 |
double cycleMean() const { |
439 |
return static_cast<double>( |
|
439 |
return static_cast<double>(_best_cost) / _best_size; |
|
440 | 440 |
} |
441 | 441 |
|
442 | 442 |
/// \brief Return the found cycle. |
443 | 443 |
/// |
444 | 444 |
/// This function returns a const reference to the path structure |
445 | 445 |
/// storing the found cycle. |
446 | 446 |
/// |
447 | 447 |
/// \pre \ref run() or \ref findCycle() must be called before using |
448 | 448 |
/// this function. |
449 | 449 |
const Path& cycle() const { |
450 | 450 |
return *_cycle_path; |
451 | 451 |
} |
452 | 452 |
|
453 | 453 |
///@} |
454 | 454 |
|
455 | 455 |
private: |
456 | 456 |
|
457 | 457 |
// Initialization |
458 | 458 |
void init() { |
459 | 459 |
if (!_cycle_path) { |
460 | 460 |
_local_path = true; |
461 | 461 |
_cycle_path = new Path; |
462 | 462 |
} |
463 | 463 |
_cycle_path->clear(); |
464 | 464 |
_best_found = false; |
465 |
|
|
465 |
_best_cost = 0; |
|
466 | 466 |
_best_size = 1; |
467 | 467 |
_cycle_path->clear(); |
468 | 468 |
for (NodeIt u(_gr); u != INVALID; ++u) |
469 | 469 |
_data[u].clear(); |
470 | 470 |
} |
471 | 471 |
|
472 | 472 |
// Find strongly connected components and initialize _comp_nodes |
473 | 473 |
// and _out_arcs |
474 | 474 |
void findComponents() { |
475 | 475 |
_comp_num = stronglyConnectedComponents(_gr, _comp); |
476 | 476 |
_comp_nodes.resize(_comp_num); |
477 | 477 |
if (_comp_num == 1) { |
478 | 478 |
_comp_nodes[0].clear(); |
479 | 479 |
for (NodeIt n(_gr); n != INVALID; ++n) { |
480 | 480 |
_comp_nodes[0].push_back(n); |
481 | 481 |
_out_arcs[n].clear(); |
482 | 482 |
for (OutArcIt a(_gr, n); a != INVALID; ++a) { |
483 | 483 |
_out_arcs[n].push_back(a); |
484 | 484 |
} |
485 | 485 |
} |
486 | 486 |
} else { |
487 | 487 |
for (int i = 0; i < _comp_num; ++i) |
488 | 488 |
_comp_nodes[i].clear(); |
489 | 489 |
for (NodeIt n(_gr); n != INVALID; ++n) { |
490 | 490 |
int k = _comp[n]; |
491 | 491 |
_comp_nodes[k].push_back(n); |
492 | 492 |
_out_arcs[n].clear(); |
493 | 493 |
for (OutArcIt a(_gr, n); a != INVALID; ++a) { |
494 | 494 |
if (_comp[_gr.target(a)] == k) _out_arcs[n].push_back(a); |
495 | 495 |
} |
496 | 496 |
} |
497 | 497 |
} |
498 | 498 |
} |
499 | 499 |
|
500 | 500 |
// Initialize path data for the current component |
501 | 501 |
bool initComponent(int comp) { |
502 | 502 |
_nodes = &(_comp_nodes[comp]); |
503 | 503 |
int n = _nodes->size(); |
504 | 504 |
if (n < 1 || (n == 1 && _out_arcs[(*_nodes)[0]].size() == 0)) { |
505 | 505 |
return false; |
506 | 506 |
} |
507 | 507 |
for (int i = 0; i < n; ++i) { |
508 | 508 |
_data[(*_nodes)[i]].resize(n + 1, PathData(INF)); |
509 | 509 |
} |
510 | 510 |
return true; |
511 | 511 |
} |
512 | 512 |
|
513 | 513 |
// Process all rounds of computing path data for the current component. |
514 |
// _data[v][k] is the |
|
514 |
// _data[v][k] is the cost of a shortest directed walk from the root |
|
515 | 515 |
// node to node v containing exactly k arcs. |
516 | 516 |
void processRounds() { |
517 | 517 |
Node start = (*_nodes)[0]; |
518 | 518 |
_data[start][0] = PathData(0); |
519 | 519 |
_process.clear(); |
520 | 520 |
_process.push_back(start); |
521 | 521 |
|
522 | 522 |
int k, n = _nodes->size(); |
523 | 523 |
int next_check = 4; |
524 | 524 |
bool terminate = false; |
525 | 525 |
for (k = 1; k <= n && int(_process.size()) < n && !terminate; ++k) { |
526 | 526 |
processNextBuildRound(k); |
527 | 527 |
if (k == next_check || k == n) { |
528 | 528 |
terminate = checkTermination(k); |
529 | 529 |
next_check = next_check * 3 / 2; |
530 | 530 |
} |
531 | 531 |
} |
532 | 532 |
for ( ; k <= n && !terminate; ++k) { |
533 | 533 |
processNextFullRound(k); |
534 | 534 |
if (k == next_check || k == n) { |
535 | 535 |
terminate = checkTermination(k); |
536 | 536 |
next_check = next_check * 3 / 2; |
537 | 537 |
} |
538 | 538 |
} |
539 | 539 |
} |
540 | 540 |
|
541 | 541 |
// Process one round and rebuild _process |
542 | 542 |
void processNextBuildRound(int k) { |
543 | 543 |
std::vector<Node> next; |
544 | 544 |
Node u, v; |
545 | 545 |
Arc e; |
546 |
|
|
546 |
LargeCost d; |
|
547 | 547 |
for (int i = 0; i < int(_process.size()); ++i) { |
548 | 548 |
u = _process[i]; |
549 | 549 |
for (int j = 0; j < int(_out_arcs[u].size()); ++j) { |
550 | 550 |
e = _out_arcs[u][j]; |
551 | 551 |
v = _gr.target(e); |
552 |
d = _data[u][k-1].dist + |
|
552 |
d = _data[u][k-1].dist + _cost[e]; |
|
553 | 553 |
if (_tolerance.less(d, _data[v][k].dist)) { |
554 | 554 |
if (_data[v][k].dist == INF) next.push_back(v); |
555 | 555 |
_data[v][k] = PathData(d, e); |
556 | 556 |
} |
557 | 557 |
} |
558 | 558 |
} |
559 | 559 |
_process.swap(next); |
560 | 560 |
} |
561 | 561 |
|
562 | 562 |
// Process one round using _nodes instead of _process |
563 | 563 |
void processNextFullRound(int k) { |
564 | 564 |
Node u, v; |
565 | 565 |
Arc e; |
566 |
|
|
566 |
LargeCost d; |
|
567 | 567 |
for (int i = 0; i < int(_nodes->size()); ++i) { |
568 | 568 |
u = (*_nodes)[i]; |
569 | 569 |
for (int j = 0; j < int(_out_arcs[u].size()); ++j) { |
570 | 570 |
e = _out_arcs[u][j]; |
571 | 571 |
v = _gr.target(e); |
572 |
d = _data[u][k-1].dist + |
|
572 |
d = _data[u][k-1].dist + _cost[e]; |
|
573 | 573 |
if (_tolerance.less(d, _data[v][k].dist)) { |
574 | 574 |
_data[v][k] = PathData(d, e); |
575 | 575 |
} |
576 | 576 |
} |
577 | 577 |
} |
578 | 578 |
} |
579 | 579 |
|
580 | 580 |
// Check early termination |
581 | 581 |
bool checkTermination(int k) { |
582 | 582 |
typedef std::pair<int, int> Pair; |
583 | 583 |
typename GR::template NodeMap<Pair> level(_gr, Pair(-1, 0)); |
584 |
typename GR::template NodeMap< |
|
584 |
typename GR::template NodeMap<LargeCost> pi(_gr); |
|
585 | 585 |
int n = _nodes->size(); |
586 |
|
|
586 |
LargeCost cost; |
|
587 | 587 |
int size; |
588 | 588 |
Node u; |
589 | 589 |
|
590 | 590 |
// Search for cycles that are already found |
591 | 591 |
_curr_found = false; |
592 | 592 |
for (int i = 0; i < n; ++i) { |
593 | 593 |
u = (*_nodes)[i]; |
594 | 594 |
if (_data[u][k].dist == INF) continue; |
595 | 595 |
for (int j = k; j >= 0; --j) { |
596 | 596 |
if (level[u].first == i && level[u].second > 0) { |
597 | 597 |
// A cycle is found |
598 |
|
|
598 |
cost = _data[u][level[u].second].dist - _data[u][j].dist; |
|
599 | 599 |
size = level[u].second - j; |
600 |
if (!_curr_found || length * _curr_size < _curr_length * size) { |
|
601 |
_curr_length = length; |
|
600 |
if (!_curr_found || cost * _curr_size < _curr_cost * size) { |
|
601 |
_curr_cost = cost; |
|
602 | 602 |
_curr_size = size; |
603 | 603 |
_curr_node = u; |
604 | 604 |
_curr_level = level[u].second; |
605 | 605 |
_curr_found = true; |
606 | 606 |
} |
607 | 607 |
} |
608 | 608 |
level[u] = Pair(i, j); |
609 | 609 |
if (j != 0) { |
610 | 610 |
u = _gr.source(_data[u][j].pred); |
611 | 611 |
} |
612 | 612 |
} |
613 | 613 |
} |
614 | 614 |
|
615 | 615 |
// If at least one cycle is found, check the optimality condition |
616 |
|
|
616 |
LargeCost d; |
|
617 | 617 |
if (_curr_found && k < n) { |
618 | 618 |
// Find node potentials |
619 | 619 |
for (int i = 0; i < n; ++i) { |
620 | 620 |
u = (*_nodes)[i]; |
621 | 621 |
pi[u] = INF; |
622 | 622 |
for (int j = 0; j <= k; ++j) { |
623 | 623 |
if (_data[u][j].dist < INF) { |
624 |
d = _data[u][j].dist * _curr_size - j * |
|
624 |
d = _data[u][j].dist * _curr_size - j * _curr_cost; |
|
625 | 625 |
if (_tolerance.less(d, pi[u])) pi[u] = d; |
626 | 626 |
} |
627 | 627 |
} |
628 | 628 |
} |
629 | 629 |
|
630 | 630 |
// Check the optimality condition for all arcs |
631 | 631 |
bool done = true; |
632 | 632 |
for (ArcIt a(_gr); a != INVALID; ++a) { |
633 |
if (_tolerance.less( |
|
633 |
if (_tolerance.less(_cost[a] * _curr_size - _curr_cost, |
|
634 | 634 |
pi[_gr.target(a)] - pi[_gr.source(a)]) ) { |
635 | 635 |
done = false; |
636 | 636 |
break; |
637 | 637 |
} |
638 | 638 |
} |
639 | 639 |
return done; |
640 | 640 |
} |
641 | 641 |
return (k == n); |
642 | 642 |
} |
643 | 643 |
|
644 |
}; //class |
|
644 |
}; //class HartmannOrlinMmc |
|
645 | 645 |
|
646 | 646 |
///@} |
647 | 647 |
|
648 | 648 |
} //namespace lemon |
649 | 649 |
|
650 |
#endif // |
|
650 |
#endif //LEMON_HARTMANN_ORLIN_MMC_H |
1 | 1 |
/* -*- C++ -*- |
2 | 2 |
* |
3 | 3 |
* This file is a part of LEMON, a generic C++ optimization library |
4 | 4 |
* |
5 | 5 |
* Copyright (C) 2003-2008 |
6 | 6 |
* Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport |
7 | 7 |
* (Egervary Research Group on Combinatorial Optimization, EGRES). |
8 | 8 |
* |
9 | 9 |
* Permission to use, modify and distribute this software is granted |
10 | 10 |
* provided that this copyright notice appears in all copies. For |
11 | 11 |
* precise terms see the accompanying LICENSE file. |
12 | 12 |
* |
13 | 13 |
* This software is provided "AS IS" with no warranty of any kind, |
14 | 14 |
* express or implied, and with no claim as to its suitability for any |
15 | 15 |
* purpose. |
16 | 16 |
* |
17 | 17 |
*/ |
18 | 18 |
|
19 |
#ifndef LEMON_HOWARD_H |
|
20 |
#define LEMON_HOWARD_H |
|
19 |
#ifndef LEMON_HOWARD_MMC_H |
|
20 |
#define LEMON_HOWARD_MMC_H |
|
21 | 21 |
|
22 | 22 |
/// \ingroup min_mean_cycle |
23 | 23 |
/// |
24 | 24 |
/// \file |
25 | 25 |
/// \brief Howard's algorithm for finding a minimum mean cycle. |
26 | 26 |
|
27 | 27 |
#include <vector> |
28 | 28 |
#include <limits> |
29 | 29 |
#include <lemon/core.h> |
30 | 30 |
#include <lemon/path.h> |
31 | 31 |
#include <lemon/tolerance.h> |
32 | 32 |
#include <lemon/connectivity.h> |
33 | 33 |
|
34 | 34 |
namespace lemon { |
35 | 35 |
|
36 |
/// \brief Default traits class of |
|
36 |
/// \brief Default traits class of HowardMmc class. |
|
37 | 37 |
/// |
38 |
/// Default traits class of |
|
38 |
/// Default traits class of HowardMmc class. |
|
39 | 39 |
/// \tparam GR The type of the digraph. |
40 |
/// \tparam |
|
40 |
/// \tparam CM The type of the cost map. |
|
41 | 41 |
/// It must conform to the \ref concepts::ReadMap "ReadMap" concept. |
42 | 42 |
#ifdef DOXYGEN |
43 |
template <typename GR, typename |
|
43 |
template <typename GR, typename CM> |
|
44 | 44 |
#else |
45 |
template <typename GR, typename LEN, |
|
46 |
bool integer = std::numeric_limits<typename LEN::Value>::is_integer> |
|
45 |
template <typename GR, typename CM, |
|
46 |
bool integer = std::numeric_limits<typename CM::Value>::is_integer> |
|
47 | 47 |
#endif |
48 |
struct |
|
48 |
struct HowardMmcDefaultTraits |
|
49 | 49 |
{ |
50 | 50 |
/// The type of the digraph |
51 | 51 |
typedef GR Digraph; |
52 |
/// The type of the length map |
|
53 |
typedef LEN LengthMap; |
|
54 |
/// The type of the arc lengths |
|
55 |
typedef typename LengthMap::Value Value; |
|
52 |
/// The type of the cost map |
|
53 |
typedef CM CostMap; |
|
54 |
/// The type of the arc costs |
|
55 |
typedef typename CostMap::Value Cost; |
|
56 | 56 |
|
57 |
/// \brief The large |
|
57 |
/// \brief The large cost type used for internal computations |
|
58 | 58 |
/// |
59 |
/// The large value type used for internal computations. |
|
60 |
/// It is \c long \c long if the \c Value type is integer, |
|
59 |
/// The large cost type used for internal computations. |
|
60 |
/// It is \c long \c long if the \c Cost type is integer, |
|
61 | 61 |
/// otherwise it is \c double. |
62 |
/// \c Value must be convertible to \c LargeValue. |
|
63 |
typedef double LargeValue; |
|
62 |
/// \c Cost must be convertible to \c LargeCost. |
|
63 |
typedef double LargeCost; |
|
64 | 64 |
|
65 | 65 |
/// The tolerance type used for internal computations |
66 |
typedef lemon::Tolerance< |
|
66 |
typedef lemon::Tolerance<LargeCost> Tolerance; |
|
67 | 67 |
|
68 | 68 |
/// \brief The path type of the found cycles |
69 | 69 |
/// |
70 | 70 |
/// The path type of the found cycles. |
71 | 71 |
/// It must conform to the \ref lemon::concepts::Path "Path" concept |
72 | 72 |
/// and it must have an \c addBack() function. |
73 | 73 |
typedef lemon::Path<Digraph> Path; |
74 | 74 |
}; |
75 | 75 |
|
76 |
// Default traits class for integer value types |
|
77 |
template <typename GR, typename LEN> |
|
78 |
|
|
76 |
// Default traits class for integer cost types |
|
77 |
template <typename GR, typename CM> |
|
78 |
struct HowardMmcDefaultTraits<GR, CM, true> |
|
79 | 79 |
{ |
80 | 80 |
typedef GR Digraph; |
81 |
typedef LEN LengthMap; |
|
82 |
typedef typename LengthMap::Value Value; |
|
81 |
typedef CM CostMap; |
|
82 |
typedef typename CostMap::Value Cost; |
|
83 | 83 |
#ifdef LEMON_HAVE_LONG_LONG |
84 |
typedef long long |
|
84 |
typedef long long LargeCost; |
|
85 | 85 |
#else |
86 |
typedef long |
|
86 |
typedef long LargeCost; |
|
87 | 87 |
#endif |
88 |
typedef lemon::Tolerance< |
|
88 |
typedef lemon::Tolerance<LargeCost> Tolerance; |
|
89 | 89 |
typedef lemon::Path<Digraph> Path; |
90 | 90 |
}; |
91 | 91 |
|
92 | 92 |
|
93 | 93 |
/// \addtogroup min_mean_cycle |
94 | 94 |
/// @{ |
95 | 95 |
|
96 | 96 |
/// \brief Implementation of Howard's algorithm for finding a minimum |
97 | 97 |
/// mean cycle. |
98 | 98 |
/// |
99 | 99 |
/// This class implements Howard's policy iteration algorithm for finding |
100 |
/// a directed cycle of minimum mean |
|
100 |
/// a directed cycle of minimum mean cost in a digraph |
|
101 | 101 |
/// \ref amo93networkflows, \ref dasdan98minmeancycle. |
102 | 102 |
/// This class provides the most efficient algorithm for the |
103 | 103 |
/// minimum mean cycle problem, though the best known theoretical |
104 | 104 |
/// bound on its running time is exponential. |
105 | 105 |
/// |
106 | 106 |
/// \tparam GR The type of the digraph the algorithm runs on. |
107 |
/// \tparam |
|
107 |
/// \tparam CM The type of the cost map. The default |
|
108 | 108 |
/// map type is \ref concepts::Digraph::ArcMap "GR::ArcMap<int>". |
109 | 109 |
/// \tparam TR The traits class that defines various types used by the |
110 |
/// algorithm. By default, it is \ref HowardDefaultTraits |
|
111 |
/// "HowardDefaultTraits<GR, LEN>". |
|
110 |
/// algorithm. By default, it is \ref HowardMmcDefaultTraits |
|
111 |
/// "HowardMmcDefaultTraits<GR, CM>". |
|
112 | 112 |
/// In most cases, this parameter should not be set directly, |
113 | 113 |
/// consider to use the named template parameters instead. |
114 | 114 |
#ifdef DOXYGEN |
115 |
template <typename GR, typename |
|
115 |
template <typename GR, typename CM, typename TR> |
|
116 | 116 |
#else |
117 | 117 |
template < typename GR, |
118 |
typename LEN = typename GR::template ArcMap<int>, |
|
119 |
typename TR = HowardDefaultTraits<GR, LEN> > |
|
118 |
typename CM = typename GR::template ArcMap<int>, |
|
119 |
typename TR = HowardMmcDefaultTraits<GR, CM> > |
|
120 | 120 |
#endif |
121 |
class |
|
121 |
class HowardMmc |
|
122 | 122 |
{ |
123 | 123 |
public: |
124 | 124 |
|
125 | 125 |
/// The type of the digraph |
126 | 126 |
typedef typename TR::Digraph Digraph; |
127 |
/// The type of the length map |
|
128 |
typedef typename TR::LengthMap LengthMap; |
|
129 |
/// The type of the arc lengths |
|
130 |
typedef typename TR::Value Value; |
|
127 |
/// The type of the cost map |
|
128 |
typedef typename TR::CostMap CostMap; |
|
129 |
/// The type of the arc costs |
|
130 |
typedef typename TR::Cost Cost; |
|
131 | 131 |
|
132 |
/// \brief The large |
|
132 |
/// \brief The large cost type |
|
133 | 133 |
/// |
134 |
/// The large value type used for internal computations. |
|
135 |
/// By default, it is \c long \c long if the \c Value type is integer, |
|
134 |
/// The large cost type used for internal computations. |
|
135 |
/// By default, it is \c long \c long if the \c Cost type is integer, |
|
136 | 136 |
/// otherwise it is \c double. |
137 |
typedef typename TR:: |
|
137 |
typedef typename TR::LargeCost LargeCost; |
|
138 | 138 |
|
139 | 139 |
/// The tolerance type |
140 | 140 |
typedef typename TR::Tolerance Tolerance; |
141 | 141 |
|
142 | 142 |
/// \brief The path type of the found cycles |
143 | 143 |
/// |
144 | 144 |
/// The path type of the found cycles. |
145 |
/// Using the \ref |
|
145 |
/// Using the \ref HowardMmcDefaultTraits "default traits class", |
|
146 | 146 |
/// it is \ref lemon::Path "Path<Digraph>". |
147 | 147 |
typedef typename TR::Path Path; |
148 | 148 |
|
149 |
/// The \ref |
|
149 |
/// The \ref HowardMmcDefaultTraits "traits class" of the algorithm |
|
150 | 150 |
typedef TR Traits; |
151 | 151 |
|
152 | 152 |
private: |
153 | 153 |
|
154 | 154 |
TEMPLATE_DIGRAPH_TYPEDEFS(Digraph); |
155 | 155 |
|
156 | 156 |
// The digraph the algorithm runs on |
157 | 157 |
const Digraph &_gr; |
158 |
// The length of the arcs |
|
159 |
const LengthMap &_length; |
|
158 |
// The cost of the arcs |
|
159 |
const CostMap &_cost; |
|
160 | 160 |
|
161 | 161 |
// Data for the found cycles |
162 | 162 |
bool _curr_found, _best_found; |
163 |
|
|
163 |
LargeCost _curr_cost, _best_cost; |
|
164 | 164 |
int _curr_size, _best_size; |
165 | 165 |
Node _curr_node, _best_node; |
166 | 166 |
|
167 | 167 |
Path *_cycle_path; |
168 | 168 |
bool _local_path; |
169 | 169 |
|
170 | 170 |
// Internal data used by the algorithm |
171 | 171 |
typename Digraph::template NodeMap<Arc> _policy; |
172 | 172 |
typename Digraph::template NodeMap<bool> _reached; |
173 | 173 |
typename Digraph::template NodeMap<int> _level; |
174 |
typename Digraph::template NodeMap< |
|
174 |
typename Digraph::template NodeMap<LargeCost> _dist; |
|
175 | 175 |
|
176 | 176 |
// Data for storing the strongly connected components |
177 | 177 |
int _comp_num; |
178 | 178 |
typename Digraph::template NodeMap<int> _comp; |
179 | 179 |
std::vector<std::vector<Node> > _comp_nodes; |
180 | 180 |
std::vector<Node>* _nodes; |
181 | 181 |
typename Digraph::template NodeMap<std::vector<Arc> > _in_arcs; |
182 | 182 |
|
183 | 183 |
// Queue used for BFS search |
184 | 184 |
std::vector<Node> _queue; |
185 | 185 |
int _qfront, _qback; |
186 | 186 |
|
187 | 187 |
Tolerance _tolerance; |
188 | 188 |
|
189 | 189 |
// Infinite constant |
190 |
const |
|
190 |
const LargeCost INF; |
|
191 | 191 |
|
192 | 192 |
public: |
193 | 193 |
|
194 | 194 |
/// \name Named Template Parameters |
195 | 195 |
/// @{ |
196 | 196 |
|
197 | 197 |
template <typename T> |
198 |
struct SetLargeValueTraits : public Traits { |
|
199 |
typedef T LargeValue; |
|
198 |
struct SetLargeCostTraits : public Traits { |
|
199 |
typedef T LargeCost; |
|
200 | 200 |
typedef lemon::Tolerance<T> Tolerance; |
201 | 201 |
}; |
202 | 202 |
|
203 | 203 |
/// \brief \ref named-templ-param "Named parameter" for setting |
204 |
/// \c |
|
204 |
/// \c LargeCost type. |
|
205 | 205 |
/// |
206 |
/// \ref named-templ-param "Named parameter" for setting \c |
|
206 |
/// \ref named-templ-param "Named parameter" for setting \c LargeCost |
|
207 | 207 |
/// type. It is used for internal computations in the algorithm. |
208 | 208 |
template <typename T> |
209 |
struct SetLargeValue |
|
210 |
: public Howard<GR, LEN, SetLargeValueTraits<T> > { |
|
211 |
|
|
209 |
struct SetLargeCost |
|
210 |
: public HowardMmc<GR, CM, SetLargeCostTraits<T> > { |
|
211 |
typedef HowardMmc<GR, CM, SetLargeCostTraits<T> > Create; |
|
212 | 212 |
}; |
213 | 213 |
|
214 | 214 |
template <typename T> |
215 | 215 |
struct SetPathTraits : public Traits { |
216 | 216 |
typedef T Path; |
217 | 217 |
}; |
218 | 218 |
|
219 | 219 |
/// \brief \ref named-templ-param "Named parameter" for setting |
220 | 220 |
/// \c %Path type. |
221 | 221 |
/// |
222 | 222 |
/// \ref named-templ-param "Named parameter" for setting the \c %Path |
223 | 223 |
/// type of the found cycles. |
224 | 224 |
/// It must conform to the \ref lemon::concepts::Path "Path" concept |
225 | 225 |
/// and it must have an \c addBack() function. |
226 | 226 |
template <typename T> |
227 | 227 |
struct SetPath |
228 |
: public Howard<GR, LEN, SetPathTraits<T> > { |
|
229 |
typedef Howard<GR, LEN, SetPathTraits<T> > Create; |
|
228 |
: public HowardMmc<GR, CM, SetPathTraits<T> > { |
|
229 |
typedef HowardMmc<GR, CM, SetPathTraits<T> > Create; |
|
230 | 230 |
}; |
231 | 231 |
|
232 | 232 |
/// @} |
233 | 233 |
|
234 | 234 |
protected: |
235 | 235 |
|
236 |
|
|
236 |
HowardMmc() {} |
|
237 | 237 |
|
238 | 238 |
public: |
239 | 239 |
|
240 | 240 |
/// \brief Constructor. |
241 | 241 |
/// |
242 | 242 |
/// The constructor of the class. |
243 | 243 |
/// |
244 | 244 |
/// \param digraph The digraph the algorithm runs on. |
245 |
/// \param length The lengths (costs) of the arcs. |
|
246 |
Howard( const Digraph &digraph, |
|
247 |
const LengthMap &length ) : |
|
248 |
_gr(digraph), _length(length), _best_found(false), |
|
249 |
|
|
245 |
/// \param cost The costs of the arcs. |
|
246 |
HowardMmc( const Digraph &digraph, |
|
247 |
const CostMap &cost ) : |
|
248 |
_gr(digraph), _cost(cost), _best_found(false), |
|
249 |
_best_cost(0), _best_size(1), _cycle_path(NULL), _local_path(false), |
|
250 | 250 |
_policy(digraph), _reached(digraph), _level(digraph), _dist(digraph), |
251 | 251 |
_comp(digraph), _in_arcs(digraph), |
252 |
INF(std::numeric_limits<LargeValue>::has_infinity ? |
|
253 |
std::numeric_limits<LargeValue>::infinity() : |
|
254 |
|
|
252 |
INF(std::numeric_limits<LargeCost>::has_infinity ? |
|
253 |
std::numeric_limits<LargeCost>::infinity() : |
|
254 |
std::numeric_limits<LargeCost>::max()) |
|
255 | 255 |
{} |
256 | 256 |
|
257 | 257 |
/// Destructor. |
258 |
~ |
|
258 |
~HowardMmc() { |
|
259 | 259 |
if (_local_path) delete _cycle_path; |
260 | 260 |
} |
261 | 261 |
|
262 | 262 |
/// \brief Set the path structure for storing the found cycle. |
263 | 263 |
/// |
264 | 264 |
/// This function sets an external path structure for storing the |
265 | 265 |
/// found cycle. |
266 | 266 |
/// |
267 | 267 |
/// If you don't call this function before calling \ref run() or |
268 |
/// \ref |
|
268 |
/// \ref findCycleMean(), it will allocate a local \ref Path "path" |
|
269 | 269 |
/// structure. The destuctor deallocates this automatically |
270 | 270 |
/// allocated object, of course. |
271 | 271 |
/// |
272 | 272 |
/// \note The algorithm calls only the \ref lemon::Path::addBack() |
273 | 273 |
/// "addBack()" function of the given path structure. |
274 | 274 |
/// |
275 | 275 |
/// \return <tt>(*this)</tt> |
276 |
|
|
276 |
HowardMmc& cycle(Path &path) { |
|
277 | 277 |
if (_local_path) { |
278 | 278 |
delete _cycle_path; |
279 | 279 |
_local_path = false; |
280 | 280 |
} |
281 | 281 |
_cycle_path = &path; |
282 | 282 |
return *this; |
283 | 283 |
} |
284 | 284 |
|
285 | 285 |
/// \brief Set the tolerance used by the algorithm. |
286 | 286 |
/// |
287 | 287 |
/// This function sets the tolerance object used by the algorithm. |
288 | 288 |
/// |
289 | 289 |
/// \return <tt>(*this)</tt> |
290 |
|
|
290 |
HowardMmc& tolerance(const Tolerance& tolerance) { |
|
291 | 291 |
_tolerance = tolerance; |
292 | 292 |
return *this; |
293 | 293 |
} |
294 | 294 |
|
295 | 295 |
/// \brief Return a const reference to the tolerance. |
296 | 296 |
/// |
297 | 297 |
/// This function returns a const reference to the tolerance object |
298 | 298 |
/// used by the algorithm. |
299 | 299 |
const Tolerance& tolerance() const { |
300 | 300 |
return _tolerance; |
301 | 301 |
} |
302 | 302 |
|
303 | 303 |
/// \name Execution control |
304 | 304 |
/// The simplest way to execute the algorithm is to call the \ref run() |
305 | 305 |
/// function.\n |
306 |
/// If you only need the minimum mean length, you may call |
|
307 |
/// \ref findMinMean(). |
|
306 |
/// If you only need the minimum mean cost, you may call |
|
307 |
/// \ref findCycleMean(). |
|
308 | 308 |
|
309 | 309 |
/// @{ |
310 | 310 |
|
311 | 311 |
/// \brief Run the algorithm. |
312 | 312 |
/// |
313 | 313 |
/// This function runs the algorithm. |
314 | 314 |
/// It can be called more than once (e.g. if the underlying digraph |
315 |
/// and/or the arc |
|
315 |
/// and/or the arc costs have been modified). |
|
316 | 316 |
/// |
317 | 317 |
/// \return \c true if a directed cycle exists in the digraph. |
318 | 318 |
/// |
319 | 319 |
/// \note <tt>mmc.run()</tt> is just a shortcut of the following code. |
320 | 320 |
/// \code |
321 |
/// return mmc. |
|
321 |
/// return mmc.findCycleMean() && mmc.findCycle(); |
|
322 | 322 |
/// \endcode |
323 | 323 |
bool run() { |
324 |
return |
|
324 |
return findCycleMean() && findCycle(); |
|
325 | 325 |
} |
326 | 326 |
|
327 | 327 |
/// \brief Find the minimum cycle mean. |
328 | 328 |
/// |
329 |
/// This function finds the minimum mean |
|
329 |
/// This function finds the minimum mean cost of the directed |
|
330 | 330 |
/// cycles in the digraph. |
331 | 331 |
/// |
332 | 332 |
/// \return \c true if a directed cycle exists in the digraph. |
333 |
bool |
|
333 |
bool findCycleMean() { |
|
334 | 334 |
// Initialize and find strongly connected components |
335 | 335 |
init(); |
336 | 336 |
findComponents(); |
337 | 337 |
|
338 | 338 |
// Find the minimum cycle mean in the components |
339 | 339 |
for (int comp = 0; comp < _comp_num; ++comp) { |
340 | 340 |
// Find the minimum mean cycle in the current component |
341 | 341 |
if (!buildPolicyGraph(comp)) continue; |
342 | 342 |
while (true) { |
343 | 343 |
findPolicyCycle(); |
344 | 344 |
if (!computeNodeDistances()) break; |
345 | 345 |
} |
346 | 346 |
// Update the best cycle (global minimum mean cycle) |
347 | 347 |
if ( _curr_found && (!_best_found || |
348 |
|
|
348 |
_curr_cost * _best_size < _best_cost * _curr_size) ) { |
|
349 | 349 |
_best_found = true; |
350 |
|
|
350 |
_best_cost = _curr_cost; |
|
351 | 351 |
_best_size = _curr_size; |
352 | 352 |
_best_node = _curr_node; |
353 | 353 |
} |
354 | 354 |
} |
355 | 355 |
return _best_found; |
356 | 356 |
} |
357 | 357 |
|
358 | 358 |
/// \brief Find a minimum mean directed cycle. |
359 | 359 |
/// |
360 |
/// This function finds a directed cycle of minimum mean length |
|
361 |
/// in the digraph using the data computed by findMinMean(). |
|
360 |
/// This function finds a directed cycle of minimum mean cost |
|
361 |
/// in the digraph using the data computed by findCycleMean(). |
|
362 | 362 |
/// |
363 | 363 |
/// \return \c true if a directed cycle exists in the digraph. |
364 | 364 |
/// |
365 |
/// \pre \ref |
|
365 |
/// \pre \ref findCycleMean() must be called before using this function. |
|
366 | 366 |
bool findCycle() { |
367 | 367 |
if (!_best_found) return false; |
368 | 368 |
_cycle_path->addBack(_policy[_best_node]); |
369 | 369 |
for ( Node v = _best_node; |
370 | 370 |
(v = _gr.target(_policy[v])) != _best_node; ) { |
371 | 371 |
_cycle_path->addBack(_policy[v]); |
372 | 372 |
} |
373 | 373 |
return true; |
374 | 374 |
} |
375 | 375 |
|
376 | 376 |
/// @} |
377 | 377 |
|
378 | 378 |
/// \name Query Functions |
379 | 379 |
/// The results of the algorithm can be obtained using these |
380 | 380 |
/// functions.\n |
381 | 381 |
/// The algorithm should be executed before using them. |
382 | 382 |
|
383 | 383 |
/// @{ |
384 | 384 |
|
385 |
/// \brief Return the total |
|
385 |
/// \brief Return the total cost of the found cycle. |
|
386 | 386 |
/// |
387 |
/// This function returns the total |
|
387 |
/// This function returns the total cost of the found cycle. |
|
388 | 388 |
/// |
389 |
/// \pre \ref run() or \ref |
|
389 |
/// \pre \ref run() or \ref findCycleMean() must be called before |
|
390 | 390 |
/// using this function. |
391 |
Value cycleLength() const { |
|
392 |
return static_cast<Value>(_best_length); |
|
391 |
Cost cycleCost() const { |
|
392 |
return static_cast<Cost>(_best_cost); |
|
393 | 393 |
} |
394 | 394 |
|
395 | 395 |
/// \brief Return the number of arcs on the found cycle. |
396 | 396 |
/// |
397 | 397 |
/// This function returns the number of arcs on the found cycle. |
398 | 398 |
/// |
399 |
/// \pre \ref run() or \ref |
|
399 |
/// \pre \ref run() or \ref findCycleMean() must be called before |
|
400 | 400 |
/// using this function. |
401 |
int |
|
401 |
int cycleSize() const { |
|
402 | 402 |
return _best_size; |
403 | 403 |
} |
404 | 404 |
|
405 |
/// \brief Return the mean |
|
405 |
/// \brief Return the mean cost of the found cycle. |
|
406 | 406 |
/// |
407 |
/// This function returns the mean |
|
407 |
/// This function returns the mean cost of the found cycle. |
|
408 | 408 |
/// |
409 | 409 |
/// \note <tt>alg.cycleMean()</tt> is just a shortcut of the |
410 | 410 |
/// following code. |
411 | 411 |
/// \code |
412 |
/// return static_cast<double>(alg. |
|
412 |
/// return static_cast<double>(alg.cycleCost()) / alg.cycleSize(); |
|
413 | 413 |
/// \endcode |
414 | 414 |
/// |
415 |
/// \pre \ref run() or \ref |
|
415 |
/// \pre \ref run() or \ref findCycleMean() must be called before |
|
416 | 416 |
/// using this function. |
417 | 417 |
double cycleMean() const { |
418 |
return static_cast<double>( |
|
418 |
return static_cast<double>(_best_cost) / _best_size; |
|
419 | 419 |
} |
420 | 420 |
|
421 | 421 |
/// \brief Return the found cycle. |
422 | 422 |
/// |
423 | 423 |
/// This function returns a const reference to the path structure |
424 | 424 |
/// storing the found cycle. |
425 | 425 |
/// |
426 | 426 |
/// \pre \ref run() or \ref findCycle() must be called before using |
427 | 427 |
/// this function. |
428 | 428 |
const Path& cycle() const { |
429 | 429 |
return *_cycle_path; |
430 | 430 |
} |
431 | 431 |
|
432 | 432 |
///@} |
433 | 433 |
|
434 | 434 |
private: |
435 | 435 |
|
436 | 436 |
// Initialize |
437 | 437 |
void init() { |
438 | 438 |
if (!_cycle_path) { |
439 | 439 |
_local_path = true; |
440 | 440 |
_cycle_path = new Path; |
441 | 441 |
} |
442 | 442 |
_queue.resize(countNodes(_gr)); |
443 | 443 |
_best_found = false; |
444 |
|
|
444 |
_best_cost = 0; |
|
445 | 445 |
_best_size = 1; |
446 | 446 |
_cycle_path->clear(); |
447 | 447 |
} |
448 | 448 |
|
449 | 449 |
// Find strongly connected components and initialize _comp_nodes |
450 | 450 |
// and _in_arcs |
451 | 451 |
void findComponents() { |
452 | 452 |
_comp_num = stronglyConnectedComponents(_gr, _comp); |
453 | 453 |
_comp_nodes.resize(_comp_num); |
454 | 454 |
if (_comp_num == 1) { |
455 | 455 |
_comp_nodes[0].clear(); |
456 | 456 |
for (NodeIt n(_gr); n != INVALID; ++n) { |
457 | 457 |
_comp_nodes[0].push_back(n); |
458 | 458 |
_in_arcs[n].clear(); |
459 | 459 |
for (InArcIt a(_gr, n); a != INVALID; ++a) { |
460 | 460 |
_in_arcs[n].push_back(a); |
461 | 461 |
} |
462 | 462 |
} |
463 | 463 |
} else { |
464 | 464 |
for (int i = 0; i < _comp_num; ++i) |
465 | 465 |
_comp_nodes[i].clear(); |
466 | 466 |
for (NodeIt n(_gr); n != INVALID; ++n) { |
467 | 467 |
int k = _comp[n]; |
468 | 468 |
_comp_nodes[k].push_back(n); |
469 | 469 |
_in_arcs[n].clear(); |
470 | 470 |
for (InArcIt a(_gr, n); a != INVALID; ++a) { |
471 | 471 |
if (_comp[_gr.source(a)] == k) _in_arcs[n].push_back(a); |
472 | 472 |
} |
473 | 473 |
} |
474 | 474 |
} |
475 | 475 |
} |
476 | 476 |
|
477 | 477 |
// Build the policy graph in the given strongly connected component |
478 | 478 |
// (the out-degree of every node is 1) |
479 | 479 |
bool buildPolicyGraph(int comp) { |
480 | 480 |
_nodes = &(_comp_nodes[comp]); |
481 | 481 |
if (_nodes->size() < 1 || |
482 | 482 |
(_nodes->size() == 1 && _in_arcs[(*_nodes)[0]].size() == 0)) { |
483 | 483 |
return false; |
484 | 484 |
} |
485 | 485 |
for (int i = 0; i < int(_nodes->size()); ++i) { |
486 | 486 |
_dist[(*_nodes)[i]] = INF; |
487 | 487 |
} |
488 | 488 |
Node u, v; |
489 | 489 |
Arc e; |
490 | 490 |
for (int i = 0; i < int(_nodes->size()); ++i) { |
491 | 491 |
v = (*_nodes)[i]; |
492 | 492 |
for (int j = 0; j < int(_in_arcs[v].size()); ++j) { |
493 | 493 |
e = _in_arcs[v][j]; |
494 | 494 |
u = _gr.source(e); |
495 |
if (_length[e] < _dist[u]) { |
|
496 |
_dist[u] = _length[e]; |
|
495 |
if (_cost[e] < _dist[u]) { |
|
496 |
_dist[u] = _cost[e]; |
|
497 | 497 |
_policy[u] = e; |
498 | 498 |
} |
499 | 499 |
} |
500 | 500 |
} |
501 | 501 |
return true; |
502 | 502 |
} |
503 | 503 |
|
504 | 504 |
// Find the minimum mean cycle in the policy graph |
505 | 505 |
void findPolicyCycle() { |
506 | 506 |
for (int i = 0; i < int(_nodes->size()); ++i) { |
507 | 507 |
_level[(*_nodes)[i]] = -1; |
508 | 508 |
} |
509 |
|
|
509 |
LargeCost ccost; |
|
510 | 510 |
int csize; |
511 | 511 |
Node u, v; |
512 | 512 |
_curr_found = false; |
513 | 513 |
for (int i = 0; i < int(_nodes->size()); ++i) { |
514 | 514 |
u = (*_nodes)[i]; |
515 | 515 |
if (_level[u] >= 0) continue; |
516 | 516 |
for (; _level[u] < 0; u = _gr.target(_policy[u])) { |
517 | 517 |
_level[u] = i; |
518 | 518 |
} |
519 | 519 |
if (_level[u] == i) { |
520 | 520 |
// A cycle is found |
521 |
|
|
521 |
ccost = _cost[_policy[u]]; |
|
522 | 522 |
csize = 1; |
523 | 523 |
for (v = u; (v = _gr.target(_policy[v])) != u; ) { |
524 |
|
|
524 |
ccost += _cost[_policy[v]]; |
|
525 | 525 |
++csize; |
526 | 526 |
} |
527 | 527 |
if ( !_curr_found || |
528 |
( |
|
528 |
(ccost * _curr_size < _curr_cost * csize) ) { |
|
529 | 529 |
_curr_found = true; |
530 |
|
|
530 |
_curr_cost = ccost; |
|
531 | 531 |
_curr_size = csize; |
532 | 532 |
_curr_node = u; |
533 | 533 |
} |
534 | 534 |
} |
535 | 535 |
} |
536 | 536 |
} |
537 | 537 |
|
538 | 538 |
// Contract the policy graph and compute node distances |
539 | 539 |
bool computeNodeDistances() { |
540 | 540 |
// Find the component of the main cycle and compute node distances |
541 | 541 |
// using reverse BFS |
542 | 542 |
for (int i = 0; i < int(_nodes->size()); ++i) { |
543 | 543 |
_reached[(*_nodes)[i]] = false; |
544 | 544 |
} |
545 | 545 |
_qfront = _qback = 0; |
546 | 546 |
_queue[0] = _curr_node; |
547 | 547 |
_reached[_curr_node] = true; |
548 | 548 |
_dist[_curr_node] = 0; |
549 | 549 |
Node u, v; |
550 | 550 |
Arc e; |
551 | 551 |
while (_qfront <= _qback) { |
552 | 552 |
v = _queue[_qfront++]; |
553 | 553 |
for (int j = 0; j < int(_in_arcs[v].size()); ++j) { |
554 | 554 |
e = _in_arcs[v][j]; |
555 | 555 |
u = _gr.source(e); |
556 | 556 |
if (_policy[u] == e && !_reached[u]) { |
557 | 557 |
_reached[u] = true; |
558 |
_dist[u] = _dist[v] + |
|
558 |
_dist[u] = _dist[v] + _cost[e] * _curr_size - _curr_cost; |
|
559 | 559 |
_queue[++_qback] = u; |
560 | 560 |
} |
561 | 561 |
} |
562 | 562 |
} |
563 | 563 |
|
564 | 564 |
// Connect all other nodes to this component and compute node |
565 | 565 |
// distances using reverse BFS |
566 | 566 |
_qfront = 0; |
567 | 567 |
while (_qback < int(_nodes->size())-1) { |
568 | 568 |
v = _queue[_qfront++]; |
569 | 569 |
for (int j = 0; j < int(_in_arcs[v].size()); ++j) { |
570 | 570 |
e = _in_arcs[v][j]; |
571 | 571 |
u = _gr.source(e); |
572 | 572 |
if (!_reached[u]) { |
573 | 573 |
_reached[u] = true; |
574 | 574 |
_policy[u] = e; |
575 |
_dist[u] = _dist[v] + |
|
575 |
_dist[u] = _dist[v] + _cost[e] * _curr_size - _curr_cost; |
|
576 | 576 |
_queue[++_qback] = u; |
577 | 577 |
} |
578 | 578 |
} |
579 | 579 |
} |
580 | 580 |
|
581 | 581 |
// Improve node distances |
582 | 582 |
bool improved = false; |
583 | 583 |
for (int i = 0; i < int(_nodes->size()); ++i) { |
584 | 584 |
v = (*_nodes)[i]; |
585 | 585 |
for (int j = 0; j < int(_in_arcs[v].size()); ++j) { |
586 | 586 |
e = _in_arcs[v][j]; |
587 | 587 |
u = _gr.source(e); |
588 |
|
|
588 |
LargeCost delta = _dist[v] + _cost[e] * _curr_size - _curr_cost; |
|
589 | 589 |
if (_tolerance.less(delta, _dist[u])) { |
590 | 590 |
_dist[u] = delta; |
591 | 591 |
_policy[u] = e; |
592 | 592 |
improved = true; |
593 | 593 |
} |
594 | 594 |
} |
595 | 595 |
} |
596 | 596 |
return improved; |
597 | 597 |
} |
598 | 598 |
|
599 |
}; //class |
|
599 |
}; //class HowardMmc |
|
600 | 600 |
|
601 | 601 |
///@} |
602 | 602 |
|
603 | 603 |
} //namespace lemon |
604 | 604 |
|
605 |
#endif // |
|
605 |
#endif //LEMON_HOWARD_MMC_H |
1 | 1 |
/* -*- C++ -*- |
2 | 2 |
* |
3 | 3 |
* This file is a part of LEMON, a generic C++ optimization library |
4 | 4 |
* |
5 | 5 |
* Copyright (C) 2003-2008 |
6 | 6 |
* Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport |
7 | 7 |
* (Egervary Research Group on Combinatorial Optimization, EGRES). |
8 | 8 |
* |
9 | 9 |
* Permission to use, modify and distribute this software is granted |
10 | 10 |
* provided that this copyright notice appears in all copies. For |
11 | 11 |
* precise terms see the accompanying LICENSE file. |
12 | 12 |
* |
13 | 13 |
* This software is provided "AS IS" with no warranty of any kind, |
14 | 14 |
* express or implied, and with no claim as to its suitability for any |
15 | 15 |
* purpose. |
16 | 16 |
* |
17 | 17 |
*/ |
18 | 18 |
|
19 |
#ifndef LEMON_KARP_H |
|
20 |
#define LEMON_KARP_H |
|
19 |
#ifndef LEMON_KARP_MMC_H |
|
20 |
#define LEMON_KARP_MMC_H |
|
21 | 21 |
|
22 | 22 |
/// \ingroup min_mean_cycle |
23 | 23 |
/// |
24 | 24 |
/// \file |
25 | 25 |
/// \brief Karp's algorithm for finding a minimum mean cycle. |
26 | 26 |
|
27 | 27 |
#include <vector> |
28 | 28 |
#include <limits> |
29 | 29 |
#include <lemon/core.h> |
30 | 30 |
#include <lemon/path.h> |
31 | 31 |
#include <lemon/tolerance.h> |
32 | 32 |
#include <lemon/connectivity.h> |
33 | 33 |
|
34 | 34 |
namespace lemon { |
35 | 35 |
|
36 |
/// \brief Default traits class of |
|
36 |
/// \brief Default traits class of KarpMmc class. |
|
37 | 37 |
/// |
38 |
/// Default traits class of |
|
38 |
/// Default traits class of KarpMmc class. |
|
39 | 39 |
/// \tparam GR The type of the digraph. |
40 |
/// \tparam |
|
40 |
/// \tparam CM The type of the cost map. |
|
41 | 41 |
/// It must conform to the \ref concepts::ReadMap "ReadMap" concept. |
42 | 42 |
#ifdef DOXYGEN |
43 |
template <typename GR, typename |
|
43 |
template <typename GR, typename CM> |
|
44 | 44 |
#else |
45 |
template <typename GR, typename LEN, |
|
46 |
bool integer = std::numeric_limits<typename LEN::Value>::is_integer> |
|
45 |
template <typename GR, typename CM, |
|
46 |
bool integer = std::numeric_limits<typename CM::Value>::is_integer> |
|
47 | 47 |
#endif |
48 |
struct |
|
48 |
struct KarpMmcDefaultTraits |
|
49 | 49 |
{ |
50 | 50 |
/// The type of the digraph |
51 | 51 |
typedef GR Digraph; |
52 |
/// The type of the length map |
|
53 |
typedef LEN LengthMap; |
|
54 |
/// The type of the arc lengths |
|
55 |
typedef typename LengthMap::Value Value; |
|
52 |
/// The type of the cost map |
|
53 |
typedef CM CostMap; |
|
54 |
/// The type of the arc costs |
|
55 |
typedef typename CostMap::Value Cost; |
|
56 | 56 |
|
57 |
/// \brief The large |
|
57 |
/// \brief The large cost type used for internal computations |
|
58 | 58 |
/// |
59 |
/// The large value type used for internal computations. |
|
60 |
/// It is \c long \c long if the \c Value type is integer, |
|
59 |
/// The large cost type used for internal computations. |
|
60 |
/// It is \c long \c long if the \c Cost type is integer, |
|
61 | 61 |
/// otherwise it is \c double. |
62 |
/// \c Value must be convertible to \c LargeValue. |
|
63 |
typedef double LargeValue; |
|
62 |
/// \c Cost must be convertible to \c LargeCost. |
|
63 |
typedef double LargeCost; |
|
64 | 64 |
|
65 | 65 |
/// The tolerance type used for internal computations |
66 |
typedef lemon::Tolerance< |
|
66 |
typedef lemon::Tolerance<LargeCost> Tolerance; |
|
67 | 67 |
|
68 | 68 |
/// \brief The path type of the found cycles |
69 | 69 |
/// |
70 | 70 |
/// The path type of the found cycles. |
71 | 71 |
/// It must conform to the \ref lemon::concepts::Path "Path" concept |
72 | 72 |
/// and it must have an \c addFront() function. |
73 | 73 |
typedef lemon::Path<Digraph> Path; |
74 | 74 |
}; |
75 | 75 |
|
76 |
// Default traits class for integer value types |
|
77 |
template <typename GR, typename LEN> |
|
78 |
|
|
76 |
// Default traits class for integer cost types |
|
77 |
template <typename GR, typename CM> |
|
78 |
struct KarpMmcDefaultTraits<GR, CM, true> |
|
79 | 79 |
{ |
80 | 80 |
typedef GR Digraph; |
81 |
typedef LEN LengthMap; |
|
82 |
typedef typename LengthMap::Value Value; |
|
81 |
typedef CM CostMap; |
|
82 |
typedef typename CostMap::Value Cost; |
|
83 | 83 |
#ifdef LEMON_HAVE_LONG_LONG |
84 |
typedef long long |
|
84 |
typedef long long LargeCost; |
|
85 | 85 |
#else |
86 |
typedef long |
|
86 |
typedef long LargeCost; |
|
87 | 87 |
#endif |
88 |
typedef lemon::Tolerance< |
|
88 |
typedef lemon::Tolerance<LargeCost> Tolerance; |
|
89 | 89 |
typedef lemon::Path<Digraph> Path; |
90 | 90 |
}; |
91 | 91 |
|
92 | 92 |
|
93 | 93 |
/// \addtogroup min_mean_cycle |
94 | 94 |
/// @{ |
95 | 95 |
|
96 | 96 |
/// \brief Implementation of Karp's algorithm for finding a minimum |
97 | 97 |
/// mean cycle. |
98 | 98 |
/// |
99 | 99 |
/// This class implements Karp's algorithm for finding a directed |
100 |
/// cycle of minimum mean |
|
100 |
/// cycle of minimum mean cost in a digraph |
|
101 | 101 |
/// \ref amo93networkflows, \ref dasdan98minmeancycle. |
102 | 102 |
/// It runs in time O(ne) and uses space O(n<sup>2</sup>+e). |
103 | 103 |
/// |
104 | 104 |
/// \tparam GR The type of the digraph the algorithm runs on. |
105 |
/// \tparam |
|
105 |
/// \tparam CM The type of the cost map. The default |
|
106 | 106 |
/// map type is \ref concepts::Digraph::ArcMap "GR::ArcMap<int>". |
107 | 107 |
/// \tparam TR The traits class that defines various types used by the |
108 |
/// algorithm. By default, it is \ref KarpDefaultTraits |
|
109 |
/// "KarpDefaultTraits<GR, LEN>". |
|
108 |
/// algorithm. By default, it is \ref KarpMmcDefaultTraits |
|
109 |
/// "KarpMmcDefaultTraits<GR, CM>". |
|
110 | 110 |
/// In most cases, this parameter should not be set directly, |
111 | 111 |
/// consider to use the named template parameters instead. |
112 | 112 |
#ifdef DOXYGEN |
113 |
template <typename GR, typename |
|
113 |
template <typename GR, typename CM, typename TR> |
|
114 | 114 |
#else |
115 | 115 |
template < typename GR, |
116 |
typename LEN = typename GR::template ArcMap<int>, |
|
117 |
typename TR = KarpDefaultTraits<GR, LEN> > |
|
116 |
typename CM = typename GR::template ArcMap<int>, |
|
117 |
typename TR = KarpMmcDefaultTraits<GR, CM> > |
|
118 | 118 |
#endif |
119 |
class |
|
119 |
class KarpMmc |
|
120 | 120 |
{ |
121 | 121 |
public: |
122 | 122 |
|
123 | 123 |
/// The type of the digraph |
124 | 124 |
typedef typename TR::Digraph Digraph; |
125 |
/// The type of the length map |
|
126 |
typedef typename TR::LengthMap LengthMap; |
|
127 |
/// The type of the arc lengths |
|
128 |
typedef typename TR::Value Value; |
|
125 |
/// The type of the cost map |
|
126 |
typedef typename TR::CostMap CostMap; |
|
127 |
/// The type of the arc costs |
|
128 |
typedef typename TR::Cost Cost; |
|
129 | 129 |
|
130 |
/// \brief The large |
|
130 |
/// \brief The large cost type |
|
131 | 131 |
/// |
132 |
/// The large value type used for internal computations. |
|
133 |
/// By default, it is \c long \c long if the \c Value type is integer, |
|
132 |
/// The large cost type used for internal computations. |
|
133 |
/// By default, it is \c long \c long if the \c Cost type is integer, |
|
134 | 134 |
/// otherwise it is \c double. |
135 |
typedef typename TR:: |
|
135 |
typedef typename TR::LargeCost LargeCost; |
|
136 | 136 |
|
137 | 137 |
/// The tolerance type |
138 | 138 |
typedef typename TR::Tolerance Tolerance; |
139 | 139 |
|
140 | 140 |
/// \brief The path type of the found cycles |
141 | 141 |
/// |
142 | 142 |
/// The path type of the found cycles. |
143 |
/// Using the \ref |
|
143 |
/// Using the \ref KarpMmcDefaultTraits "default traits class", |
|
144 | 144 |
/// it is \ref lemon::Path "Path<Digraph>". |
145 | 145 |
typedef typename TR::Path Path; |
146 | 146 |
|
147 |
/// The \ref |
|
147 |
/// The \ref KarpMmcDefaultTraits "traits class" of the algorithm |
|
148 | 148 |
typedef TR Traits; |
149 | 149 |
|
150 | 150 |
private: |
151 | 151 |
|
152 | 152 |
TEMPLATE_DIGRAPH_TYPEDEFS(Digraph); |
153 | 153 |
|
154 | 154 |
// Data sturcture for path data |
155 | 155 |
struct PathData |
156 | 156 |
{ |
157 |
|
|
157 |
LargeCost dist; |
|
158 | 158 |
Arc pred; |
159 |
PathData( |
|
159 |
PathData(LargeCost d, Arc p = INVALID) : |
|
160 | 160 |
dist(d), pred(p) {} |
161 | 161 |
}; |
162 | 162 |
|
163 | 163 |
typedef typename Digraph::template NodeMap<std::vector<PathData> > |
164 | 164 |
PathDataNodeMap; |
165 | 165 |
|
166 | 166 |
private: |
167 | 167 |
|
168 | 168 |
// The digraph the algorithm runs on |
169 | 169 |
const Digraph &_gr; |
170 |
// The length of the arcs |
|
171 |
const LengthMap &_length; |
|
170 |
// The cost of the arcs |
|
171 |
const CostMap &_cost; |
|
172 | 172 |
|
173 | 173 |
// Data for storing the strongly connected components |
174 | 174 |
int _comp_num; |
175 | 175 |
typename Digraph::template NodeMap<int> _comp; |
176 | 176 |
std::vector<std::vector<Node> > _comp_nodes; |
177 | 177 |
std::vector<Node>* _nodes; |
178 | 178 |
typename Digraph::template NodeMap<std::vector<Arc> > _out_arcs; |
179 | 179 |
|
180 | 180 |
// Data for the found cycle |
181 |
|
|
181 |
LargeCost _cycle_cost; |
|
182 | 182 |
int _cycle_size; |
183 | 183 |
Node _cycle_node; |
184 | 184 |
|
185 | 185 |
Path *_cycle_path; |
186 | 186 |
bool _local_path; |
187 | 187 |
|
188 | 188 |
// Node map for storing path data |
189 | 189 |
PathDataNodeMap _data; |
190 | 190 |
// The processed nodes in the last round |
191 | 191 |
std::vector<Node> _process; |
192 | 192 |
|
193 | 193 |
Tolerance _tolerance; |
194 | 194 |
|
195 | 195 |
// Infinite constant |
196 |
const |
|
196 |
const LargeCost INF; |
|
197 | 197 |
|
198 | 198 |
public: |
199 | 199 |
|
200 | 200 |
/// \name Named Template Parameters |
201 | 201 |
/// @{ |
202 | 202 |
|
203 | 203 |
template <typename T> |
204 |
struct SetLargeValueTraits : public Traits { |
|
205 |
typedef T LargeValue; |
|
204 |
struct SetLargeCostTraits : public Traits { |
|
205 |
typedef T LargeCost; |
|
206 | 206 |
typedef lemon::Tolerance<T> Tolerance; |
207 | 207 |
}; |
208 | 208 |
|
209 | 209 |
/// \brief \ref named-templ-param "Named parameter" for setting |
210 |
/// \c |
|
210 |
/// \c LargeCost type. |
|
211 | 211 |
/// |
212 |
/// \ref named-templ-param "Named parameter" for setting \c |
|
212 |
/// \ref named-templ-param "Named parameter" for setting \c LargeCost |
|
213 | 213 |
/// type. It is used for internal computations in the algorithm. |
214 | 214 |
template <typename T> |
215 |
struct SetLargeValue |
|
216 |
: public Karp<GR, LEN, SetLargeValueTraits<T> > { |
|
217 |
|
|
215 |
struct SetLargeCost |
|
216 |
: public KarpMmc<GR, CM, SetLargeCostTraits<T> > { |
|
217 |
typedef KarpMmc<GR, CM, SetLargeCostTraits<T> > Create; |
|
218 | 218 |
}; |
219 | 219 |
|
220 | 220 |
template <typename T> |
221 | 221 |
struct SetPathTraits : public Traits { |
222 | 222 |
typedef T Path; |
223 | 223 |
}; |
224 | 224 |
|
225 | 225 |
/// \brief \ref named-templ-param "Named parameter" for setting |
226 | 226 |
/// \c %Path type. |
227 | 227 |
/// |
228 | 228 |
/// \ref named-templ-param "Named parameter" for setting the \c %Path |
229 | 229 |
/// type of the found cycles. |
230 | 230 |
/// It must conform to the \ref lemon::concepts::Path "Path" concept |
231 | 231 |
/// and it must have an \c addFront() function. |
232 | 232 |
template <typename T> |
233 | 233 |
struct SetPath |
234 |
: public Karp<GR, LEN, SetPathTraits<T> > { |
|
235 |
typedef Karp<GR, LEN, SetPathTraits<T> > Create; |
|
234 |
: public KarpMmc<GR, CM, SetPathTraits<T> > { |
|
235 |
typedef KarpMmc<GR, CM, SetPathTraits<T> > Create; |
|
236 | 236 |
}; |
237 | 237 |
|
238 | 238 |
/// @} |
239 | 239 |
|
240 | 240 |
protected: |
241 | 241 |
|
242 |
|
|
242 |
KarpMmc() {} |
|
243 | 243 |
|
244 | 244 |
public: |
245 | 245 |
|
246 | 246 |
/// \brief Constructor. |
247 | 247 |
/// |
248 | 248 |
/// The constructor of the class. |
249 | 249 |
/// |
250 | 250 |
/// \param digraph The digraph the algorithm runs on. |
251 |
/// \param length The lengths (costs) of the arcs. |
|
252 |
Karp( const Digraph &digraph, |
|
253 |
const LengthMap &length ) : |
|
254 |
_gr(digraph), _length(length), _comp(digraph), _out_arcs(digraph), |
|
255 |
|
|
251 |
/// \param cost The costs of the arcs. |
|
252 |
KarpMmc( const Digraph &digraph, |
|
253 |
const CostMap &cost ) : |
|
254 |
_gr(digraph), _cost(cost), _comp(digraph), _out_arcs(digraph), |
|
255 |
_cycle_cost(0), _cycle_size(1), _cycle_node(INVALID), |
|
256 | 256 |
_cycle_path(NULL), _local_path(false), _data(digraph), |
257 |
INF(std::numeric_limits<LargeValue>::has_infinity ? |
|
258 |
std::numeric_limits<LargeValue>::infinity() : |
|
259 |
|
|
257 |
INF(std::numeric_limits<LargeCost>::has_infinity ? |
|
258 |
std::numeric_limits<LargeCost>::infinity() : |
|
259 |
std::numeric_limits<LargeCost>::max()) |
|
260 | 260 |
{} |
261 | 261 |
|
262 | 262 |
/// Destructor. |
263 |
~ |
|
263 |
~KarpMmc() { |
|
264 | 264 |
if (_local_path) delete _cycle_path; |
265 | 265 |
} |
266 | 266 |
|
267 | 267 |
/// \brief Set the path structure for storing the found cycle. |
268 | 268 |
/// |
269 | 269 |
/// This function sets an external path structure for storing the |
270 | 270 |
/// found cycle. |
271 | 271 |
/// |
272 | 272 |
/// If you don't call this function before calling \ref run() or |
273 |
/// \ref |
|
273 |
/// \ref findCycleMean(), it will allocate a local \ref Path "path" |
|
274 | 274 |
/// structure. The destuctor deallocates this automatically |
275 | 275 |
/// allocated object, of course. |
276 | 276 |
/// |
277 | 277 |
/// \note The algorithm calls only the \ref lemon::Path::addFront() |
278 | 278 |
/// "addFront()" function of the given path structure. |
279 | 279 |
/// |
280 | 280 |
/// \return <tt>(*this)</tt> |
281 |
|
|
281 |
KarpMmc& cycle(Path &path) { |
|
282 | 282 |
if (_local_path) { |
283 | 283 |
delete _cycle_path; |
284 | 284 |
_local_path = false; |
285 | 285 |
} |
286 | 286 |
_cycle_path = &path; |
287 | 287 |
return *this; |
288 | 288 |
} |
289 | 289 |
|
290 | 290 |
/// \brief Set the tolerance used by the algorithm. |
291 | 291 |
/// |
292 | 292 |
/// This function sets the tolerance object used by the algorithm. |
293 | 293 |
/// |
294 | 294 |
/// \return <tt>(*this)</tt> |
295 |
|
|
295 |
KarpMmc& tolerance(const Tolerance& tolerance) { |
|
296 | 296 |
_tolerance = tolerance; |
297 | 297 |
return *this; |
298 | 298 |
} |
299 | 299 |
|
300 | 300 |
/// \brief Return a const reference to the tolerance. |
301 | 301 |
/// |
302 | 302 |
/// This function returns a const reference to the tolerance object |
303 | 303 |
/// used by the algorithm. |
304 | 304 |
const Tolerance& tolerance() const { |
305 | 305 |
return _tolerance; |
306 | 306 |
} |
307 | 307 |
|
308 | 308 |
/// \name Execution control |
309 | 309 |
/// The simplest way to execute the algorithm is to call the \ref run() |
310 | 310 |
/// function.\n |
311 |
/// If you only need the minimum mean length, you may call |
|
312 |
/// \ref findMinMean(). |
|
311 |
/// If you only need the minimum mean cost, you may call |
|
312 |
/// \ref findCycleMean(). |
|
313 | 313 |
|
314 | 314 |
/// @{ |
315 | 315 |
|
316 | 316 |
/// \brief Run the algorithm. |
317 | 317 |
/// |
318 | 318 |
/// This function runs the algorithm. |
319 | 319 |
/// It can be called more than once (e.g. if the underlying digraph |
320 |
/// and/or the arc |
|
320 |
/// and/or the arc costs have been modified). |
|
321 | 321 |
/// |
322 | 322 |
/// \return \c true if a directed cycle exists in the digraph. |
323 | 323 |
/// |
324 | 324 |
/// \note <tt>mmc.run()</tt> is just a shortcut of the following code. |
325 | 325 |
/// \code |
326 |
/// return mmc. |
|
326 |
/// return mmc.findCycleMean() && mmc.findCycle(); |
|
327 | 327 |
/// \endcode |
328 | 328 |
bool run() { |
329 |
return |
|
329 |
return findCycleMean() && findCycle(); |
|
330 | 330 |
} |
331 | 331 |
|
332 | 332 |
/// \brief Find the minimum cycle mean. |
333 | 333 |
/// |
334 |
/// This function finds the minimum mean |
|
334 |
/// This function finds the minimum mean cost of the directed |
|
335 | 335 |
/// cycles in the digraph. |
336 | 336 |
/// |
337 | 337 |
/// \return \c true if a directed cycle exists in the digraph. |
338 |
bool |
|
338 |
bool findCycleMean() { |
|
339 | 339 |
// Initialization and find strongly connected components |
340 | 340 |
init(); |
341 | 341 |
findComponents(); |
342 | 342 |
|
343 | 343 |
// Find the minimum cycle mean in the components |
344 | 344 |
for (int comp = 0; comp < _comp_num; ++comp) { |
345 | 345 |
if (!initComponent(comp)) continue; |
346 | 346 |
processRounds(); |
347 | 347 |
updateMinMean(); |
348 | 348 |
} |
349 | 349 |
return (_cycle_node != INVALID); |
350 | 350 |
} |
351 | 351 |
|
352 | 352 |
/// \brief Find a minimum mean directed cycle. |
353 | 353 |
/// |
354 |
/// This function finds a directed cycle of minimum mean length |
|
355 |
/// in the digraph using the data computed by findMinMean(). |
|
354 |
/// This function finds a directed cycle of minimum mean cost |
|
355 |
/// in the digraph using the data computed by findCycleMean(). |
|
356 | 356 |
/// |
357 | 357 |
/// \return \c true if a directed cycle exists in the digraph. |
358 | 358 |
/// |
359 |
/// \pre \ref |
|
359 |
/// \pre \ref findCycleMean() must be called before using this function. |
|
360 | 360 |
bool findCycle() { |
361 | 361 |
if (_cycle_node == INVALID) return false; |
362 | 362 |
IntNodeMap reached(_gr, -1); |
363 | 363 |
int r = _data[_cycle_node].size(); |
364 | 364 |
Node u = _cycle_node; |
365 | 365 |
while (reached[u] < 0) { |
366 | 366 |
reached[u] = --r; |
367 | 367 |
u = _gr.source(_data[u][r].pred); |
368 | 368 |
} |
369 | 369 |
r = reached[u]; |
370 | 370 |
Arc e = _data[u][r].pred; |
371 | 371 |
_cycle_path->addFront(e); |
372 |
|
|
372 |
_cycle_cost = _cost[e]; |
|
373 | 373 |
_cycle_size = 1; |
374 | 374 |
Node v; |
375 | 375 |
while ((v = _gr.source(e)) != u) { |
376 | 376 |
e = _data[v][--r].pred; |
377 | 377 |
_cycle_path->addFront(e); |
378 |
|
|
378 |
_cycle_cost += _cost[e]; |
|
379 | 379 |
++_cycle_size; |
380 | 380 |
} |
381 | 381 |
return true; |
382 | 382 |
} |
383 | 383 |
|
384 | 384 |
/// @} |
385 | 385 |
|
386 | 386 |
/// \name Query Functions |
387 | 387 |
/// The results of the algorithm can be obtained using these |
388 | 388 |
/// functions.\n |
389 | 389 |
/// The algorithm should be executed before using them. |
390 | 390 |
|
391 | 391 |
/// @{ |
392 | 392 |
|
393 |
/// \brief Return the total |
|
393 |
/// \brief Return the total cost of the found cycle. |
|
394 | 394 |
/// |
395 |
/// This function returns the total |
|
395 |
/// This function returns the total cost of the found cycle. |
|
396 | 396 |
/// |
397 |
/// \pre \ref run() or \ref |
|
397 |
/// \pre \ref run() or \ref findCycleMean() must be called before |
|
398 | 398 |
/// using this function. |
399 |
Value cycleLength() const { |
|
400 |
return static_cast<Value>(_cycle_length); |
|
399 |
Cost cycleCost() const { |
|
400 |
return static_cast<Cost>(_cycle_cost); |
|
401 | 401 |
} |
402 | 402 |
|
403 | 403 |
/// \brief Return the number of arcs on the found cycle. |
404 | 404 |
/// |
405 | 405 |
/// This function returns the number of arcs on the found cycle. |
406 | 406 |
/// |
407 |
/// \pre \ref run() or \ref |
|
407 |
/// \pre \ref run() or \ref findCycleMean() must be called before |
|
408 | 408 |
/// using this function. |
409 |
int |
|
409 |
int cycleSize() const { |
|
410 | 410 |
return _cycle_size; |
411 | 411 |
} |
412 | 412 |
|
413 |
/// \brief Return the mean |
|
413 |
/// \brief Return the mean cost of the found cycle. |
|
414 | 414 |
/// |
415 |
/// This function returns the mean |
|
415 |
/// This function returns the mean cost of the found cycle. |
|
416 | 416 |
/// |
417 | 417 |
/// \note <tt>alg.cycleMean()</tt> is just a shortcut of the |
418 | 418 |
/// following code. |
419 | 419 |
/// \code |
420 |
/// return static_cast<double>(alg. |
|
420 |
/// return static_cast<double>(alg.cycleCost()) / alg.cycleSize(); |
|
421 | 421 |
/// \endcode |
422 | 422 |
/// |
423 |
/// \pre \ref run() or \ref |
|
423 |
/// \pre \ref run() or \ref findCycleMean() must be called before |
|
424 | 424 |
/// using this function. |
425 | 425 |
double cycleMean() const { |
426 |
return static_cast<double>( |
|
426 |
return static_cast<double>(_cycle_cost) / _cycle_size; |
|
427 | 427 |
} |
428 | 428 |
|
429 | 429 |
/// \brief Return the found cycle. |
430 | 430 |
/// |
431 | 431 |
/// This function returns a const reference to the path structure |
432 | 432 |
/// storing the found cycle. |
433 | 433 |
/// |
434 | 434 |
/// \pre \ref run() or \ref findCycle() must be called before using |
435 | 435 |
/// this function. |
436 | 436 |
const Path& cycle() const { |
437 | 437 |
return *_cycle_path; |
438 | 438 |
} |
439 | 439 |
|
440 | 440 |
///@} |
441 | 441 |
|
442 | 442 |
private: |
443 | 443 |
|
444 | 444 |
// Initialization |
445 | 445 |
void init() { |
446 | 446 |
if (!_cycle_path) { |
447 | 447 |
_local_path = true; |
448 | 448 |
_cycle_path = new Path; |
449 | 449 |
} |
450 | 450 |
_cycle_path->clear(); |
451 |
|
|
451 |
_cycle_cost = 0; |
|
452 | 452 |
_cycle_size = 1; |
453 | 453 |
_cycle_node = INVALID; |
454 | 454 |
for (NodeIt u(_gr); u != INVALID; ++u) |
455 | 455 |
_data[u].clear(); |
456 | 456 |
} |
457 | 457 |
|
458 | 458 |
// Find strongly connected components and initialize _comp_nodes |
459 | 459 |
// and _out_arcs |
460 | 460 |
void findComponents() { |
461 | 461 |
_comp_num = stronglyConnectedComponents(_gr, _comp); |
462 | 462 |
_comp_nodes.resize(_comp_num); |
463 | 463 |
if (_comp_num == 1) { |
464 | 464 |
_comp_nodes[0].clear(); |
465 | 465 |
for (NodeIt n(_gr); n != INVALID; ++n) { |
466 | 466 |
_comp_nodes[0].push_back(n); |
467 | 467 |
_out_arcs[n].clear(); |
468 | 468 |
for (OutArcIt a(_gr, n); a != INVALID; ++a) { |
469 | 469 |
_out_arcs[n].push_back(a); |
470 | 470 |
} |
471 | 471 |
} |
472 | 472 |
} else { |
473 | 473 |
for (int i = 0; i < _comp_num; ++i) |
474 | 474 |
_comp_nodes[i].clear(); |
475 | 475 |
for (NodeIt n(_gr); n != INVALID; ++n) { |
476 | 476 |
int k = _comp[n]; |
477 | 477 |
_comp_nodes[k].push_back(n); |
478 | 478 |
_out_arcs[n].clear(); |
479 | 479 |
for (OutArcIt a(_gr, n); a != INVALID; ++a) { |
480 | 480 |
if (_comp[_gr.target(a)] == k) _out_arcs[n].push_back(a); |
481 | 481 |
} |
482 | 482 |
} |
483 | 483 |
} |
484 | 484 |
} |
485 | 485 |
|
486 | 486 |
// Initialize path data for the current component |
487 | 487 |
bool initComponent(int comp) { |
488 | 488 |
_nodes = &(_comp_nodes[comp]); |
489 | 489 |
int n = _nodes->size(); |
490 | 490 |
if (n < 1 || (n == 1 && _out_arcs[(*_nodes)[0]].size() == 0)) { |
491 | 491 |
return false; |
492 | 492 |
} |
493 | 493 |
for (int i = 0; i < n; ++i) { |
494 | 494 |
_data[(*_nodes)[i]].resize(n + 1, PathData(INF)); |
495 | 495 |
} |
496 | 496 |
return true; |
497 | 497 |
} |
498 | 498 |
|
499 | 499 |
// Process all rounds of computing path data for the current component. |
500 |
// _data[v][k] is the |
|
500 |
// _data[v][k] is the cost of a shortest directed walk from the root |
|
501 | 501 |
// node to node v containing exactly k arcs. |
502 | 502 |
void processRounds() { |
503 | 503 |
Node start = (*_nodes)[0]; |
504 | 504 |
_data[start][0] = PathData(0); |
505 | 505 |
_process.clear(); |
506 | 506 |
_process.push_back(start); |
507 | 507 |
|
508 | 508 |
int k, n = _nodes->size(); |
509 | 509 |
for (k = 1; k <= n && int(_process.size()) < n; ++k) { |
510 | 510 |
processNextBuildRound(k); |
511 | 511 |
} |
512 | 512 |
for ( ; k <= n; ++k) { |
513 | 513 |
processNextFullRound(k); |
514 | 514 |
} |
515 | 515 |
} |
516 | 516 |
|
517 | 517 |
// Process one round and rebuild _process |
518 | 518 |
void processNextBuildRound(int k) { |
519 | 519 |
std::vector<Node> next; |
520 | 520 |
Node u, v; |
521 | 521 |
Arc e; |
522 |
|
|
522 |
LargeCost d; |
|
523 | 523 |
for (int i = 0; i < int(_process.size()); ++i) { |
524 | 524 |
u = _process[i]; |
525 | 525 |
for (int j = 0; j < int(_out_arcs[u].size()); ++j) { |
526 | 526 |
e = _out_arcs[u][j]; |
527 | 527 |
v = _gr.target(e); |
528 |
d = _data[u][k-1].dist + |
|
528 |
d = _data[u][k-1].dist + _cost[e]; |
|
529 | 529 |
if (_tolerance.less(d, _data[v][k].dist)) { |
530 | 530 |
if (_data[v][k].dist == INF) next.push_back(v); |
531 | 531 |
_data[v][k] = PathData(d, e); |
532 | 532 |
} |
533 | 533 |
} |
534 | 534 |
} |
535 | 535 |
_process.swap(next); |
536 | 536 |
} |
537 | 537 |
|
538 | 538 |
// Process one round using _nodes instead of _process |
539 | 539 |
void processNextFullRound(int k) { |
540 | 540 |
Node u, v; |
541 | 541 |
Arc e; |
542 |
|
|
542 |
LargeCost d; |
|
543 | 543 |
for (int i = 0; i < int(_nodes->size()); ++i) { |
544 | 544 |
u = (*_nodes)[i]; |
545 | 545 |
for (int j = 0; j < int(_out_arcs[u].size()); ++j) { |
546 | 546 |
e = _out_arcs[u][j]; |
547 | 547 |
v = _gr.target(e); |
548 |
d = _data[u][k-1].dist + |
|
548 |
d = _data[u][k-1].dist + _cost[e]; |
|
549 | 549 |
if (_tolerance.less(d, _data[v][k].dist)) { |
550 | 550 |
_data[v][k] = PathData(d, e); |
551 | 551 |
} |
552 | 552 |
} |
553 | 553 |
} |
554 | 554 |
} |
555 | 555 |
|
556 | 556 |
// Update the minimum cycle mean |
557 | 557 |
void updateMinMean() { |
558 | 558 |
int n = _nodes->size(); |
559 | 559 |
for (int i = 0; i < n; ++i) { |
560 | 560 |
Node u = (*_nodes)[i]; |
561 | 561 |
if (_data[u][n].dist == INF) continue; |
562 |
|
|
562 |
LargeCost cost, max_cost = 0; |
|
563 | 563 |
int size, max_size = 1; |
564 | 564 |
bool found_curr = false; |
565 | 565 |
for (int k = 0; k < n; ++k) { |
566 | 566 |
if (_data[u][k].dist == INF) continue; |
567 |
|
|
567 |
cost = _data[u][n].dist - _data[u][k].dist; |
|
568 | 568 |
size = n - k; |
569 |
if (!found_curr || |
|
569 |
if (!found_curr || cost * max_size > max_cost * size) { |
|
570 | 570 |
found_curr = true; |
571 |
|
|
571 |
max_cost = cost; |
|
572 | 572 |
max_size = size; |
573 | 573 |
} |
574 | 574 |
} |
575 | 575 |
if ( found_curr && (_cycle_node == INVALID || |
576 |
max_length * _cycle_size < _cycle_length * max_size) ) { |
|
577 |
_cycle_length = max_length; |
|
576 |
max_cost * _cycle_size < _cycle_cost * max_size) ) { |
|
577 |
_cycle_cost = max_cost; |
|
578 | 578 |
_cycle_size = max_size; |
579 | 579 |
_cycle_node = u; |
580 | 580 |
} |
581 | 581 |
} |
582 | 582 |
} |
583 | 583 |
|
584 |
}; //class |
|
584 |
}; //class KarpMmc |
|
585 | 585 |
|
586 | 586 |
///@} |
587 | 587 |
|
588 | 588 |
} //namespace lemon |
589 | 589 |
|
590 |
#endif // |
|
590 |
#endif //LEMON_KARP_MMC_H |
1 | 1 |
/* -*- mode: C++; indent-tabs-mode: nil; -*- |
2 | 2 |
* |
3 | 3 |
* This file is a part of LEMON, a generic C++ optimization library. |
4 | 4 |
* |
5 | 5 |
* Copyright (C) 2003-2009 |
6 | 6 |
* Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport |
7 | 7 |
* (Egervary Research Group on Combinatorial Optimization, EGRES). |
8 | 8 |
* |
9 | 9 |
* Permission to use, modify and distribute this software is granted |
10 | 10 |
* provided that this copyright notice appears in all copies. For |
11 | 11 |
* precise terms see the accompanying LICENSE file. |
12 | 12 |
* |
13 | 13 |
* This software is provided "AS IS" with no warranty of any kind, |
14 | 14 |
* express or implied, and with no claim as to its suitability for any |
15 | 15 |
* purpose. |
16 | 16 |
* |
17 | 17 |
*/ |
18 | 18 |
|
19 | 19 |
#include <iostream> |
20 | 20 |
#include <sstream> |
21 | 21 |
|
22 | 22 |
#include <lemon/smart_graph.h> |
23 | 23 |
#include <lemon/lgf_reader.h> |
24 | 24 |
#include <lemon/path.h> |
25 | 25 |
#include <lemon/concepts/digraph.h> |
26 | 26 |
#include <lemon/concept_check.h> |
27 | 27 |
|
28 |
#include <lemon/karp.h> |
|
29 |
#include <lemon/hartmann_orlin.h> |
|
30 |
#include <lemon/ |
|
28 |
#include <lemon/karp_mmc.h> |
|
29 |
#include <lemon/hartmann_orlin_mmc.h> |
|
30 |
#include <lemon/howard_mmc.h> |
|
31 | 31 |
|
32 | 32 |
#include "test_tools.h" |
33 | 33 |
|
34 | 34 |
using namespace lemon; |
35 | 35 |
|
36 | 36 |
char test_lgf[] = |
37 | 37 |
"@nodes\n" |
38 | 38 |
"label\n" |
39 | 39 |
"1\n" |
40 | 40 |
"2\n" |
41 | 41 |
"3\n" |
42 | 42 |
"4\n" |
43 | 43 |
"5\n" |
44 | 44 |
"6\n" |
45 | 45 |
"7\n" |
46 | 46 |
"@arcs\n" |
47 | 47 |
" len1 len2 len3 len4 c1 c2 c3 c4\n" |
48 | 48 |
"1 2 1 1 1 1 0 0 0 0\n" |
49 | 49 |
"2 4 5 5 5 5 1 0 0 0\n" |
50 | 50 |
"2 3 8 8 8 8 0 0 0 0\n" |
51 | 51 |
"3 2 -2 0 0 0 1 0 0 0\n" |
52 | 52 |
"3 4 4 4 4 4 0 0 0 0\n" |
53 | 53 |
"3 7 -4 -4 -4 -4 0 0 0 0\n" |
54 | 54 |
"4 1 2 2 2 2 0 0 0 0\n" |
55 | 55 |
"4 3 3 3 3 3 1 0 0 0\n" |
56 | 56 |
"4 4 3 3 0 0 0 0 1 0\n" |
57 | 57 |
"5 2 4 4 4 4 0 0 0 0\n" |
58 | 58 |
"5 6 3 3 3 3 0 1 0 0\n" |
59 | 59 |
"6 5 2 2 2 2 0 1 0 0\n" |
60 | 60 |
"6 4 -1 -1 -1 -1 0 0 0 0\n" |
61 | 61 |
"6 7 1 1 1 1 0 0 0 0\n" |
62 | 62 |
"7 7 4 4 4 -1 0 0 0 1\n"; |
63 | 63 |
|
64 | 64 |
|
65 | 65 |
// Check the interface of an MMC algorithm |
66 |
template <typename GR, typename |
|
66 |
template <typename GR, typename Cost> |
|
67 | 67 |
struct MmcClassConcept |
68 | 68 |
{ |
69 | 69 |
template <typename MMC> |
70 | 70 |
struct Constraints { |
71 | 71 |
void constraints() { |
72 | 72 |
const Constraints& me = *this; |
73 | 73 |
|
74 | 74 |
typedef typename MMC |
75 | 75 |
::template SetPath<ListPath<GR> > |
76 |
::template |
|
76 |
::template SetLargeCost<Cost> |
|
77 | 77 |
::Create MmcAlg; |
78 |
MmcAlg mmc(me.g, me. |
|
78 |
MmcAlg mmc(me.g, me.cost); |
|
79 | 79 |
const MmcAlg& const_mmc = mmc; |
80 | 80 |
|
81 | 81 |
typename MmcAlg::Tolerance tol = const_mmc.tolerance(); |
82 | 82 |
mmc.tolerance(tol); |
83 | 83 |
|
84 | 84 |
b = mmc.cycle(p).run(); |
85 |
b = mmc. |
|
85 |
b = mmc.findCycleMean(); |
|
86 | 86 |
b = mmc.findCycle(); |
87 | 87 |
|
88 |
v = const_mmc.cycleLength(); |
|
89 |
i = const_mmc.cycleArcNum(); |
|
88 |
v = const_mmc.cycleCost(); |
|
89 |
i = const_mmc.cycleSize(); |
|
90 | 90 |
d = const_mmc.cycleMean(); |
91 | 91 |
p = const_mmc.cycle(); |
92 | 92 |
} |
93 | 93 |
|
94 |
typedef concepts::ReadMap<typename GR::Arc, |
|
94 |
typedef concepts::ReadMap<typename GR::Arc, Cost> CM; |
|
95 | 95 |
|
96 | 96 |
GR g; |
97 |
|
|
97 |
CM cost; |
|
98 | 98 |
ListPath<GR> p; |
99 |
|
|
99 |
Cost v; |
|
100 | 100 |
int i; |
101 | 101 |
double d; |
102 | 102 |
bool b; |
103 | 103 |
}; |
104 | 104 |
}; |
105 | 105 |
|
106 | 106 |
// Perform a test with the given parameters |
107 | 107 |
template <typename MMC> |
108 | 108 |
void checkMmcAlg(const SmartDigraph& gr, |
109 | 109 |
const SmartDigraph::ArcMap<int>& lm, |
110 | 110 |
const SmartDigraph::ArcMap<int>& cm, |
111 |
int |
|
111 |
int cost, int size) { |
|
112 | 112 |
MMC alg(gr, lm); |
113 |
alg.findMinMean(); |
|
114 |
check(alg.cycleMean() == static_cast<double>(length) / size, |
|
113 |
alg.findCycleMean(); |
|
114 |
check(alg.cycleMean() == static_cast<double>(cost) / size, |
|
115 | 115 |
"Wrong cycle mean"); |
116 | 116 |
alg.findCycle(); |
117 |
check(alg. |
|
117 |
check(alg.cycleCost() == cost && alg.cycleSize() == size, |
|
118 | 118 |
"Wrong path"); |
119 | 119 |
SmartDigraph::ArcMap<int> cycle(gr, 0); |
120 | 120 |
for (typename MMC::Path::ArcIt a(alg.cycle()); a != INVALID; ++a) { |
121 | 121 |
++cycle[a]; |
122 | 122 |
} |
123 | 123 |
for (SmartDigraph::ArcIt a(gr); a != INVALID; ++a) { |
124 | 124 |
check(cm[a] == cycle[a], "Wrong path"); |
125 | 125 |
} |
126 | 126 |
} |
127 | 127 |
|
128 | 128 |
// Class for comparing types |
129 | 129 |
template <typename T1, typename T2> |
130 | 130 |
struct IsSameType { |
131 | 131 |
static const int result = 0; |
132 | 132 |
}; |
133 | 133 |
|
134 | 134 |
template <typename T> |
135 | 135 |
struct IsSameType<T,T> { |
136 | 136 |
static const int result = 1; |
137 | 137 |
}; |
138 | 138 |
|
139 | 139 |
|
140 | 140 |
int main() { |
141 | 141 |
#ifdef LEMON_HAVE_LONG_LONG |
142 | 142 |
typedef long long long_int; |
143 | 143 |
#else |
144 | 144 |
typedef long long_int; |
145 | 145 |
#endif |
146 | 146 |
|
147 | 147 |
// Check the interface |
148 | 148 |
{ |
149 | 149 |
typedef concepts::Digraph GR; |
150 | 150 |
|
151 |
// |
|
151 |
// KarpMmc |
|
152 | 152 |
checkConcept< MmcClassConcept<GR, int>, |
153 |
|
|
153 |
KarpMmc<GR, concepts::ReadMap<GR::Arc, int> > >(); |
|
154 | 154 |
checkConcept< MmcClassConcept<GR, float>, |
155 |
|
|
155 |
KarpMmc<GR, concepts::ReadMap<GR::Arc, float> > >(); |
|
156 | 156 |
|
157 |
// |
|
157 |
// HartmannOrlinMmc |
|
158 | 158 |
checkConcept< MmcClassConcept<GR, int>, |
159 |
|
|
159 |
HartmannOrlinMmc<GR, concepts::ReadMap<GR::Arc, int> > >(); |
|
160 | 160 |
checkConcept< MmcClassConcept<GR, float>, |
161 |
|
|
161 |
HartmannOrlinMmc<GR, concepts::ReadMap<GR::Arc, float> > >(); |
|
162 | 162 |
|
163 |
// |
|
163 |
// HowardMmc |
|
164 | 164 |
checkConcept< MmcClassConcept<GR, int>, |
165 |
|
|
165 |
HowardMmc<GR, concepts::ReadMap<GR::Arc, int> > >(); |
|
166 | 166 |
checkConcept< MmcClassConcept<GR, float>, |
167 |
|
|
167 |
HowardMmc<GR, concepts::ReadMap<GR::Arc, float> > >(); |
|
168 | 168 |
|
169 |
if (IsSameType<Howard<GR, concepts::ReadMap<GR::Arc, int> >::LargeValue, |
|
170 |
long_int>::result == 0) check(false, "Wrong LargeValue type"); |
|
171 |
if (IsSameType<Howard<GR, concepts::ReadMap<GR::Arc, float> >::LargeValue, |
|
172 |
double>::result == 0) check(false, "Wrong LargeValue type"); |
|
169 |
check((IsSameType<HowardMmc<GR, concepts::ReadMap<GR::Arc, int> > |
|
170 |
::LargeCost, long_int>::result == 1), "Wrong LargeCost type"); |
|
171 |
check((IsSameType<HowardMmc<GR, concepts::ReadMap<GR::Arc, float> > |
|
172 |
::LargeCost, double>::result == 1), "Wrong LargeCost type"); |
|
173 | 173 |
} |
174 | 174 |
|
175 | 175 |
// Run various tests |
176 | 176 |
{ |
177 | 177 |
typedef SmartDigraph GR; |
178 | 178 |
DIGRAPH_TYPEDEFS(GR); |
179 | 179 |
|
180 | 180 |
GR gr; |
181 | 181 |
IntArcMap l1(gr), l2(gr), l3(gr), l4(gr); |
182 | 182 |
IntArcMap c1(gr), c2(gr), c3(gr), c4(gr); |
183 | 183 |
|
184 | 184 |
std::istringstream input(test_lgf); |
185 | 185 |
digraphReader(gr, input). |
186 | 186 |
arcMap("len1", l1). |
187 | 187 |
arcMap("len2", l2). |
188 | 188 |
arcMap("len3", l3). |
189 | 189 |
arcMap("len4", l4). |
190 | 190 |
arcMap("c1", c1). |
191 | 191 |
arcMap("c2", c2). |
192 | 192 |
arcMap("c3", c3). |
193 | 193 |
arcMap("c4", c4). |
194 | 194 |
run(); |
195 | 195 |
|
196 | 196 |
// Karp |
197 |
checkMmcAlg<Karp<GR, IntArcMap> >(gr, l1, c1, 6, 3); |
|
198 |
checkMmcAlg<Karp<GR, IntArcMap> >(gr, l2, c2, 5, 2); |
|
199 |
checkMmcAlg<Karp<GR, IntArcMap> >(gr, l3, c3, 0, 1); |
|
200 |
checkMmcAlg<Karp<GR, IntArcMap> >(gr, l4, c4, -1, 1); |
|
197 |
checkMmcAlg<KarpMmc<GR, IntArcMap> >(gr, l1, c1, 6, 3); |
|
198 |
checkMmcAlg<KarpMmc<GR, IntArcMap> >(gr, l2, c2, 5, 2); |
|
199 |
checkMmcAlg<KarpMmc<GR, IntArcMap> >(gr, l3, c3, 0, 1); |
|
200 |
checkMmcAlg<KarpMmc<GR, IntArcMap> >(gr, l4, c4, -1, 1); |
|
201 | 201 |
|
202 | 202 |
// HartmannOrlin |
203 |
checkMmcAlg<HartmannOrlin<GR, IntArcMap> >(gr, l1, c1, 6, 3); |
|
204 |
checkMmcAlg<HartmannOrlin<GR, IntArcMap> >(gr, l2, c2, 5, 2); |
|
205 |
checkMmcAlg<HartmannOrlin<GR, IntArcMap> >(gr, l3, c3, 0, 1); |
|
206 |
checkMmcAlg<HartmannOrlin<GR, IntArcMap> >(gr, l4, c4, -1, 1); |
|
203 |
checkMmcAlg<HartmannOrlinMmc<GR, IntArcMap> >(gr, l1, c1, 6, 3); |
|
204 |
checkMmcAlg<HartmannOrlinMmc<GR, IntArcMap> >(gr, l2, c2, 5, 2); |
|
205 |
checkMmcAlg<HartmannOrlinMmc<GR, IntArcMap> >(gr, l3, c3, 0, 1); |
|
206 |
checkMmcAlg<HartmannOrlinMmc<GR, IntArcMap> >(gr, l4, c4, -1, 1); |
|
207 | 207 |
|
208 | 208 |
// Howard |
209 |
checkMmcAlg<Howard<GR, IntArcMap> >(gr, l1, c1, 6, 3); |
|
210 |
checkMmcAlg<Howard<GR, IntArcMap> >(gr, l2, c2, 5, 2); |
|
211 |
checkMmcAlg<Howard<GR, IntArcMap> >(gr, l3, c3, 0, 1); |
|
212 |
checkMmcAlg<Howard<GR, IntArcMap> >(gr, l4, c4, -1, 1); |
|
209 |
checkMmcAlg<HowardMmc<GR, IntArcMap> >(gr, l1, c1, 6, 3); |
|
210 |
checkMmcAlg<HowardMmc<GR, IntArcMap> >(gr, l2, c2, 5, 2); |
|
211 |
checkMmcAlg<HowardMmc<GR, IntArcMap> >(gr, l3, c3, 0, 1); |
|
212 |
checkMmcAlg<HowardMmc<GR, IntArcMap> >(gr, l4, c4, -1, 1); |
|
213 | 213 |
} |
214 | 214 |
|
215 | 215 |
return 0; |
216 | 216 |
} |
0 comments (0 inline)