gravatar
kpeter (Peter Kovacs)
kpeter@inf.elte.hu
Fix the GEQ/LEQ handling in NetworkSimplex + improve doc (#291) - Fix the optimality conditions for the GEQ/LEQ form. - Fix the initialization of the algortihm. It ensures correct solutions and it is much faster for the inequality forms. - Fix the pivot rules to search all the arcs that have to be allowed to get in the basis. - Better block size for the Block Search pivot rule. - Improve documentation of the problem and move it to a separate page.
0 3 1
default
4 files changed with 338 insertions and 188 deletions:
↑ Collapse diff ↑
Ignore white space 6 line context
1
/* -*- mode: C++; indent-tabs-mode: nil; -*-
2
 *
3
 * This file is a part of LEMON, a generic C++ optimization library.
4
 *
5
 * Copyright (C) 2003-2009
6
 * Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
7
 * (Egervary Research Group on Combinatorial Optimization, EGRES).
8
 *
9
 * Permission to use, modify and distribute this software is granted
10
 * provided that this copyright notice appears in all copies. For
11
 * precise terms see the accompanying LICENSE file.
12
 *
13
 * This software is provided "AS IS" with no warranty of any kind,
14
 * express or implied, and with no claim as to its suitability for any
15
 * purpose.
16
 *
17
 */
18

	
19
namespace lemon {
20

	
21
/**
22
\page min_cost_flow Minimum Cost Flow Problem
23

	
24
\section mcf_def Definition (GEQ form)
25

	
26
The \e minimum \e cost \e flow \e problem is to find a feasible flow of
27
minimum total cost from a set of supply nodes to a set of demand nodes
28
in a network with capacity constraints (lower and upper bounds)
29
and arc costs.
30

	
31
Formally, let \f$G=(V,A)\f$ be a digraph, \f$lower: A\rightarrow\mathbf{R}\f$,
32
\f$upper: A\rightarrow\mathbf{R}\cup\{+\infty\}\f$ denote the lower and
33
upper bounds for the flow values on the arcs, for which
34
\f$lower(uv) \leq upper(uv)\f$ must hold for all \f$uv\in A\f$,
35
\f$cost: A\rightarrow\mathbf{R}\f$ denotes the cost per unit flow
36
on the arcs and \f$sup: V\rightarrow\mathbf{R}\f$ denotes the
37
signed supply values of the nodes.
38
If \f$sup(u)>0\f$, then \f$u\f$ is a supply node with \f$sup(u)\f$
39
supply, if \f$sup(u)<0\f$, then \f$u\f$ is a demand node with
40
\f$-sup(u)\f$ demand.
41
A minimum cost flow is an \f$f: A\rightarrow\mathbf{R}\f$ solution
42
of the following optimization problem.
43

	
44
\f[ \min\sum_{uv\in A} f(uv) \cdot cost(uv) \f]
45
\f[ \sum_{uv\in A} f(uv) - \sum_{vu\in A} f(vu) \geq
46
    sup(u) \quad \forall u\in V \f]
47
\f[ lower(uv) \leq f(uv) \leq upper(uv) \quad \forall uv\in A \f]
48

	
49
The sum of the supply values, i.e. \f$\sum_{u\in V} sup(u)\f$ must be
50
zero or negative in order to have a feasible solution (since the sum
51
of the expressions on the left-hand side of the inequalities is zero).
52
It means that the total demand must be greater or equal to the total
53
supply and all the supplies have to be carried out from the supply nodes,
54
but there could be demands that are not satisfied.
55
If \f$\sum_{u\in V} sup(u)\f$ is zero, then all the supply/demand
56
constraints have to be satisfied with equality, i.e. all demands
57
have to be satisfied and all supplies have to be used.
58

	
59

	
60
\section mcf_algs Algorithms
61

	
62
LEMON contains several algorithms for solving this problem, for more
63
information see \ref min_cost_flow_algs "Minimum Cost Flow Algorithms".
64

	
65
A feasible solution for this problem can be found using \ref Circulation.
66

	
67

	
68
\section mcf_dual Dual Solution
69

	
70
The dual solution of the minimum cost flow problem is represented by
71
node potentials \f$\pi: V\rightarrow\mathbf{R}\f$.
72
An \f$f: A\rightarrow\mathbf{R}\f$ primal feasible solution is optimal
73
if and only if for some \f$\pi: V\rightarrow\mathbf{R}\f$ node potentials
74
the following \e complementary \e slackness optimality conditions hold.
75

	
76
 - For all \f$uv\in A\f$ arcs:
77
   - if \f$cost^\pi(uv)>0\f$, then \f$f(uv)=lower(uv)\f$;
78
   - if \f$lower(uv)<f(uv)<upper(uv)\f$, then \f$cost^\pi(uv)=0\f$;
79
   - if \f$cost^\pi(uv)<0\f$, then \f$f(uv)=upper(uv)\f$.
80
 - For all \f$u\in V\f$ nodes:
81
   - \f$\pi(u)<=0\f$;
82
   - if \f$\sum_{uv\in A} f(uv) - \sum_{vu\in A} f(vu) \neq sup(u)\f$,
83
     then \f$\pi(u)=0\f$.
84
 
85
Here \f$cost^\pi(uv)\f$ denotes the \e reduced \e cost of the arc
86
\f$uv\in A\f$ with respect to the potential function \f$\pi\f$, i.e.
87
\f[ cost^\pi(uv) = cost(uv) + \pi(u) - \pi(v).\f]
88

	
89
All algorithms provide dual solution (node potentials), as well,
90
if an optimal flow is found.
91

	
92

	
93
\section mcf_eq Equality Form
94

	
95
The above \ref mcf_def "definition" is actually more general than the
96
usual formulation of the minimum cost flow problem, in which strict
97
equalities are required in the supply/demand contraints.
98

	
99
\f[ \min\sum_{uv\in A} f(uv) \cdot cost(uv) \f]
100
\f[ \sum_{uv\in A} f(uv) - \sum_{vu\in A} f(vu) =
101
    sup(u) \quad \forall u\in V \f]
102
\f[ lower(uv) \leq f(uv) \leq upper(uv) \quad \forall uv\in A \f]
103

	
104
However if the sum of the supply values is zero, then these two problems
105
are equivalent.
106
The \ref min_cost_flow_algs "algorithms" in LEMON support the general
107
form, so if you need the equality form, you have to ensure this additional
108
contraint manually.
109

	
110

	
111
\section mcf_leq Opposite Inequalites (LEQ Form)
112

	
113
Another possible definition of the minimum cost flow problem is
114
when there are <em>"less or equal"</em> (LEQ) supply/demand constraints,
115
instead of the <em>"greater or equal"</em> (GEQ) constraints.
116

	
117
\f[ \min\sum_{uv\in A} f(uv) \cdot cost(uv) \f]
118
\f[ \sum_{uv\in A} f(uv) - \sum_{vu\in A} f(vu) \leq
119
    sup(u) \quad \forall u\in V \f]
120
\f[ lower(uv) \leq f(uv) \leq upper(uv) \quad \forall uv\in A \f]
121

	
122
It means that the total demand must be less or equal to the 
123
total supply (i.e. \f$\sum_{u\in V} sup(u)\f$ must be zero or
124
positive) and all the demands have to be satisfied, but there
125
could be supplies that are not carried out from the supply
126
nodes.
127
The equality form is also a special case of this form, of course.
128

	
129
You could easily transform this case to the \ref mcf_def "GEQ form"
130
of the problem by reversing the direction of the arcs and taking the
131
negative of the supply values (e.g. using \ref ReverseDigraph and
132
\ref NegMap adaptors).
133
However \ref NetworkSimplex algorithm also supports this form directly
134
for the sake of convenience.
135

	
136
Note that the optimality conditions for this supply constraint type are
137
slightly differ from the conditions that are discussed for the GEQ form,
138
namely the potentials have to be non-negative instead of non-positive.
139
An \f$f: A\rightarrow\mathbf{R}\f$ feasible solution of this problem
140
is optimal if and only if for some \f$\pi: V\rightarrow\mathbf{R}\f$
141
node potentials the following conditions hold.
142

	
143
 - For all \f$uv\in A\f$ arcs:
144
   - if \f$cost^\pi(uv)>0\f$, then \f$f(uv)=lower(uv)\f$;
145
   - if \f$lower(uv)<f(uv)<upper(uv)\f$, then \f$cost^\pi(uv)=0\f$;
146
   - if \f$cost^\pi(uv)<0\f$, then \f$f(uv)=upper(uv)\f$.
147
 - For all \f$u\in V\f$ nodes:
148
   - \f$\pi(u)>=0\f$;
149
   - if \f$\sum_{uv\in A} f(uv) - \sum_{vu\in A} f(vu) \neq sup(u)\f$,
150
     then \f$\pi(u)=0\f$.
151

	
152
*/
153
}
Ignore white space 6 line context
1 1
EXTRA_DIST += \
2 2
	doc/Doxyfile.in \
3 3
	doc/DoxygenLayout.xml \
4 4
	doc/coding_style.dox \
5 5
	doc/dirs.dox \
6 6
	doc/groups.dox \
7 7
	doc/lgf.dox \
8 8
	doc/license.dox \
9 9
	doc/mainpage.dox \
10 10
	doc/migration.dox \
11
	doc/min_cost_flow.dox \
11 12
	doc/named-param.dox \
12 13
	doc/namespaces.dox \
13 14
	doc/html \
14 15
	doc/CMakeLists.txt
15 16

	
16 17
DOC_EPS_IMAGES18 = \
17 18
	grid_graph.eps \
18 19
	nodeshape_0.eps \
19 20
	nodeshape_1.eps \
20 21
	nodeshape_2.eps \
21 22
	nodeshape_3.eps \
22 23
	nodeshape_4.eps
23 24

	
24 25
DOC_EPS_IMAGES27 = \
25 26
	bipartite_matching.eps \
26 27
	bipartite_partitions.eps \
27 28
	connected_components.eps \
28 29
	edge_biconnected_components.eps \
29 30
	node_biconnected_components.eps \
30 31
	strongly_connected_components.eps
31 32

	
32 33
DOC_EPS_IMAGES = \
33 34
	$(DOC_EPS_IMAGES18) \
34 35
	$(DOC_EPS_IMAGES27)
35 36

	
36 37
DOC_PNG_IMAGES = \
37 38
	$(DOC_EPS_IMAGES:%.eps=doc/gen-images/%.png)
38 39

	
39 40
EXTRA_DIST += $(DOC_EPS_IMAGES:%=doc/images/%)
40 41

	
41 42
doc/html:
42 43
	$(MAKE) $(AM_MAKEFLAGS) html
43 44

	
44 45
GS_COMMAND=gs -dNOPAUSE -dBATCH -q -dEPSCrop -dTextAlphaBits=4 -dGraphicsAlphaBits=4
45 46

	
46 47
$(DOC_EPS_IMAGES18:%.eps=doc/gen-images/%.png): doc/gen-images/%.png: doc/images/%.eps
47 48
	-mkdir doc/gen-images
48 49
	if test ${gs_found} = yes; then \
49 50
	  $(GS_COMMAND) -sDEVICE=pngalpha -r18 -sOutputFile=$@ $<; \
50 51
	else \
51 52
	  echo; \
52 53
	  echo "Ghostscript not found."; \
53 54
	  echo; \
54 55
	  exit 1; \
55 56
	fi
56 57

	
57 58
$(DOC_EPS_IMAGES27:%.eps=doc/gen-images/%.png): doc/gen-images/%.png: doc/images/%.eps
58 59
	-mkdir doc/gen-images
59 60
	if test ${gs_found} = yes; then \
60 61
	  $(GS_COMMAND) -sDEVICE=pngalpha -r27 -sOutputFile=$@ $<; \
61 62
	else \
62 63
	  echo; \
63 64
	  echo "Ghostscript not found."; \
64 65
	  echo; \
65 66
	  exit 1; \
66 67
	fi
67 68

	
68 69
html-local: $(DOC_PNG_IMAGES)
69 70
	if test ${doxygen_found} = yes; then \
70 71
	  cd doc; \
71 72
	  doxygen Doxyfile; \
72 73
	  cd ..; \
73 74
	else \
74 75
	  echo; \
75 76
	  echo "Doxygen not found."; \
76 77
	  echo; \
77 78
	  exit 1; \
78 79
	fi
79 80

	
80 81
clean-local:
81 82
	-rm -rf doc/html
82 83
	-rm -f doc/doxygen.log
83 84
	-rm -f $(DOC_PNG_IMAGES)
84 85
	-rm -rf doc/gen-images
85 86

	
86 87
update-external-tags:
87 88
	wget -O doc/libstdc++.tag.tmp http://gcc.gnu.org/onlinedocs/libstdc++/latest-doxygen/libstdc++.tag && \
88 89
	mv doc/libstdc++.tag.tmp doc/libstdc++.tag || \
89 90
	rm doc/libstdc++.tag.tmp
90 91

	
91 92
install-html-local: doc/html
92 93
	@$(NORMAL_INSTALL)
93 94
	$(mkinstalldirs) $(DESTDIR)$(htmldir)/docs
94 95
	for p in doc/html/*.{html,css,png,map,gif,tag} ; do \
95 96
	  f="`echo $$p | sed -e 's|^.*/||'`"; \
96 97
	  echo " $(INSTALL_DATA) $$p $(DESTDIR)$(htmldir)/docs/$$f"; \
97 98
	  $(INSTALL_DATA) $$p $(DESTDIR)$(htmldir)/docs/$$f; \
98 99
	done
99 100

	
100 101
uninstall-local:
101 102
	@$(NORMAL_UNINSTALL)
102 103
	for p in doc/html/*.{html,css,png,map,gif,tag} ; do \
103 104
	  f="`echo $$p | sed -e 's|^.*/||'`"; \
104 105
	  echo " rm -f $(DESTDIR)$(htmldir)/docs/$$f"; \
105 106
	  rm -f $(DESTDIR)$(htmldir)/docs/$$f; \
106 107
	done
107 108

	
108 109
.PHONY: update-external-tags
Ignore white space 8192 line context
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
namespace lemon {
20 20

	
21 21
/**
22 22
@defgroup datas Data Structures
23 23
This group contains the several data structures implemented in LEMON.
24 24
*/
25 25

	
26 26
/**
27 27
@defgroup graphs Graph Structures
28 28
@ingroup datas
29 29
\brief Graph structures implemented in LEMON.
30 30

	
31 31
The implementation of combinatorial algorithms heavily relies on
32 32
efficient graph implementations. LEMON offers data structures which are
33 33
planned to be easily used in an experimental phase of implementation studies,
34 34
and thereafter the program code can be made efficient by small modifications.
35 35

	
36 36
The most efficient implementation of diverse applications require the
37 37
usage of different physical graph implementations. These differences
38 38
appear in the size of graph we require to handle, memory or time usage
39 39
limitations or in the set of operations through which the graph can be
40 40
accessed.  LEMON provides several physical graph structures to meet
41 41
the diverging requirements of the possible users.  In order to save on
42 42
running time or on memory usage, some structures may fail to provide
43 43
some graph features like arc/edge or node deletion.
44 44

	
45 45
Alteration of standard containers need a very limited number of
46 46
operations, these together satisfy the everyday requirements.
47 47
In the case of graph structures, different operations are needed which do
48 48
not alter the physical graph, but gives another view. If some nodes or
49 49
arcs have to be hidden or the reverse oriented graph have to be used, then
50 50
this is the case. It also may happen that in a flow implementation
51 51
the residual graph can be accessed by another algorithm, or a node-set
52 52
is to be shrunk for another algorithm.
53 53
LEMON also provides a variety of graphs for these requirements called
54 54
\ref graph_adaptors "graph adaptors". Adaptors cannot be used alone but only
55 55
in conjunction with other graph representations.
56 56

	
57 57
You are free to use the graph structure that fit your requirements
58 58
the best, most graph algorithms and auxiliary data structures can be used
59 59
with any graph structure.
60 60

	
61 61
<b>See also:</b> \ref graph_concepts "Graph Structure Concepts".
62 62
*/
63 63

	
64 64
/**
65 65
@defgroup graph_adaptors Adaptor Classes for Graphs
66 66
@ingroup graphs
67 67
\brief Adaptor classes for digraphs and graphs
68 68

	
69 69
This group contains several useful adaptor classes for digraphs and graphs.
70 70

	
71 71
The main parts of LEMON are the different graph structures, generic
72 72
graph algorithms, graph concepts, which couple them, and graph
73 73
adaptors. While the previous notions are more or less clear, the
74 74
latter one needs further explanation. Graph adaptors are graph classes
75 75
which serve for considering graph structures in different ways.
76 76

	
77 77
A short example makes this much clearer.  Suppose that we have an
78 78
instance \c g of a directed graph type, say ListDigraph and an algorithm
79 79
\code
80 80
template <typename Digraph>
81 81
int algorithm(const Digraph&);
82 82
\endcode
83 83
is needed to run on the reverse oriented graph.  It may be expensive
84 84
(in time or in memory usage) to copy \c g with the reversed
85 85
arcs.  In this case, an adaptor class is used, which (according
86 86
to LEMON \ref concepts::Digraph "digraph concepts") works as a digraph.
87 87
The adaptor uses the original digraph structure and digraph operations when
88 88
methods of the reversed oriented graph are called.  This means that the adaptor
89 89
have minor memory usage, and do not perform sophisticated algorithmic
90 90
actions.  The purpose of it is to give a tool for the cases when a
91 91
graph have to be used in a specific alteration.  If this alteration is
92 92
obtained by a usual construction like filtering the node or the arc set or
93 93
considering a new orientation, then an adaptor is worthwhile to use.
94 94
To come back to the reverse oriented graph, in this situation
95 95
\code
96 96
template<typename Digraph> class ReverseDigraph;
97 97
\endcode
98 98
template class can be used. The code looks as follows
99 99
\code
100 100
ListDigraph g;
101 101
ReverseDigraph<ListDigraph> rg(g);
102 102
int result = algorithm(rg);
103 103
\endcode
104 104
During running the algorithm, the original digraph \c g is untouched.
105 105
This techniques give rise to an elegant code, and based on stable
106 106
graph adaptors, complex algorithms can be implemented easily.
107 107

	
108 108
In flow, circulation and matching problems, the residual
109 109
graph is of particular importance. Combining an adaptor implementing
110 110
this with shortest path algorithms or minimum mean cycle algorithms,
111 111
a range of weighted and cardinality optimization algorithms can be
112 112
obtained. For other examples, the interested user is referred to the
113 113
detailed documentation of particular adaptors.
114 114

	
115 115
The behavior of graph adaptors can be very different. Some of them keep
116 116
capabilities of the original graph while in other cases this would be
117 117
meaningless. This means that the concepts that they meet depend
118 118
on the graph adaptor, and the wrapped graph.
119 119
For example, if an arc of a reversed digraph is deleted, this is carried
120 120
out by deleting the corresponding arc of the original digraph, thus the
121 121
adaptor modifies the original digraph.
122 122
However in case of a residual digraph, this operation has no sense.
123 123

	
124 124
Let us stand one more example here to simplify your work.
125 125
ReverseDigraph has constructor
126 126
\code
127 127
ReverseDigraph(Digraph& digraph);
128 128
\endcode
129 129
This means that in a situation, when a <tt>const %ListDigraph&</tt>
130 130
reference to a graph is given, then it have to be instantiated with
131 131
<tt>Digraph=const %ListDigraph</tt>.
132 132
\code
133 133
int algorithm1(const ListDigraph& g) {
134 134
  ReverseDigraph<const ListDigraph> rg(g);
135 135
  return algorithm2(rg);
136 136
}
137 137
\endcode
138 138
*/
139 139

	
140 140
/**
141 141
@defgroup maps Maps
142 142
@ingroup datas
143 143
\brief Map structures implemented in LEMON.
144 144

	
145 145
This group contains the map structures implemented in LEMON.
146 146

	
147 147
LEMON provides several special purpose maps and map adaptors that e.g. combine
148 148
new maps from existing ones.
149 149

	
150 150
<b>See also:</b> \ref map_concepts "Map Concepts".
151 151
*/
152 152

	
153 153
/**
154 154
@defgroup graph_maps Graph Maps
155 155
@ingroup maps
156 156
\brief Special graph-related maps.
157 157

	
158 158
This group contains maps that are specifically designed to assign
159 159
values to the nodes and arcs/edges of graphs.
160 160

	
161 161
If you are looking for the standard graph maps (\c NodeMap, \c ArcMap,
162 162
\c EdgeMap), see the \ref graph_concepts "Graph Structure Concepts".
163 163
*/
164 164

	
165 165
/**
166 166
\defgroup map_adaptors Map Adaptors
167 167
\ingroup maps
168 168
\brief Tools to create new maps from existing ones
169 169

	
170 170
This group contains map adaptors that are used to create "implicit"
171 171
maps from other maps.
172 172

	
173 173
Most of them are \ref concepts::ReadMap "read-only maps".
174 174
They can make arithmetic and logical operations between one or two maps
175 175
(negation, shifting, addition, multiplication, logical 'and', 'or',
176 176
'not' etc.) or e.g. convert a map to another one of different Value type.
177 177

	
178 178
The typical usage of this classes is passing implicit maps to
179 179
algorithms.  If a function type algorithm is called then the function
180 180
type map adaptors can be used comfortable. For example let's see the
181 181
usage of map adaptors with the \c graphToEps() function.
182 182
\code
183 183
  Color nodeColor(int deg) {
184 184
    if (deg >= 2) {
185 185
      return Color(0.5, 0.0, 0.5);
186 186
    } else if (deg == 1) {
187 187
      return Color(1.0, 0.5, 1.0);
188 188
    } else {
189 189
      return Color(0.0, 0.0, 0.0);
190 190
    }
191 191
  }
192 192

	
193 193
  Digraph::NodeMap<int> degree_map(graph);
194 194

	
195 195
  graphToEps(graph, "graph.eps")
196 196
    .coords(coords).scaleToA4().undirected()
197 197
    .nodeColors(composeMap(functorToMap(nodeColor), degree_map))
198 198
    .run();
199 199
\endcode
200 200
The \c functorToMap() function makes an \c int to \c Color map from the
201 201
\c nodeColor() function. The \c composeMap() compose the \c degree_map
202 202
and the previously created map. The composed map is a proper function to
203 203
get the color of each node.
204 204

	
205 205
The usage with class type algorithms is little bit harder. In this
206 206
case the function type map adaptors can not be used, because the
207 207
function map adaptors give back temporary objects.
208 208
\code
209 209
  Digraph graph;
210 210

	
211 211
  typedef Digraph::ArcMap<double> DoubleArcMap;
212 212
  DoubleArcMap length(graph);
213 213
  DoubleArcMap speed(graph);
214 214

	
215 215
  typedef DivMap<DoubleArcMap, DoubleArcMap> TimeMap;
216 216
  TimeMap time(length, speed);
217 217

	
218 218
  Dijkstra<Digraph, TimeMap> dijkstra(graph, time);
219 219
  dijkstra.run(source, target);
220 220
\endcode
221 221
We have a length map and a maximum speed map on the arcs of a digraph.
222 222
The minimum time to pass the arc can be calculated as the division of
223 223
the two maps which can be done implicitly with the \c DivMap template
224 224
class. We use the implicit minimum time map as the length map of the
225 225
\c Dijkstra algorithm.
226 226
*/
227 227

	
228 228
/**
229 229
@defgroup matrices Matrices
230 230
@ingroup datas
231 231
\brief Two dimensional data storages implemented in LEMON.
232 232

	
233 233
This group contains two dimensional data storages implemented in LEMON.
234 234
*/
235 235

	
236 236
/**
237 237
@defgroup paths Path Structures
238 238
@ingroup datas
239 239
\brief %Path structures implemented in LEMON.
240 240

	
241 241
This group contains the path structures implemented in LEMON.
242 242

	
243 243
LEMON provides flexible data structures to work with paths.
244 244
All of them have similar interfaces and they can be copied easily with
245 245
assignment operators and copy constructors. This makes it easy and
246 246
efficient to have e.g. the Dijkstra algorithm to store its result in
247 247
any kind of path structure.
248 248

	
249 249
\sa lemon::concepts::Path
250 250
*/
251 251

	
252 252
/**
253 253
@defgroup auxdat Auxiliary Data Structures
254 254
@ingroup datas
255 255
\brief Auxiliary data structures implemented in LEMON.
256 256

	
257 257
This group contains some data structures implemented in LEMON in
258 258
order to make it easier to implement combinatorial algorithms.
259 259
*/
260 260

	
261 261
/**
262 262
@defgroup algs Algorithms
263 263
\brief This group contains the several algorithms
264 264
implemented in LEMON.
265 265

	
266 266
This group contains the several algorithms
267 267
implemented in LEMON.
268 268
*/
269 269

	
270 270
/**
271 271
@defgroup search Graph Search
272 272
@ingroup algs
273 273
\brief Common graph search algorithms.
274 274

	
275 275
This group contains the common graph search algorithms, namely
276 276
\e breadth-first \e search (BFS) and \e depth-first \e search (DFS).
277 277
*/
278 278

	
279 279
/**
280 280
@defgroup shortest_path Shortest Path Algorithms
281 281
@ingroup algs
282 282
\brief Algorithms for finding shortest paths.
283 283

	
284 284
This group contains the algorithms for finding shortest paths in digraphs.
285 285

	
286 286
 - \ref Dijkstra algorithm for finding shortest paths from a source node
287 287
   when all arc lengths are non-negative.
288 288
 - \ref BellmanFord "Bellman-Ford" algorithm for finding shortest paths
289 289
   from a source node when arc lenghts can be either positive or negative,
290 290
   but the digraph should not contain directed cycles with negative total
291 291
   length.
292 292
 - \ref FloydWarshall "Floyd-Warshall" and \ref Johnson "Johnson" algorithms
293 293
   for solving the \e all-pairs \e shortest \e paths \e problem when arc
294 294
   lenghts can be either positive or negative, but the digraph should
295 295
   not contain directed cycles with negative total length.
296 296
 - \ref Suurballe A successive shortest path algorithm for finding
297 297
   arc-disjoint paths between two nodes having minimum total length.
298 298
*/
299 299

	
300 300
/**
301 301
@defgroup max_flow Maximum Flow Algorithms
302 302
@ingroup algs
303 303
\brief Algorithms for finding maximum flows.
304 304

	
305 305
This group contains the algorithms for finding maximum flows and
306 306
feasible circulations.
307 307

	
308 308
The \e maximum \e flow \e problem is to find a flow of maximum value between
309 309
a single source and a single target. Formally, there is a \f$G=(V,A)\f$
310 310
digraph, a \f$cap: A\rightarrow\mathbf{R}^+_0\f$ capacity function and
311 311
\f$s, t \in V\f$ source and target nodes.
312 312
A maximum flow is an \f$f: A\rightarrow\mathbf{R}^+_0\f$ solution of the
313 313
following optimization problem.
314 314

	
315 315
\f[ \max\sum_{sv\in A} f(sv) - \sum_{vs\in A} f(vs) \f]
316 316
\f[ \sum_{uv\in A} f(uv) = \sum_{vu\in A} f(vu)
317 317
    \quad \forall u\in V\setminus\{s,t\} \f]
318 318
\f[ 0 \leq f(uv) \leq cap(uv) \quad \forall uv\in A \f]
319 319

	
320 320
LEMON contains several algorithms for solving maximum flow problems:
321 321
- \ref EdmondsKarp Edmonds-Karp algorithm.
322 322
- \ref Preflow Goldberg-Tarjan's preflow push-relabel algorithm.
323 323
- \ref DinitzSleatorTarjan Dinitz's blocking flow algorithm with dynamic trees.
324 324
- \ref GoldbergTarjan Preflow push-relabel algorithm with dynamic trees.
325 325

	
326 326
In most cases the \ref Preflow "Preflow" algorithm provides the
327 327
fastest method for computing a maximum flow. All implementations
328 328
also provide functions to query the minimum cut, which is the dual
329 329
problem of maximum flow.
330 330

	
331 331
\ref Circulation is a preflow push-relabel algorithm implemented directly 
332 332
for finding feasible circulations, which is a somewhat different problem,
333 333
but it is strongly related to maximum flow.
334 334
For more information, see \ref Circulation.
335 335
*/
336 336

	
337 337
/**
338
@defgroup min_cost_flow Minimum Cost Flow Algorithms
338
@defgroup min_cost_flow_algs Minimum Cost Flow Algorithms
339 339
@ingroup algs
340 340

	
341 341
\brief Algorithms for finding minimum cost flows and circulations.
342 342

	
343 343
This group contains the algorithms for finding minimum cost flows and
344
circulations.
344
circulations. For more information about this problem and its dual
345
solution see \ref min_cost_flow "Minimum Cost Flow Problem".
345 346

	
346
The \e minimum \e cost \e flow \e problem is to find a feasible flow of
347
minimum total cost from a set of supply nodes to a set of demand nodes
348
in a network with capacity constraints (lower and upper bounds)
349
and arc costs.
350
Formally, let \f$G=(V,A)\f$ be a digraph, \f$lower: A\rightarrow\mathbf{Z}\f$,
351
\f$upper: A\rightarrow\mathbf{Z}\cup\{+\infty\}\f$ denote the lower and
352
upper bounds for the flow values on the arcs, for which
353
\f$lower(uv) \leq upper(uv)\f$ must hold for all \f$uv\in A\f$,
354
\f$cost: A\rightarrow\mathbf{Z}\f$ denotes the cost per unit flow
355
on the arcs and \f$sup: V\rightarrow\mathbf{Z}\f$ denotes the
356
signed supply values of the nodes.
357
If \f$sup(u)>0\f$, then \f$u\f$ is a supply node with \f$sup(u)\f$
358
supply, if \f$sup(u)<0\f$, then \f$u\f$ is a demand node with
359
\f$-sup(u)\f$ demand.
360
A minimum cost flow is an \f$f: A\rightarrow\mathbf{Z}\f$ solution
361
of the following optimization problem.
362

	
363
\f[ \min\sum_{uv\in A} f(uv) \cdot cost(uv) \f]
364
\f[ \sum_{uv\in A} f(uv) - \sum_{vu\in A} f(vu) \geq
365
    sup(u) \quad \forall u\in V \f]
366
\f[ lower(uv) \leq f(uv) \leq upper(uv) \quad \forall uv\in A \f]
367

	
368
The sum of the supply values, i.e. \f$\sum_{u\in V} sup(u)\f$ must be
369
zero or negative in order to have a feasible solution (since the sum
370
of the expressions on the left-hand side of the inequalities is zero).
371
It means that the total demand must be greater or equal to the total
372
supply and all the supplies have to be carried out from the supply nodes,
373
but there could be demands that are not satisfied.
374
If \f$\sum_{u\in V} sup(u)\f$ is zero, then all the supply/demand
375
constraints have to be satisfied with equality, i.e. all demands
376
have to be satisfied and all supplies have to be used.
377

	
378
If you need the opposite inequalities in the supply/demand constraints
379
(i.e. the total demand is less than the total supply and all the demands
380
have to be satisfied while there could be supplies that are not used),
381
then you could easily transform the problem to the above form by reversing
382
the direction of the arcs and taking the negative of the supply values
383
(e.g. using \ref ReverseDigraph and \ref NegMap adaptors).
384
However \ref NetworkSimplex algorithm also supports this form directly
385
for the sake of convenience.
386

	
387
A feasible solution for this problem can be found using \ref Circulation.
388

	
389
Note that the above formulation is actually more general than the usual
390
definition of the minimum cost flow problem, in which strict equalities
391
are required in the supply/demand contraints, i.e.
392

	
393
\f[ \sum_{uv\in A} f(uv) - \sum_{vu\in A} f(vu) =
394
    sup(u) \quad \forall u\in V. \f]
395

	
396
However if the sum of the supply values is zero, then these two problems
397
are equivalent. So if you need the equality form, you have to ensure this
398
additional contraint for the algorithms.
399

	
400
The dual solution of the minimum cost flow problem is represented by node 
401
potentials \f$\pi: V\rightarrow\mathbf{Z}\f$.
402
An \f$f: A\rightarrow\mathbf{Z}\f$ feasible solution of the problem
403
is optimal if and only if for some \f$\pi: V\rightarrow\mathbf{Z}\f$
404
node potentials the following \e complementary \e slackness optimality
405
conditions hold.
406

	
407
 - For all \f$uv\in A\f$ arcs:
408
   - if \f$cost^\pi(uv)>0\f$, then \f$f(uv)=lower(uv)\f$;
409
   - if \f$lower(uv)<f(uv)<upper(uv)\f$, then \f$cost^\pi(uv)=0\f$;
410
   - if \f$cost^\pi(uv)<0\f$, then \f$f(uv)=upper(uv)\f$.
411
 - For all \f$u\in V\f$ nodes:
412
   - if \f$\sum_{uv\in A} f(uv) - \sum_{vu\in A} f(vu) \neq sup(u)\f$,
413
     then \f$\pi(u)=0\f$.
414
 
415
Here \f$cost^\pi(uv)\f$ denotes the \e reduced \e cost of the arc
416
\f$uv\in A\f$ with respect to the potential function \f$\pi\f$, i.e.
417
\f[ cost^\pi(uv) = cost(uv) + \pi(u) - \pi(v).\f]
418

	
419
All algorithms provide dual solution (node potentials) as well,
420
if an optimal flow is found.
421

	
422
LEMON contains several algorithms for solving minimum cost flow problems.
347
LEMON contains several algorithms for this problem.
423 348
 - \ref NetworkSimplex Primal Network Simplex algorithm with various
424 349
   pivot strategies.
425 350
 - \ref CostScaling Push-Relabel and Augment-Relabel algorithms based on
426 351
   cost scaling.
427 352
 - \ref CapacityScaling Successive Shortest %Path algorithm with optional
428 353
   capacity scaling.
429 354
 - \ref CancelAndTighten The Cancel and Tighten algorithm.
430 355
 - \ref CycleCanceling Cycle-Canceling algorithms.
431 356

	
432
Most of these implementations support the general inequality form of the
433
minimum cost flow problem, but CancelAndTighten and CycleCanceling
434
only support the equality form due to the primal method they use.
435

	
436 357
In general NetworkSimplex is the most efficient implementation,
437 358
but in special cases other algorithms could be faster.
438 359
For example, if the total supply and/or capacities are rather small,
439 360
CapacityScaling is usually the fastest algorithm (without effective scaling).
440 361
*/
441 362

	
442 363
/**
443 364
@defgroup min_cut Minimum Cut Algorithms
444 365
@ingroup algs
445 366

	
446 367
\brief Algorithms for finding minimum cut in graphs.
447 368

	
448 369
This group contains the algorithms for finding minimum cut in graphs.
449 370

	
450 371
The \e minimum \e cut \e problem is to find a non-empty and non-complete
451 372
\f$X\f$ subset of the nodes with minimum overall capacity on
452 373
outgoing arcs. Formally, there is a \f$G=(V,A)\f$ digraph, a
453 374
\f$cap: A\rightarrow\mathbf{R}^+_0\f$ capacity function. The minimum
454 375
cut is the \f$X\f$ solution of the next optimization problem:
455 376

	
456 377
\f[ \min_{X \subset V, X\not\in \{\emptyset, V\}}
457 378
    \sum_{uv\in A, u\in X, v\not\in X}cap(uv) \f]
458 379

	
459 380
LEMON contains several algorithms related to minimum cut problems:
460 381

	
461 382
- \ref HaoOrlin "Hao-Orlin algorithm" for calculating minimum cut
462 383
  in directed graphs.
463 384
- \ref NagamochiIbaraki "Nagamochi-Ibaraki algorithm" for
464 385
  calculating minimum cut in undirected graphs.
465 386
- \ref GomoryHu "Gomory-Hu tree computation" for calculating
466 387
  all-pairs minimum cut in undirected graphs.
467 388

	
468 389
If you want to find minimum cut just between two distinict nodes,
469 390
see the \ref max_flow "maximum flow problem".
470 391
*/
471 392

	
472 393
/**
473 394
@defgroup graph_properties Connectivity and Other Graph Properties
474 395
@ingroup algs
475 396
\brief Algorithms for discovering the graph properties
476 397

	
477 398
This group contains the algorithms for discovering the graph properties
478 399
like connectivity, bipartiteness, euler property, simplicity etc.
479 400

	
480 401
\image html edge_biconnected_components.png
481 402
\image latex edge_biconnected_components.eps "bi-edge-connected components" width=\textwidth
482 403
*/
483 404

	
484 405
/**
485 406
@defgroup planar Planarity Embedding and Drawing
486 407
@ingroup algs
487 408
\brief Algorithms for planarity checking, embedding and drawing
488 409

	
489 410
This group contains the algorithms for planarity checking,
490 411
embedding and drawing.
491 412

	
492 413
\image html planar.png
493 414
\image latex planar.eps "Plane graph" width=\textwidth
494 415
*/
495 416

	
496 417
/**
497 418
@defgroup matching Matching Algorithms
498 419
@ingroup algs
499 420
\brief Algorithms for finding matchings in graphs and bipartite graphs.
500 421

	
501 422
This group contains the algorithms for calculating
502 423
matchings in graphs and bipartite graphs. The general matching problem is
503 424
finding a subset of the edges for which each node has at most one incident
504 425
edge.
505 426

	
506 427
There are several different algorithms for calculate matchings in
507 428
graphs.  The matching problems in bipartite graphs are generally
508 429
easier than in general graphs. The goal of the matching optimization
509 430
can be finding maximum cardinality, maximum weight or minimum cost
510 431
matching. The search can be constrained to find perfect or
511 432
maximum cardinality matching.
512 433

	
513 434
The matching algorithms implemented in LEMON:
514 435
- \ref MaxBipartiteMatching Hopcroft-Karp augmenting path algorithm
515 436
  for calculating maximum cardinality matching in bipartite graphs.
516 437
- \ref PrBipartiteMatching Push-relabel algorithm
517 438
  for calculating maximum cardinality matching in bipartite graphs.
518 439
- \ref MaxWeightedBipartiteMatching
519 440
  Successive shortest path algorithm for calculating maximum weighted
520 441
  matching and maximum weighted bipartite matching in bipartite graphs.
521 442
- \ref MinCostMaxBipartiteMatching
522 443
  Successive shortest path algorithm for calculating minimum cost maximum
523 444
  matching in bipartite graphs.
524 445
- \ref MaxMatching Edmond's blossom shrinking algorithm for calculating
525 446
  maximum cardinality matching in general graphs.
526 447
- \ref MaxWeightedMatching Edmond's blossom shrinking algorithm for calculating
527 448
  maximum weighted matching in general graphs.
528 449
- \ref MaxWeightedPerfectMatching
529 450
  Edmond's blossom shrinking algorithm for calculating maximum weighted
530 451
  perfect matching in general graphs.
531 452

	
532 453
\image html bipartite_matching.png
533 454
\image latex bipartite_matching.eps "Bipartite Matching" width=\textwidth
534 455
*/
535 456

	
536 457
/**
537 458
@defgroup spantree Minimum Spanning Tree Algorithms
538 459
@ingroup algs
539 460
\brief Algorithms for finding minimum cost spanning trees and arborescences.
540 461

	
541 462
This group contains the algorithms for finding minimum cost spanning
542 463
trees and arborescences.
543 464
*/
544 465

	
545 466
/**
546 467
@defgroup auxalg Auxiliary Algorithms
547 468
@ingroup algs
548 469
\brief Auxiliary algorithms implemented in LEMON.
549 470

	
550 471
This group contains some algorithms implemented in LEMON
551 472
in order to make it easier to implement complex algorithms.
552 473
*/
553 474

	
554 475
/**
555 476
@defgroup approx Approximation Algorithms
556 477
@ingroup algs
557 478
\brief Approximation algorithms.
558 479

	
559 480
This group contains the approximation and heuristic algorithms
560 481
implemented in LEMON.
561 482
*/
562 483

	
563 484
/**
564 485
@defgroup gen_opt_group General Optimization Tools
565 486
\brief This group contains some general optimization frameworks
566 487
implemented in LEMON.
567 488

	
568 489
This group contains some general optimization frameworks
569 490
implemented in LEMON.
570 491
*/
571 492

	
572 493
/**
573 494
@defgroup lp_group Lp and Mip Solvers
574 495
@ingroup gen_opt_group
575 496
\brief Lp and Mip solver interfaces for LEMON.
576 497

	
577 498
This group contains Lp and Mip solver interfaces for LEMON. The
578 499
various LP solvers could be used in the same manner with this
579 500
interface.
580 501
*/
581 502

	
582 503
/**
583 504
@defgroup lp_utils Tools for Lp and Mip Solvers
584 505
@ingroup lp_group
585 506
\brief Helper tools to the Lp and Mip solvers.
586 507

	
587 508
This group adds some helper tools to general optimization framework
588 509
implemented in LEMON.
589 510
*/
590 511

	
591 512
/**
592 513
@defgroup metah Metaheuristics
593 514
@ingroup gen_opt_group
594 515
\brief Metaheuristics for LEMON library.
595 516

	
596 517
This group contains some metaheuristic optimization tools.
597 518
*/
598 519

	
599 520
/**
600 521
@defgroup utils Tools and Utilities
601 522
\brief Tools and utilities for programming in LEMON
602 523

	
603 524
Tools and utilities for programming in LEMON.
604 525
*/
605 526

	
606 527
/**
607 528
@defgroup gutils Basic Graph Utilities
608 529
@ingroup utils
609 530
\brief Simple basic graph utilities.
610 531

	
611 532
This group contains some simple basic graph utilities.
612 533
*/
613 534

	
614 535
/**
615 536
@defgroup misc Miscellaneous Tools
616 537
@ingroup utils
617 538
\brief Tools for development, debugging and testing.
618 539

	
619 540
This group contains several useful tools for development,
620 541
debugging and testing.
621 542
*/
622 543

	
623 544
/**
624 545
@defgroup timecount Time Measuring and Counting
625 546
@ingroup misc
626 547
\brief Simple tools for measuring the performance of algorithms.
627 548

	
628 549
This group contains simple tools for measuring the performance
629 550
of algorithms.
630 551
*/
631 552

	
632 553
/**
633 554
@defgroup exceptions Exceptions
634 555
@ingroup utils
635 556
\brief Exceptions defined in LEMON.
636 557

	
637 558
This group contains the exceptions defined in LEMON.
638 559
*/
639 560

	
640 561
/**
641 562
@defgroup io_group Input-Output
642 563
\brief Graph Input-Output methods
643 564

	
644 565
This group contains the tools for importing and exporting graphs
645 566
and graph related data. Now it supports the \ref lgf-format
646 567
"LEMON Graph Format", the \c DIMACS format and the encapsulated
647 568
postscript (EPS) format.
648 569
*/
649 570

	
650 571
/**
651 572
@defgroup lemon_io LEMON Graph Format
652 573
@ingroup io_group
653 574
\brief Reading and writing LEMON Graph Format.
654 575

	
655 576
This group contains methods for reading and writing
656 577
\ref lgf-format "LEMON Graph Format".
657 578
*/
658 579

	
659 580
/**
660 581
@defgroup eps_io Postscript Exporting
661 582
@ingroup io_group
662 583
\brief General \c EPS drawer and graph exporter
663 584

	
664 585
This group contains general \c EPS drawing methods and special
665 586
graph exporting tools.
666 587
*/
667 588

	
668 589
/**
669 590
@defgroup dimacs_group DIMACS format
670 591
@ingroup io_group
671 592
\brief Read and write files in DIMACS format
672 593

	
673 594
Tools to read a digraph from or write it to a file in DIMACS format data.
674 595
*/
675 596

	
676 597
/**
677 598
@defgroup nauty_group NAUTY Format
678 599
@ingroup io_group
679 600
\brief Read \e Nauty format
680 601

	
681 602
Tool to read graphs from \e Nauty format data.
682 603
*/
683 604

	
684 605
/**
685 606
@defgroup concept Concepts
686 607
\brief Skeleton classes and concept checking classes
687 608

	
688 609
This group contains the data/algorithm skeletons and concept checking
689 610
classes implemented in LEMON.
690 611

	
691 612
The purpose of the classes in this group is fourfold.
692 613

	
693 614
- These classes contain the documentations of the %concepts. In order
694 615
  to avoid document multiplications, an implementation of a concept
695 616
  simply refers to the corresponding concept class.
696 617

	
697 618
- These classes declare every functions, <tt>typedef</tt>s etc. an
698 619
  implementation of the %concepts should provide, however completely
699 620
  without implementations and real data structures behind the
700 621
  interface. On the other hand they should provide nothing else. All
701 622
  the algorithms working on a data structure meeting a certain concept
702 623
  should compile with these classes. (Though it will not run properly,
703 624
  of course.) In this way it is easily to check if an algorithm
704 625
  doesn't use any extra feature of a certain implementation.
705 626

	
706 627
- The concept descriptor classes also provide a <em>checker class</em>
707 628
  that makes it possible to check whether a certain implementation of a
708 629
  concept indeed provides all the required features.
709 630

	
710 631
- Finally, They can serve as a skeleton of a new implementation of a concept.
711 632
*/
712 633

	
713 634
/**
714 635
@defgroup graph_concepts Graph Structure Concepts
715 636
@ingroup concept
716 637
\brief Skeleton and concept checking classes for graph structures
717 638

	
718 639
This group contains the skeletons and concept checking classes of LEMON's
719 640
graph structures and helper classes used to implement these.
720 641
*/
721 642

	
722 643
/**
723 644
@defgroup map_concepts Map Concepts
724 645
@ingroup concept
725 646
\brief Skeleton and concept checking classes for maps
726 647

	
727 648
This group contains the skeletons and concept checking classes of maps.
728 649
*/
729 650

	
730 651
/**
731 652
\anchor demoprograms
732 653

	
733 654
@defgroup demos Demo Programs
734 655

	
735 656
Some demo programs are listed here. Their full source codes can be found in
736 657
the \c demo subdirectory of the source tree.
737 658

	
738 659
In order to compile them, use the <tt>make demo</tt> or the
739 660
<tt>make check</tt> commands.
740 661
*/
741 662

	
742 663
/**
743 664
@defgroup tools Standalone Utility Applications
744 665

	
745 666
Some utility applications are listed here.
746 667

	
747 668
The standard compilation procedure (<tt>./configure;make</tt>) will compile
748 669
them, as well.
749 670
*/
750 671

	
751 672
}
Ignore white space 6 line context
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
#ifndef LEMON_NETWORK_SIMPLEX_H
20 20
#define LEMON_NETWORK_SIMPLEX_H
21 21

	
22
/// \ingroup min_cost_flow
22
/// \ingroup min_cost_flow_algs
23 23
///
24 24
/// \file
25 25
/// \brief Network Simplex algorithm for finding a minimum cost flow.
26 26

	
27 27
#include <vector>
28 28
#include <limits>
29 29
#include <algorithm>
30 30

	
31 31
#include <lemon/core.h>
32 32
#include <lemon/math.h>
33 33

	
34 34
namespace lemon {
35 35

	
36
  /// \addtogroup min_cost_flow
36
  /// \addtogroup min_cost_flow_algs
37 37
  /// @{
38 38

	
39 39
  /// \brief Implementation of the primal Network Simplex algorithm
40 40
  /// for finding a \ref min_cost_flow "minimum cost flow".
41 41
  ///
42 42
  /// \ref NetworkSimplex implements the primal Network Simplex algorithm
43 43
  /// for finding a \ref min_cost_flow "minimum cost flow".
44 44
  /// This algorithm is a specialized version of the linear programming
45 45
  /// simplex method directly for the minimum cost flow problem.
46 46
  /// It is one of the most efficient solution methods.
47 47
  ///
48 48
  /// In general this class is the fastest implementation available
49 49
  /// in LEMON for the minimum cost flow problem.
50 50
  /// Moreover it supports both directions of the supply/demand inequality
51 51
  /// constraints. For more information see \ref SupplyType.
52 52
  ///
53 53
  /// Most of the parameters of the problem (except for the digraph)
54 54
  /// can be given using separate functions, and the algorithm can be
55 55
  /// executed using the \ref run() function. If some parameters are not
56 56
  /// specified, then default values will be used.
57 57
  ///
58 58
  /// \tparam GR The digraph type the algorithm runs on.
59 59
  /// \tparam V The value type used for flow amounts, capacity bounds
60 60
  /// and supply values in the algorithm. By default it is \c int.
61 61
  /// \tparam C The value type used for costs and potentials in the
62 62
  /// algorithm. By default it is the same as \c V.
63 63
  ///
64 64
  /// \warning Both value types must be signed and all input data must
65 65
  /// be integer.
66 66
  ///
67 67
  /// \note %NetworkSimplex provides five different pivot rule
68 68
  /// implementations, from which the most efficient one is used
69 69
  /// by default. For more information see \ref PivotRule.
70 70
  template <typename GR, typename V = int, typename C = V>
71 71
  class NetworkSimplex
72 72
  {
73 73
  public:
74 74

	
75 75
    /// The type of the flow amounts, capacity bounds and supply values
76 76
    typedef V Value;
77 77
    /// The type of the arc costs
78 78
    typedef C Cost;
79 79

	
80 80
  public:
81 81

	
82 82
    /// \brief Problem type constants for the \c run() function.
83 83
    ///
84 84
    /// Enum type containing the problem type constants that can be
85 85
    /// returned by the \ref run() function of the algorithm.
86 86
    enum ProblemType {
87 87
      /// The problem has no feasible solution (flow).
88 88
      INFEASIBLE,
89 89
      /// The problem has optimal solution (i.e. it is feasible and
90 90
      /// bounded), and the algorithm has found optimal flow and node
91 91
      /// potentials (primal and dual solutions).
92 92
      OPTIMAL,
93 93
      /// The objective function of the problem is unbounded, i.e.
94 94
      /// there is a directed cycle having negative total cost and
95 95
      /// infinite upper bound.
96 96
      UNBOUNDED
97 97
    };
98 98
    
99 99
    /// \brief Constants for selecting the type of the supply constraints.
100 100
    ///
101 101
    /// Enum type containing constants for selecting the supply type,
102 102
    /// i.e. the direction of the inequalities in the supply/demand
103 103
    /// constraints of the \ref min_cost_flow "minimum cost flow problem".
104 104
    ///
105
    /// The default supply type is \c GEQ, since this form is supported
106
    /// by other minimum cost flow algorithms and the \ref Circulation
107
    /// algorithm, as well.
108
    /// The \c LEQ problem type can be selected using the \ref supplyType()
109
    /// function.
110
    ///
111
    /// Note that the equality form is a special case of both supply types.
105
    /// The default supply type is \c GEQ, the \c LEQ type can be
106
    /// selected using \ref supplyType().
107
    /// The equality form is a special case of both supply types.
112 108
    enum SupplyType {
113

	
114 109
      /// This option means that there are <em>"greater or equal"</em>
115
      /// supply/demand constraints in the definition, i.e. the exact
116
      /// formulation of the problem is the following.
117
      /**
118
          \f[ \min\sum_{uv\in A} f(uv) \cdot cost(uv) \f]
119
          \f[ \sum_{uv\in A} f(uv) - \sum_{vu\in A} f(vu) \geq
120
              sup(u) \quad \forall u\in V \f]
121
          \f[ lower(uv) \leq f(uv) \leq upper(uv) \quad \forall uv\in A \f]
122
      */
123
      /// It means that the total demand must be greater or equal to the 
124
      /// total supply (i.e. \f$\sum_{u\in V} sup(u)\f$ must be zero or
125
      /// negative) and all the supplies have to be carried out from 
126
      /// the supply nodes, but there could be demands that are not 
127
      /// satisfied.
110
      /// supply/demand constraints in the definition of the problem.
128 111
      GEQ,
129
      /// It is just an alias for the \c GEQ option.
130
      CARRY_SUPPLIES = GEQ,
131

	
132 112
      /// This option means that there are <em>"less or equal"</em>
133
      /// supply/demand constraints in the definition, i.e. the exact
134
      /// formulation of the problem is the following.
135
      /**
136
          \f[ \min\sum_{uv\in A} f(uv) \cdot cost(uv) \f]
137
          \f[ \sum_{uv\in A} f(uv) - \sum_{vu\in A} f(vu) \leq
138
              sup(u) \quad \forall u\in V \f]
139
          \f[ lower(uv) \leq f(uv) \leq upper(uv) \quad \forall uv\in A \f]
140
      */
141
      /// It means that the total demand must be less or equal to the 
142
      /// total supply (i.e. \f$\sum_{u\in V} sup(u)\f$ must be zero or
143
      /// positive) and all the demands have to be satisfied, but there
144
      /// could be supplies that are not carried out from the supply
145
      /// nodes.
146
      LEQ,
147
      /// It is just an alias for the \c LEQ option.
148
      SATISFY_DEMANDS = LEQ
113
      /// supply/demand constraints in the definition of the problem.
114
      LEQ
149 115
    };
150 116
    
151 117
    /// \brief Constants for selecting the pivot rule.
152 118
    ///
153 119
    /// Enum type containing constants for selecting the pivot rule for
154 120
    /// the \ref run() function.
155 121
    ///
156 122
    /// \ref NetworkSimplex provides five different pivot rule
157 123
    /// implementations that significantly affect the running time
158 124
    /// of the algorithm.
159 125
    /// By default \ref BLOCK_SEARCH "Block Search" is used, which
160 126
    /// proved to be the most efficient and the most robust on various
161 127
    /// test inputs according to our benchmark tests.
162 128
    /// However another pivot rule can be selected using the \ref run()
163 129
    /// function with the proper parameter.
164 130
    enum PivotRule {
165 131

	
166 132
      /// The First Eligible pivot rule.
167 133
      /// The next eligible arc is selected in a wraparound fashion
168 134
      /// in every iteration.
169 135
      FIRST_ELIGIBLE,
170 136

	
171 137
      /// The Best Eligible pivot rule.
172 138
      /// The best eligible arc is selected in every iteration.
173 139
      BEST_ELIGIBLE,
174 140

	
175 141
      /// The Block Search pivot rule.
176 142
      /// A specified number of arcs are examined in every iteration
177 143
      /// in a wraparound fashion and the best eligible arc is selected
178 144
      /// from this block.
179 145
      BLOCK_SEARCH,
180 146

	
181 147
      /// The Candidate List pivot rule.
182 148
      /// In a major iteration a candidate list is built from eligible arcs
183 149
      /// in a wraparound fashion and in the following minor iterations
184 150
      /// the best eligible arc is selected from this list.
185 151
      CANDIDATE_LIST,
186 152

	
187 153
      /// The Altering Candidate List pivot rule.
188 154
      /// It is a modified version of the Candidate List method.
189 155
      /// It keeps only the several best eligible arcs from the former
190 156
      /// candidate list and extends this list in every iteration.
191 157
      ALTERING_LIST
192 158
    };
193 159
    
194 160
  private:
195 161

	
196 162
    TEMPLATE_DIGRAPH_TYPEDEFS(GR);
197 163

	
198 164
    typedef std::vector<Arc> ArcVector;
199 165
    typedef std::vector<Node> NodeVector;
200 166
    typedef std::vector<int> IntVector;
201 167
    typedef std::vector<bool> BoolVector;
202 168
    typedef std::vector<Value> ValueVector;
203 169
    typedef std::vector<Cost> CostVector;
204 170

	
205 171
    // State constants for arcs
206 172
    enum ArcStateEnum {
207 173
      STATE_UPPER = -1,
208 174
      STATE_TREE  =  0,
209 175
      STATE_LOWER =  1
210 176
    };
211 177

	
212 178
  private:
213 179

	
214 180
    // Data related to the underlying digraph
215 181
    const GR &_graph;
216 182
    int _node_num;
217 183
    int _arc_num;
184
    int _all_arc_num;
185
    int _search_arc_num;
218 186

	
219 187
    // Parameters of the problem
220 188
    bool _have_lower;
221 189
    SupplyType _stype;
222 190
    Value _sum_supply;
223 191

	
224 192
    // Data structures for storing the digraph
225 193
    IntNodeMap _node_id;
226 194
    IntArcMap _arc_id;
227 195
    IntVector _source;
228 196
    IntVector _target;
229 197

	
230 198
    // Node and arc data
231 199
    ValueVector _lower;
232 200
    ValueVector _upper;
233 201
    ValueVector _cap;
234 202
    CostVector _cost;
235 203
    ValueVector _supply;
236 204
    ValueVector _flow;
237 205
    CostVector _pi;
238 206

	
239 207
    // Data for storing the spanning tree structure
240 208
    IntVector _parent;
241 209
    IntVector _pred;
242 210
    IntVector _thread;
243 211
    IntVector _rev_thread;
244 212
    IntVector _succ_num;
245 213
    IntVector _last_succ;
246 214
    IntVector _dirty_revs;
247 215
    BoolVector _forward;
248 216
    IntVector _state;
249 217
    int _root;
250 218

	
251 219
    // Temporary data used in the current pivot iteration
252 220
    int in_arc, join, u_in, v_in, u_out, v_out;
253 221
    int first, second, right, last;
254 222
    int stem, par_stem, new_stem;
255 223
    Value delta;
256 224

	
257 225
  public:
258 226
  
259 227
    /// \brief Constant for infinite upper bounds (capacities).
260 228
    ///
261 229
    /// Constant for infinite upper bounds (capacities).
262 230
    /// It is \c std::numeric_limits<Value>::infinity() if available,
263 231
    /// \c std::numeric_limits<Value>::max() otherwise.
264 232
    const Value INF;
265 233

	
266 234
  private:
267 235

	
268 236
    // Implementation of the First Eligible pivot rule
269 237
    class FirstEligiblePivotRule
270 238
    {
271 239
    private:
272 240

	
273 241
      // References to the NetworkSimplex class
274 242
      const IntVector  &_source;
275 243
      const IntVector  &_target;
276 244
      const CostVector &_cost;
277 245
      const IntVector  &_state;
278 246
      const CostVector &_pi;
279 247
      int &_in_arc;
280
      int _arc_num;
248
      int _search_arc_num;
281 249

	
282 250
      // Pivot rule data
283 251
      int _next_arc;
284 252

	
285 253
    public:
286 254

	
287 255
      // Constructor
288 256
      FirstEligiblePivotRule(NetworkSimplex &ns) :
289 257
        _source(ns._source), _target(ns._target),
290 258
        _cost(ns._cost), _state(ns._state), _pi(ns._pi),
291
        _in_arc(ns.in_arc), _arc_num(ns._arc_num), _next_arc(0)
259
        _in_arc(ns.in_arc), _search_arc_num(ns._search_arc_num),
260
        _next_arc(0)
292 261
      {}
293 262

	
294 263
      // Find next entering arc
295 264
      bool findEnteringArc() {
296 265
        Cost c;
297
        for (int e = _next_arc; e < _arc_num; ++e) {
266
        for (int e = _next_arc; e < _search_arc_num; ++e) {
298 267
          c = _state[e] * (_cost[e] + _pi[_source[e]] - _pi[_target[e]]);
299 268
          if (c < 0) {
300 269
            _in_arc = e;
301 270
            _next_arc = e + 1;
302 271
            return true;
303 272
          }
304 273
        }
305 274
        for (int e = 0; e < _next_arc; ++e) {
306 275
          c = _state[e] * (_cost[e] + _pi[_source[e]] - _pi[_target[e]]);
307 276
          if (c < 0) {
308 277
            _in_arc = e;
309 278
            _next_arc = e + 1;
310 279
            return true;
311 280
          }
312 281
        }
313 282
        return false;
314 283
      }
315 284

	
316 285
    }; //class FirstEligiblePivotRule
317 286

	
318 287

	
319 288
    // Implementation of the Best Eligible pivot rule
320 289
    class BestEligiblePivotRule
321 290
    {
322 291
    private:
323 292

	
324 293
      // References to the NetworkSimplex class
325 294
      const IntVector  &_source;
326 295
      const IntVector  &_target;
327 296
      const CostVector &_cost;
328 297
      const IntVector  &_state;
329 298
      const CostVector &_pi;
330 299
      int &_in_arc;
331
      int _arc_num;
300
      int _search_arc_num;
332 301

	
333 302
    public:
334 303

	
335 304
      // Constructor
336 305
      BestEligiblePivotRule(NetworkSimplex &ns) :
337 306
        _source(ns._source), _target(ns._target),
338 307
        _cost(ns._cost), _state(ns._state), _pi(ns._pi),
339
        _in_arc(ns.in_arc), _arc_num(ns._arc_num)
308
        _in_arc(ns.in_arc), _search_arc_num(ns._search_arc_num)
340 309
      {}
341 310

	
342 311
      // Find next entering arc
343 312
      bool findEnteringArc() {
344 313
        Cost c, min = 0;
345
        for (int e = 0; e < _arc_num; ++e) {
314
        for (int e = 0; e < _search_arc_num; ++e) {
346 315
          c = _state[e] * (_cost[e] + _pi[_source[e]] - _pi[_target[e]]);
347 316
          if (c < min) {
348 317
            min = c;
349 318
            _in_arc = e;
350 319
          }
351 320
        }
352 321
        return min < 0;
353 322
      }
354 323

	
355 324
    }; //class BestEligiblePivotRule
356 325

	
357 326

	
358 327
    // Implementation of the Block Search pivot rule
359 328
    class BlockSearchPivotRule
360 329
    {
361 330
    private:
362 331

	
363 332
      // References to the NetworkSimplex class
364 333
      const IntVector  &_source;
365 334
      const IntVector  &_target;
366 335
      const CostVector &_cost;
367 336
      const IntVector  &_state;
368 337
      const CostVector &_pi;
369 338
      int &_in_arc;
370
      int _arc_num;
339
      int _search_arc_num;
371 340

	
372 341
      // Pivot rule data
373 342
      int _block_size;
374 343
      int _next_arc;
375 344

	
376 345
    public:
377 346

	
378 347
      // Constructor
379 348
      BlockSearchPivotRule(NetworkSimplex &ns) :
380 349
        _source(ns._source), _target(ns._target),
381 350
        _cost(ns._cost), _state(ns._state), _pi(ns._pi),
382
        _in_arc(ns.in_arc), _arc_num(ns._arc_num), _next_arc(0)
351
        _in_arc(ns.in_arc), _search_arc_num(ns._search_arc_num),
352
        _next_arc(0)
383 353
      {
384 354
        // The main parameters of the pivot rule
385
        const double BLOCK_SIZE_FACTOR = 2.0;
355
        const double BLOCK_SIZE_FACTOR = 0.5;
386 356
        const int MIN_BLOCK_SIZE = 10;
387 357

	
388 358
        _block_size = std::max( int(BLOCK_SIZE_FACTOR *
389
                                    std::sqrt(double(_arc_num))),
359
                                    std::sqrt(double(_search_arc_num))),
390 360
                                MIN_BLOCK_SIZE );
391 361
      }
392 362

	
393 363
      // Find next entering arc
394 364
      bool findEnteringArc() {
395 365
        Cost c, min = 0;
396 366
        int cnt = _block_size;
397 367
        int e, min_arc = _next_arc;
398
        for (e = _next_arc; e < _arc_num; ++e) {
368
        for (e = _next_arc; e < _search_arc_num; ++e) {
399 369
          c = _state[e] * (_cost[e] + _pi[_source[e]] - _pi[_target[e]]);
400 370
          if (c < min) {
401 371
            min = c;
402 372
            min_arc = e;
403 373
          }
404 374
          if (--cnt == 0) {
405 375
            if (min < 0) break;
406 376
            cnt = _block_size;
407 377
          }
408 378
        }
409 379
        if (min == 0 || cnt > 0) {
410 380
          for (e = 0; e < _next_arc; ++e) {
411 381
            c = _state[e] * (_cost[e] + _pi[_source[e]] - _pi[_target[e]]);
412 382
            if (c < min) {
413 383
              min = c;
414 384
              min_arc = e;
415 385
            }
416 386
            if (--cnt == 0) {
417 387
              if (min < 0) break;
418 388
              cnt = _block_size;
419 389
            }
420 390
          }
421 391
        }
422 392
        if (min >= 0) return false;
423 393
        _in_arc = min_arc;
424 394
        _next_arc = e;
425 395
        return true;
426 396
      }
427 397

	
428 398
    }; //class BlockSearchPivotRule
429 399

	
430 400

	
431 401
    // Implementation of the Candidate List pivot rule
432 402
    class CandidateListPivotRule
433 403
    {
434 404
    private:
435 405

	
436 406
      // References to the NetworkSimplex class
437 407
      const IntVector  &_source;
438 408
      const IntVector  &_target;
439 409
      const CostVector &_cost;
440 410
      const IntVector  &_state;
441 411
      const CostVector &_pi;
442 412
      int &_in_arc;
443
      int _arc_num;
413
      int _search_arc_num;
444 414

	
445 415
      // Pivot rule data
446 416
      IntVector _candidates;
447 417
      int _list_length, _minor_limit;
448 418
      int _curr_length, _minor_count;
449 419
      int _next_arc;
450 420

	
451 421
    public:
452 422

	
453 423
      /// Constructor
454 424
      CandidateListPivotRule(NetworkSimplex &ns) :
455 425
        _source(ns._source), _target(ns._target),
456 426
        _cost(ns._cost), _state(ns._state), _pi(ns._pi),
457
        _in_arc(ns.in_arc), _arc_num(ns._arc_num), _next_arc(0)
427
        _in_arc(ns.in_arc), _search_arc_num(ns._search_arc_num),
428
        _next_arc(0)
458 429
      {
459 430
        // The main parameters of the pivot rule
460 431
        const double LIST_LENGTH_FACTOR = 1.0;
461 432
        const int MIN_LIST_LENGTH = 10;
462 433
        const double MINOR_LIMIT_FACTOR = 0.1;
463 434
        const int MIN_MINOR_LIMIT = 3;
464 435

	
465 436
        _list_length = std::max( int(LIST_LENGTH_FACTOR *
466
                                     std::sqrt(double(_arc_num))),
437
                                     std::sqrt(double(_search_arc_num))),
467 438
                                 MIN_LIST_LENGTH );
468 439
        _minor_limit = std::max( int(MINOR_LIMIT_FACTOR * _list_length),
469 440
                                 MIN_MINOR_LIMIT );
470 441
        _curr_length = _minor_count = 0;
471 442
        _candidates.resize(_list_length);
472 443
      }
473 444

	
474 445
      /// Find next entering arc
475 446
      bool findEnteringArc() {
476 447
        Cost min, c;
477 448
        int e, min_arc = _next_arc;
478 449
        if (_curr_length > 0 && _minor_count < _minor_limit) {
479 450
          // Minor iteration: select the best eligible arc from the
480 451
          // current candidate list
481 452
          ++_minor_count;
482 453
          min = 0;
483 454
          for (int i = 0; i < _curr_length; ++i) {
484 455
            e = _candidates[i];
485 456
            c = _state[e] * (_cost[e] + _pi[_source[e]] - _pi[_target[e]]);
486 457
            if (c < min) {
487 458
              min = c;
488 459
              min_arc = e;
489 460
            }
490 461
            if (c >= 0) {
491 462
              _candidates[i--] = _candidates[--_curr_length];
492 463
            }
493 464
          }
494 465
          if (min < 0) {
495 466
            _in_arc = min_arc;
496 467
            return true;
497 468
          }
498 469
        }
499 470

	
500 471
        // Major iteration: build a new candidate list
501 472
        min = 0;
502 473
        _curr_length = 0;
503
        for (e = _next_arc; e < _arc_num; ++e) {
474
        for (e = _next_arc; e < _search_arc_num; ++e) {
504 475
          c = _state[e] * (_cost[e] + _pi[_source[e]] - _pi[_target[e]]);
505 476
          if (c < 0) {
506 477
            _candidates[_curr_length++] = e;
507 478
            if (c < min) {
508 479
              min = c;
509 480
              min_arc = e;
510 481
            }
511 482
            if (_curr_length == _list_length) break;
512 483
          }
513 484
        }
514 485
        if (_curr_length < _list_length) {
515 486
          for (e = 0; e < _next_arc; ++e) {
516 487
            c = _state[e] * (_cost[e] + _pi[_source[e]] - _pi[_target[e]]);
517 488
            if (c < 0) {
518 489
              _candidates[_curr_length++] = e;
519 490
              if (c < min) {
520 491
                min = c;
521 492
                min_arc = e;
522 493
              }
523 494
              if (_curr_length == _list_length) break;
524 495
            }
525 496
          }
526 497
        }
527 498
        if (_curr_length == 0) return false;
528 499
        _minor_count = 1;
529 500
        _in_arc = min_arc;
530 501
        _next_arc = e;
531 502
        return true;
532 503
      }
533 504

	
534 505
    }; //class CandidateListPivotRule
535 506

	
536 507

	
537 508
    // Implementation of the Altering Candidate List pivot rule
538 509
    class AlteringListPivotRule
539 510
    {
540 511
    private:
541 512

	
542 513
      // References to the NetworkSimplex class
543 514
      const IntVector  &_source;
544 515
      const IntVector  &_target;
545 516
      const CostVector &_cost;
546 517
      const IntVector  &_state;
547 518
      const CostVector &_pi;
548 519
      int &_in_arc;
549
      int _arc_num;
520
      int _search_arc_num;
550 521

	
551 522
      // Pivot rule data
552 523
      int _block_size, _head_length, _curr_length;
553 524
      int _next_arc;
554 525
      IntVector _candidates;
555 526
      CostVector _cand_cost;
556 527

	
557 528
      // Functor class to compare arcs during sort of the candidate list
558 529
      class SortFunc
559 530
      {
560 531
      private:
561 532
        const CostVector &_map;
562 533
      public:
563 534
        SortFunc(const CostVector &map) : _map(map) {}
564 535
        bool operator()(int left, int right) {
565 536
          return _map[left] > _map[right];
566 537
        }
567 538
      };
568 539

	
569 540
      SortFunc _sort_func;
570 541

	
571 542
    public:
572 543

	
573 544
      // Constructor
574 545
      AlteringListPivotRule(NetworkSimplex &ns) :
575 546
        _source(ns._source), _target(ns._target),
576 547
        _cost(ns._cost), _state(ns._state), _pi(ns._pi),
577
        _in_arc(ns.in_arc), _arc_num(ns._arc_num),
578
        _next_arc(0), _cand_cost(ns._arc_num), _sort_func(_cand_cost)
548
        _in_arc(ns.in_arc), _search_arc_num(ns._search_arc_num),
549
        _next_arc(0), _cand_cost(ns._search_arc_num), _sort_func(_cand_cost)
579 550
      {
580 551
        // The main parameters of the pivot rule
581 552
        const double BLOCK_SIZE_FACTOR = 1.5;
582 553
        const int MIN_BLOCK_SIZE = 10;
583 554
        const double HEAD_LENGTH_FACTOR = 0.1;
584 555
        const int MIN_HEAD_LENGTH = 3;
585 556

	
586 557
        _block_size = std::max( int(BLOCK_SIZE_FACTOR *
587
                                    std::sqrt(double(_arc_num))),
558
                                    std::sqrt(double(_search_arc_num))),
588 559
                                MIN_BLOCK_SIZE );
589 560
        _head_length = std::max( int(HEAD_LENGTH_FACTOR * _block_size),
590 561
                                 MIN_HEAD_LENGTH );
591 562
        _candidates.resize(_head_length + _block_size);
592 563
        _curr_length = 0;
593 564
      }
594 565

	
595 566
      // Find next entering arc
596 567
      bool findEnteringArc() {
597 568
        // Check the current candidate list
598 569
        int e;
599 570
        for (int i = 0; i < _curr_length; ++i) {
600 571
          e = _candidates[i];
601 572
          _cand_cost[e] = _state[e] *
602 573
            (_cost[e] + _pi[_source[e]] - _pi[_target[e]]);
603 574
          if (_cand_cost[e] >= 0) {
604 575
            _candidates[i--] = _candidates[--_curr_length];
605 576
          }
606 577
        }
607 578

	
608 579
        // Extend the list
609 580
        int cnt = _block_size;
610 581
        int last_arc = 0;
611 582
        int limit = _head_length;
612 583

	
613
        for (int e = _next_arc; e < _arc_num; ++e) {
584
        for (int e = _next_arc; e < _search_arc_num; ++e) {
614 585
          _cand_cost[e] = _state[e] *
615 586
            (_cost[e] + _pi[_source[e]] - _pi[_target[e]]);
616 587
          if (_cand_cost[e] < 0) {
617 588
            _candidates[_curr_length++] = e;
618 589
            last_arc = e;
619 590
          }
620 591
          if (--cnt == 0) {
621 592
            if (_curr_length > limit) break;
622 593
            limit = 0;
623 594
            cnt = _block_size;
624 595
          }
625 596
        }
626 597
        if (_curr_length <= limit) {
627 598
          for (int e = 0; e < _next_arc; ++e) {
628 599
            _cand_cost[e] = _state[e] *
629 600
              (_cost[e] + _pi[_source[e]] - _pi[_target[e]]);
630 601
            if (_cand_cost[e] < 0) {
631 602
              _candidates[_curr_length++] = e;
632 603
              last_arc = e;
633 604
            }
634 605
            if (--cnt == 0) {
635 606
              if (_curr_length > limit) break;
636 607
              limit = 0;
637 608
              cnt = _block_size;
638 609
            }
639 610
          }
640 611
        }
641 612
        if (_curr_length == 0) return false;
642 613
        _next_arc = last_arc + 1;
643 614

	
644 615
        // Make heap of the candidate list (approximating a partial sort)
645 616
        make_heap( _candidates.begin(), _candidates.begin() + _curr_length,
646 617
                   _sort_func );
647 618

	
648 619
        // Pop the first element of the heap
649 620
        _in_arc = _candidates[0];
650 621
        pop_heap( _candidates.begin(), _candidates.begin() + _curr_length,
651 622
                  _sort_func );
652 623
        _curr_length = std::min(_head_length, _curr_length - 1);
653 624
        return true;
654 625
      }
655 626

	
656 627
    }; //class AlteringListPivotRule
657 628

	
658 629
  public:
659 630

	
660 631
    /// \brief Constructor.
661 632
    ///
662 633
    /// The constructor of the class.
663 634
    ///
664 635
    /// \param graph The digraph the algorithm runs on.
665 636
    NetworkSimplex(const GR& graph) :
666 637
      _graph(graph), _node_id(graph), _arc_id(graph),
667 638
      INF(std::numeric_limits<Value>::has_infinity ?
668 639
          std::numeric_limits<Value>::infinity() :
669 640
          std::numeric_limits<Value>::max())
670 641
    {
671 642
      // Check the value types
672 643
      LEMON_ASSERT(std::numeric_limits<Value>::is_signed,
673 644
        "The flow type of NetworkSimplex must be signed");
674 645
      LEMON_ASSERT(std::numeric_limits<Cost>::is_signed,
675 646
        "The cost type of NetworkSimplex must be signed");
676 647
        
677 648
      // Resize vectors
678 649
      _node_num = countNodes(_graph);
679 650
      _arc_num = countArcs(_graph);
680 651
      int all_node_num = _node_num + 1;
681
      int all_arc_num = _arc_num + _node_num;
652
      int max_arc_num = _arc_num + 2 * _node_num;
682 653

	
683
      _source.resize(all_arc_num);
684
      _target.resize(all_arc_num);
654
      _source.resize(max_arc_num);
655
      _target.resize(max_arc_num);
685 656

	
686
      _lower.resize(all_arc_num);
687
      _upper.resize(all_arc_num);
688
      _cap.resize(all_arc_num);
689
      _cost.resize(all_arc_num);
657
      _lower.resize(_arc_num);
658
      _upper.resize(_arc_num);
659
      _cap.resize(max_arc_num);
660
      _cost.resize(max_arc_num);
690 661
      _supply.resize(all_node_num);
691
      _flow.resize(all_arc_num);
662
      _flow.resize(max_arc_num);
692 663
      _pi.resize(all_node_num);
693 664

	
694 665
      _parent.resize(all_node_num);
695 666
      _pred.resize(all_node_num);
696 667
      _forward.resize(all_node_num);
697 668
      _thread.resize(all_node_num);
698 669
      _rev_thread.resize(all_node_num);
699 670
      _succ_num.resize(all_node_num);
700 671
      _last_succ.resize(all_node_num);
701
      _state.resize(all_arc_num);
672
      _state.resize(max_arc_num);
702 673

	
703 674
      // Copy the graph (store the arcs in a mixed order)
704 675
      int i = 0;
705 676
      for (NodeIt n(_graph); n != INVALID; ++n, ++i) {
706 677
        _node_id[n] = i;
707 678
      }
708 679
      int k = std::max(int(std::sqrt(double(_arc_num))), 10);
709 680
      i = 0;
710 681
      for (ArcIt a(_graph); a != INVALID; ++a) {
711 682
        _arc_id[a] = i;
712 683
        _source[i] = _node_id[_graph.source(a)];
713 684
        _target[i] = _node_id[_graph.target(a)];
714 685
        if ((i += k) >= _arc_num) i = (i % k) + 1;
715 686
      }
716 687
      
717 688
      // Initialize maps
718 689
      for (int i = 0; i != _node_num; ++i) {
719 690
        _supply[i] = 0;
720 691
      }
721 692
      for (int i = 0; i != _arc_num; ++i) {
722 693
        _lower[i] = 0;
723 694
        _upper[i] = INF;
724 695
        _cost[i] = 1;
725 696
      }
726 697
      _have_lower = false;
727 698
      _stype = GEQ;
728 699
    }
729 700

	
730 701
    /// \name Parameters
731 702
    /// The parameters of the algorithm can be specified using these
732 703
    /// functions.
733 704

	
734 705
    /// @{
735 706

	
736 707
    /// \brief Set the lower bounds on the arcs.
737 708
    ///
738 709
    /// This function sets the lower bounds on the arcs.
739 710
    /// If it is not used before calling \ref run(), the lower bounds
740 711
    /// will be set to zero on all arcs.
741 712
    ///
742 713
    /// \param map An arc map storing the lower bounds.
743 714
    /// Its \c Value type must be convertible to the \c Value type
744 715
    /// of the algorithm.
745 716
    ///
746 717
    /// \return <tt>(*this)</tt>
747 718
    template <typename LowerMap>
748 719
    NetworkSimplex& lowerMap(const LowerMap& map) {
749 720
      _have_lower = true;
750 721
      for (ArcIt a(_graph); a != INVALID; ++a) {
751 722
        _lower[_arc_id[a]] = map[a];
752 723
      }
753 724
      return *this;
754 725
    }
755 726

	
756 727
    /// \brief Set the upper bounds (capacities) on the arcs.
757 728
    ///
758 729
    /// This function sets the upper bounds (capacities) on the arcs.
759 730
    /// If it is not used before calling \ref run(), the upper bounds
760 731
    /// will be set to \ref INF on all arcs (i.e. the flow value will be
761 732
    /// unbounded from above on each arc).
762 733
    ///
763 734
    /// \param map An arc map storing the upper bounds.
764 735
    /// Its \c Value type must be convertible to the \c Value type
765 736
    /// of the algorithm.
766 737
    ///
767 738
    /// \return <tt>(*this)</tt>
768 739
    template<typename UpperMap>
769 740
    NetworkSimplex& upperMap(const UpperMap& map) {
770 741
      for (ArcIt a(_graph); a != INVALID; ++a) {
771 742
        _upper[_arc_id[a]] = map[a];
772 743
      }
773 744
      return *this;
774 745
    }
775 746

	
776 747
    /// \brief Set the costs of the arcs.
777 748
    ///
778 749
    /// This function sets the costs of the arcs.
779 750
    /// If it is not used before calling \ref run(), the costs
780 751
    /// will be set to \c 1 on all arcs.
781 752
    ///
782 753
    /// \param map An arc map storing the costs.
783 754
    /// Its \c Value type must be convertible to the \c Cost type
784 755
    /// of the algorithm.
785 756
    ///
786 757
    /// \return <tt>(*this)</tt>
787 758
    template<typename CostMap>
788 759
    NetworkSimplex& costMap(const CostMap& map) {
789 760
      for (ArcIt a(_graph); a != INVALID; ++a) {
790 761
        _cost[_arc_id[a]] = map[a];
791 762
      }
792 763
      return *this;
793 764
    }
794 765

	
795 766
    /// \brief Set the supply values of the nodes.
796 767
    ///
797 768
    /// This function sets the supply values of the nodes.
798 769
    /// If neither this function nor \ref stSupply() is used before
799 770
    /// calling \ref run(), the supply of each node will be set to zero.
800 771
    /// (It makes sense only if non-zero lower bounds are given.)
801 772
    ///
802 773
    /// \param map A node map storing the supply values.
803 774
    /// Its \c Value type must be convertible to the \c Value type
804 775
    /// of the algorithm.
805 776
    ///
806 777
    /// \return <tt>(*this)</tt>
807 778
    template<typename SupplyMap>
808 779
    NetworkSimplex& supplyMap(const SupplyMap& map) {
809 780
      for (NodeIt n(_graph); n != INVALID; ++n) {
810 781
        _supply[_node_id[n]] = map[n];
811 782
      }
812 783
      return *this;
813 784
    }
814 785

	
815 786
    /// \brief Set single source and target nodes and a supply value.
816 787
    ///
817 788
    /// This function sets a single source node and a single target node
818 789
    /// and the required flow value.
819 790
    /// If neither this function nor \ref supplyMap() is used before
820 791
    /// calling \ref run(), the supply of each node will be set to zero.
821 792
    /// (It makes sense only if non-zero lower bounds are given.)
822 793
    ///
823 794
    /// Using this function has the same effect as using \ref supplyMap()
824 795
    /// with such a map in which \c k is assigned to \c s, \c -k is
825 796
    /// assigned to \c t and all other nodes have zero supply value.
826 797
    ///
827 798
    /// \param s The source node.
828 799
    /// \param t The target node.
829 800
    /// \param k The required amount of flow from node \c s to node \c t
830 801
    /// (i.e. the supply of \c s and the demand of \c t).
831 802
    ///
832 803
    /// \return <tt>(*this)</tt>
833 804
    NetworkSimplex& stSupply(const Node& s, const Node& t, Value k) {
834 805
      for (int i = 0; i != _node_num; ++i) {
835 806
        _supply[i] = 0;
836 807
      }
837 808
      _supply[_node_id[s]] =  k;
838 809
      _supply[_node_id[t]] = -k;
839 810
      return *this;
840 811
    }
841 812
    
842 813
    /// \brief Set the type of the supply constraints.
843 814
    ///
844 815
    /// This function sets the type of the supply/demand constraints.
845 816
    /// If it is not used before calling \ref run(), the \ref GEQ supply
846 817
    /// type will be used.
847 818
    ///
848 819
    /// For more information see \ref SupplyType.
849 820
    ///
850 821
    /// \return <tt>(*this)</tt>
851 822
    NetworkSimplex& supplyType(SupplyType supply_type) {
852 823
      _stype = supply_type;
853 824
      return *this;
854 825
    }
855 826

	
856 827
    /// @}
857 828

	
858 829
    /// \name Execution Control
859 830
    /// The algorithm can be executed using \ref run().
860 831

	
861 832
    /// @{
862 833

	
863 834
    /// \brief Run the algorithm.
864 835
    ///
865 836
    /// This function runs the algorithm.
866 837
    /// The paramters can be specified using functions \ref lowerMap(),
867 838
    /// \ref upperMap(), \ref costMap(), \ref supplyMap(), \ref stSupply(), 
868 839
    /// \ref supplyType().
869 840
    /// For example,
870 841
    /// \code
871 842
    ///   NetworkSimplex<ListDigraph> ns(graph);
872 843
    ///   ns.lowerMap(lower).upperMap(upper).costMap(cost)
873 844
    ///     .supplyMap(sup).run();
874 845
    /// \endcode
875 846
    ///
876 847
    /// This function can be called more than once. All the parameters
877 848
    /// that have been given are kept for the next call, unless
878 849
    /// \ref reset() is called, thus only the modified parameters
879 850
    /// have to be set again. See \ref reset() for examples.
880 851
    /// However the underlying digraph must not be modified after this
881 852
    /// class have been constructed, since it copies and extends the graph.
882 853
    ///
883 854
    /// \param pivot_rule The pivot rule that will be used during the
884 855
    /// algorithm. For more information see \ref PivotRule.
885 856
    ///
886 857
    /// \return \c INFEASIBLE if no feasible flow exists,
887 858
    /// \n \c OPTIMAL if the problem has optimal solution
888 859
    /// (i.e. it is feasible and bounded), and the algorithm has found
889 860
    /// optimal flow and node potentials (primal and dual solutions),
890 861
    /// \n \c UNBOUNDED if the objective function of the problem is
891 862
    /// unbounded, i.e. there is a directed cycle having negative total
892 863
    /// cost and infinite upper bound.
893 864
    ///
894 865
    /// \see ProblemType, PivotRule
895 866
    ProblemType run(PivotRule pivot_rule = BLOCK_SEARCH) {
896 867
      if (!init()) return INFEASIBLE;
897 868
      return start(pivot_rule);
898 869
    }
899 870

	
900 871
    /// \brief Reset all the parameters that have been given before.
901 872
    ///
902 873
    /// This function resets all the paramaters that have been given
903 874
    /// before using functions \ref lowerMap(), \ref upperMap(),
904 875
    /// \ref costMap(), \ref supplyMap(), \ref stSupply(), \ref supplyType().
905 876
    ///
906 877
    /// It is useful for multiple run() calls. If this function is not
907 878
    /// used, all the parameters given before are kept for the next
908 879
    /// \ref run() call.
909 880
    /// However the underlying digraph must not be modified after this
910 881
    /// class have been constructed, since it copies and extends the graph.
911 882
    ///
912 883
    /// For example,
913 884
    /// \code
914 885
    ///   NetworkSimplex<ListDigraph> ns(graph);
915 886
    ///
916 887
    ///   // First run
917 888
    ///   ns.lowerMap(lower).upperMap(upper).costMap(cost)
918 889
    ///     .supplyMap(sup).run();
919 890
    ///
920 891
    ///   // Run again with modified cost map (reset() is not called,
921 892
    ///   // so only the cost map have to be set again)
922 893
    ///   cost[e] += 100;
923 894
    ///   ns.costMap(cost).run();
924 895
    ///
925 896
    ///   // Run again from scratch using reset()
926 897
    ///   // (the lower bounds will be set to zero on all arcs)
927 898
    ///   ns.reset();
928 899
    ///   ns.upperMap(capacity).costMap(cost)
929 900
    ///     .supplyMap(sup).run();
930 901
    /// \endcode
931 902
    ///
932 903
    /// \return <tt>(*this)</tt>
933 904
    NetworkSimplex& reset() {
934 905
      for (int i = 0; i != _node_num; ++i) {
935 906
        _supply[i] = 0;
936 907
      }
937 908
      for (int i = 0; i != _arc_num; ++i) {
938 909
        _lower[i] = 0;
939 910
        _upper[i] = INF;
940 911
        _cost[i] = 1;
941 912
      }
942 913
      _have_lower = false;
943 914
      _stype = GEQ;
944 915
      return *this;
945 916
    }
946 917

	
947 918
    /// @}
948 919

	
949 920
    /// \name Query Functions
950 921
    /// The results of the algorithm can be obtained using these
951 922
    /// functions.\n
952 923
    /// The \ref run() function must be called before using them.
953 924

	
954 925
    /// @{
955 926

	
956 927
    /// \brief Return the total cost of the found flow.
957 928
    ///
958 929
    /// This function returns the total cost of the found flow.
959 930
    /// Its complexity is O(e).
960 931
    ///
961 932
    /// \note The return type of the function can be specified as a
962 933
    /// template parameter. For example,
963 934
    /// \code
964 935
    ///   ns.totalCost<double>();
965 936
    /// \endcode
966 937
    /// It is useful if the total cost cannot be stored in the \c Cost
967 938
    /// type of the algorithm, which is the default return type of the
968 939
    /// function.
969 940
    ///
970 941
    /// \pre \ref run() must be called before using this function.
971 942
    template <typename Number>
972 943
    Number totalCost() const {
973 944
      Number c = 0;
974 945
      for (ArcIt a(_graph); a != INVALID; ++a) {
975 946
        int i = _arc_id[a];
976 947
        c += Number(_flow[i]) * Number(_cost[i]);
977 948
      }
978 949
      return c;
979 950
    }
980 951

	
981 952
#ifndef DOXYGEN
982 953
    Cost totalCost() const {
983 954
      return totalCost<Cost>();
984 955
    }
985 956
#endif
986 957

	
987 958
    /// \brief Return the flow on the given arc.
988 959
    ///
989 960
    /// This function returns the flow on the given arc.
990 961
    ///
991 962
    /// \pre \ref run() must be called before using this function.
992 963
    Value flow(const Arc& a) const {
993 964
      return _flow[_arc_id[a]];
994 965
    }
995 966

	
996 967
    /// \brief Return the flow map (the primal solution).
997 968
    ///
998 969
    /// This function copies the flow value on each arc into the given
999 970
    /// map. The \c Value type of the algorithm must be convertible to
1000 971
    /// the \c Value type of the map.
1001 972
    ///
1002 973
    /// \pre \ref run() must be called before using this function.
1003 974
    template <typename FlowMap>
1004 975
    void flowMap(FlowMap &map) const {
1005 976
      for (ArcIt a(_graph); a != INVALID; ++a) {
1006 977
        map.set(a, _flow[_arc_id[a]]);
1007 978
      }
1008 979
    }
1009 980

	
1010 981
    /// \brief Return the potential (dual value) of the given node.
1011 982
    ///
1012 983
    /// This function returns the potential (dual value) of the
1013 984
    /// given node.
1014 985
    ///
1015 986
    /// \pre \ref run() must be called before using this function.
1016 987
    Cost potential(const Node& n) const {
1017 988
      return _pi[_node_id[n]];
1018 989
    }
1019 990

	
1020 991
    /// \brief Return the potential map (the dual solution).
1021 992
    ///
1022 993
    /// This function copies the potential (dual value) of each node
1023 994
    /// into the given map.
1024 995
    /// The \c Cost type of the algorithm must be convertible to the
1025 996
    /// \c Value type of the map.
1026 997
    ///
1027 998
    /// \pre \ref run() must be called before using this function.
1028 999
    template <typename PotentialMap>
1029 1000
    void potentialMap(PotentialMap &map) const {
1030 1001
      for (NodeIt n(_graph); n != INVALID; ++n) {
1031 1002
        map.set(n, _pi[_node_id[n]]);
1032 1003
      }
1033 1004
    }
1034 1005

	
1035 1006
    /// @}
1036 1007

	
1037 1008
  private:
1038 1009

	
1039 1010
    // Initialize internal data structures
1040 1011
    bool init() {
1041 1012
      if (_node_num == 0) return false;
1042 1013

	
1043 1014
      // Check the sum of supply values
1044 1015
      _sum_supply = 0;
1045 1016
      for (int i = 0; i != _node_num; ++i) {
1046 1017
        _sum_supply += _supply[i];
1047 1018
      }
1048 1019
      if ( !((_stype == GEQ && _sum_supply <= 0) ||
1049 1020
             (_stype == LEQ && _sum_supply >= 0)) ) return false;
1050 1021

	
1051 1022
      // Remove non-zero lower bounds
1052 1023
      if (_have_lower) {
1053 1024
        for (int i = 0; i != _arc_num; ++i) {
1054 1025
          Value c = _lower[i];
1055 1026
          if (c >= 0) {
1056 1027
            _cap[i] = _upper[i] < INF ? _upper[i] - c : INF;
1057 1028
          } else {
1058 1029
            _cap[i] = _upper[i] < INF + c ? _upper[i] - c : INF;
1059 1030
          }
1060 1031
          _supply[_source[i]] -= c;
1061 1032
          _supply[_target[i]] += c;
1062 1033
        }
1063 1034
      } else {
1064 1035
        for (int i = 0; i != _arc_num; ++i) {
1065 1036
          _cap[i] = _upper[i];
1066 1037
        }
1067 1038
      }
1068 1039

	
1069 1040
      // Initialize artifical cost
1070 1041
      Cost ART_COST;
1071 1042
      if (std::numeric_limits<Cost>::is_exact) {
1072
        ART_COST = std::numeric_limits<Cost>::max() / 4 + 1;
1043
        ART_COST = std::numeric_limits<Cost>::max() / 2 + 1;
1073 1044
      } else {
1074 1045
        ART_COST = std::numeric_limits<Cost>::min();
1075 1046
        for (int i = 0; i != _arc_num; ++i) {
1076 1047
          if (_cost[i] > ART_COST) ART_COST = _cost[i];
1077 1048
        }
1078 1049
        ART_COST = (ART_COST + 1) * _node_num;
1079 1050
      }
1080 1051

	
1081 1052
      // Initialize arc maps
1082 1053
      for (int i = 0; i != _arc_num; ++i) {
1083 1054
        _flow[i] = 0;
1084 1055
        _state[i] = STATE_LOWER;
1085 1056
      }
1086 1057
      
1087 1058
      // Set data for the artificial root node
1088 1059
      _root = _node_num;
1089 1060
      _parent[_root] = -1;
1090 1061
      _pred[_root] = -1;
1091 1062
      _thread[_root] = 0;
1092 1063
      _rev_thread[0] = _root;
1093 1064
      _succ_num[_root] = _node_num + 1;
1094 1065
      _last_succ[_root] = _root - 1;
1095 1066
      _supply[_root] = -_sum_supply;
1096
      _pi[_root] = _sum_supply < 0 ? -ART_COST : ART_COST;
1067
      _pi[_root] = 0;
1097 1068

	
1098 1069
      // Add artificial arcs and initialize the spanning tree data structure
1099
      for (int u = 0, e = _arc_num; u != _node_num; ++u, ++e) {
1100
        _parent[u] = _root;
1101
        _pred[u] = e;
1102
        _thread[u] = u + 1;
1103
        _rev_thread[u + 1] = u;
1104
        _succ_num[u] = 1;
1105
        _last_succ[u] = u;
1106
        _cost[e] = ART_COST;
1107
        _cap[e] = INF;
1108
        _state[e] = STATE_TREE;
1109
        if (_supply[u] > 0 || (_supply[u] == 0 && _sum_supply <= 0)) {
1110
          _flow[e] = _supply[u];
1111
          _forward[u] = true;
1112
          _pi[u] = -ART_COST + _pi[_root];
1113
        } else {
1114
          _flow[e] = -_supply[u];
1115
          _forward[u] = false;
1116
          _pi[u] = ART_COST + _pi[_root];
1070
      if (_sum_supply == 0) {
1071
        // EQ supply constraints
1072
        _search_arc_num = _arc_num;
1073
        _all_arc_num = _arc_num + _node_num;
1074
        for (int u = 0, e = _arc_num; u != _node_num; ++u, ++e) {
1075
          _parent[u] = _root;
1076
          _pred[u] = e;
1077
          _thread[u] = u + 1;
1078
          _rev_thread[u + 1] = u;
1079
          _succ_num[u] = 1;
1080
          _last_succ[u] = u;
1081
          _cap[e] = INF;
1082
          _state[e] = STATE_TREE;
1083
          if (_supply[u] >= 0) {
1084
            _forward[u] = true;
1085
            _pi[u] = 0;
1086
            _source[e] = u;
1087
            _target[e] = _root;
1088
            _flow[e] = _supply[u];
1089
            _cost[e] = 0;
1090
          } else {
1091
            _forward[u] = false;
1092
            _pi[u] = ART_COST;
1093
            _source[e] = _root;
1094
            _target[e] = u;
1095
            _flow[e] = -_supply[u];
1096
            _cost[e] = ART_COST;
1097
          }
1117 1098
        }
1118 1099
      }
1100
      else if (_sum_supply > 0) {
1101
        // LEQ supply constraints
1102
        _search_arc_num = _arc_num + _node_num;
1103
        int f = _arc_num + _node_num;
1104
        for (int u = 0, e = _arc_num; u != _node_num; ++u, ++e) {
1105
          _parent[u] = _root;
1106
          _thread[u] = u + 1;
1107
          _rev_thread[u + 1] = u;
1108
          _succ_num[u] = 1;
1109
          _last_succ[u] = u;
1110
          if (_supply[u] >= 0) {
1111
            _forward[u] = true;
1112
            _pi[u] = 0;
1113
            _pred[u] = e;
1114
            _source[e] = u;
1115
            _target[e] = _root;
1116
            _cap[e] = INF;
1117
            _flow[e] = _supply[u];
1118
            _cost[e] = 0;
1119
            _state[e] = STATE_TREE;
1120
          } else {
1121
            _forward[u] = false;
1122
            _pi[u] = ART_COST;
1123
            _pred[u] = f;
1124
            _source[f] = _root;
1125
            _target[f] = u;
1126
            _cap[f] = INF;
1127
            _flow[f] = -_supply[u];
1128
            _cost[f] = ART_COST;
1129
            _state[f] = STATE_TREE;
1130
            _source[e] = u;
1131
            _target[e] = _root;
1132
            _cap[e] = INF;
1133
            _flow[e] = 0;
1134
            _cost[e] = 0;
1135
            _state[e] = STATE_LOWER;
1136
            ++f;
1137
          }
1138
        }
1139
        _all_arc_num = f;
1140
      }
1141
      else {
1142
        // GEQ supply constraints
1143
        _search_arc_num = _arc_num + _node_num;
1144
        int f = _arc_num + _node_num;
1145
        for (int u = 0, e = _arc_num; u != _node_num; ++u, ++e) {
1146
          _parent[u] = _root;
1147
          _thread[u] = u + 1;
1148
          _rev_thread[u + 1] = u;
1149
          _succ_num[u] = 1;
1150
          _last_succ[u] = u;
1151
          if (_supply[u] <= 0) {
1152
            _forward[u] = false;
1153
            _pi[u] = 0;
1154
            _pred[u] = e;
1155
            _source[e] = _root;
1156
            _target[e] = u;
1157
            _cap[e] = INF;
1158
            _flow[e] = -_supply[u];
1159
            _cost[e] = 0;
1160
            _state[e] = STATE_TREE;
1161
          } else {
1162
            _forward[u] = true;
1163
            _pi[u] = -ART_COST;
1164
            _pred[u] = f;
1165
            _source[f] = u;
1166
            _target[f] = _root;
1167
            _cap[f] = INF;
1168
            _flow[f] = _supply[u];
1169
            _state[f] = STATE_TREE;
1170
            _cost[f] = ART_COST;
1171
            _source[e] = _root;
1172
            _target[e] = u;
1173
            _cap[e] = INF;
1174
            _flow[e] = 0;
1175
            _cost[e] = 0;
1176
            _state[e] = STATE_LOWER;
1177
            ++f;
1178
          }
1179
        }
1180
        _all_arc_num = f;
1181
      }
1119 1182

	
1120 1183
      return true;
1121 1184
    }
1122 1185

	
1123 1186
    // Find the join node
1124 1187
    void findJoinNode() {
1125 1188
      int u = _source[in_arc];
1126 1189
      int v = _target[in_arc];
1127 1190
      while (u != v) {
1128 1191
        if (_succ_num[u] < _succ_num[v]) {
1129 1192
          u = _parent[u];
1130 1193
        } else {
1131 1194
          v = _parent[v];
1132 1195
        }
1133 1196
      }
1134 1197
      join = u;
1135 1198
    }
1136 1199

	
1137 1200
    // Find the leaving arc of the cycle and returns true if the
1138 1201
    // leaving arc is not the same as the entering arc
1139 1202
    bool findLeavingArc() {
1140 1203
      // Initialize first and second nodes according to the direction
1141 1204
      // of the cycle
1142 1205
      if (_state[in_arc] == STATE_LOWER) {
1143 1206
        first  = _source[in_arc];
1144 1207
        second = _target[in_arc];
1145 1208
      } else {
1146 1209
        first  = _target[in_arc];
1147 1210
        second = _source[in_arc];
1148 1211
      }
1149 1212
      delta = _cap[in_arc];
1150 1213
      int result = 0;
1151 1214
      Value d;
1152 1215
      int e;
1153 1216

	
1154 1217
      // Search the cycle along the path form the first node to the root
1155 1218
      for (int u = first; u != join; u = _parent[u]) {
1156 1219
        e = _pred[u];
1157 1220
        d = _forward[u] ?
1158 1221
          _flow[e] : (_cap[e] == INF ? INF : _cap[e] - _flow[e]);
1159 1222
        if (d < delta) {
1160 1223
          delta = d;
1161 1224
          u_out = u;
1162 1225
          result = 1;
1163 1226
        }
1164 1227
      }
1165 1228
      // Search the cycle along the path form the second node to the root
1166 1229
      for (int u = second; u != join; u = _parent[u]) {
1167 1230
        e = _pred[u];
1168 1231
        d = _forward[u] ? 
1169 1232
          (_cap[e] == INF ? INF : _cap[e] - _flow[e]) : _flow[e];
1170 1233
        if (d <= delta) {
1171 1234
          delta = d;
1172 1235
          u_out = u;
1173 1236
          result = 2;
1174 1237
        }
1175 1238
      }
1176 1239

	
1177 1240
      if (result == 1) {
1178 1241
        u_in = first;
1179 1242
        v_in = second;
1180 1243
      } else {
1181 1244
        u_in = second;
1182 1245
        v_in = first;
1183 1246
      }
1184 1247
      return result != 0;
1185 1248
    }
1186 1249

	
1187 1250
    // Change _flow and _state vectors
1188 1251
    void changeFlow(bool change) {
1189 1252
      // Augment along the cycle
1190 1253
      if (delta > 0) {
1191 1254
        Value val = _state[in_arc] * delta;
1192 1255
        _flow[in_arc] += val;
1193 1256
        for (int u = _source[in_arc]; u != join; u = _parent[u]) {
1194 1257
          _flow[_pred[u]] += _forward[u] ? -val : val;
1195 1258
        }
1196 1259
        for (int u = _target[in_arc]; u != join; u = _parent[u]) {
1197 1260
          _flow[_pred[u]] += _forward[u] ? val : -val;
1198 1261
        }
1199 1262
      }
1200 1263
      // Update the state of the entering and leaving arcs
1201 1264
      if (change) {
1202 1265
        _state[in_arc] = STATE_TREE;
1203 1266
        _state[_pred[u_out]] =
1204 1267
          (_flow[_pred[u_out]] == 0) ? STATE_LOWER : STATE_UPPER;
1205 1268
      } else {
1206 1269
        _state[in_arc] = -_state[in_arc];
1207 1270
      }
1208 1271
    }
1209 1272

	
1210 1273
    // Update the tree structure
1211 1274
    void updateTreeStructure() {
1212 1275
      int u, w;
1213 1276
      int old_rev_thread = _rev_thread[u_out];
1214 1277
      int old_succ_num = _succ_num[u_out];
1215 1278
      int old_last_succ = _last_succ[u_out];
1216 1279
      v_out = _parent[u_out];
1217 1280

	
1218 1281
      u = _last_succ[u_in];  // the last successor of u_in
1219 1282
      right = _thread[u];    // the node after it
1220 1283

	
1221 1284
      // Handle the case when old_rev_thread equals to v_in
1222 1285
      // (it also means that join and v_out coincide)
1223 1286
      if (old_rev_thread == v_in) {
1224 1287
        last = _thread[_last_succ[u_out]];
1225 1288
      } else {
1226 1289
        last = _thread[v_in];
1227 1290
      }
1228 1291

	
1229 1292
      // Update _thread and _parent along the stem nodes (i.e. the nodes
1230 1293
      // between u_in and u_out, whose parent have to be changed)
1231 1294
      _thread[v_in] = stem = u_in;
1232 1295
      _dirty_revs.clear();
1233 1296
      _dirty_revs.push_back(v_in);
1234 1297
      par_stem = v_in;
1235 1298
      while (stem != u_out) {
1236 1299
        // Insert the next stem node into the thread list
1237 1300
        new_stem = _parent[stem];
1238 1301
        _thread[u] = new_stem;
1239 1302
        _dirty_revs.push_back(u);
1240 1303

	
1241 1304
        // Remove the subtree of stem from the thread list
1242 1305
        w = _rev_thread[stem];
1243 1306
        _thread[w] = right;
1244 1307
        _rev_thread[right] = w;
1245 1308

	
1246 1309
        // Change the parent node and shift stem nodes
1247 1310
        _parent[stem] = par_stem;
1248 1311
        par_stem = stem;
1249 1312
        stem = new_stem;
1250 1313

	
1251 1314
        // Update u and right
1252 1315
        u = _last_succ[stem] == _last_succ[par_stem] ?
1253 1316
          _rev_thread[par_stem] : _last_succ[stem];
1254 1317
        right = _thread[u];
1255 1318
      }
1256 1319
      _parent[u_out] = par_stem;
1257 1320
      _thread[u] = last;
1258 1321
      _rev_thread[last] = u;
1259 1322
      _last_succ[u_out] = u;
1260 1323

	
1261 1324
      // Remove the subtree of u_out from the thread list except for
1262 1325
      // the case when old_rev_thread equals to v_in
1263 1326
      // (it also means that join and v_out coincide)
1264 1327
      if (old_rev_thread != v_in) {
1265 1328
        _thread[old_rev_thread] = right;
1266 1329
        _rev_thread[right] = old_rev_thread;
1267 1330
      }
1268 1331

	
1269 1332
      // Update _rev_thread using the new _thread values
1270 1333
      for (int i = 0; i < int(_dirty_revs.size()); ++i) {
1271 1334
        u = _dirty_revs[i];
1272 1335
        _rev_thread[_thread[u]] = u;
1273 1336
      }
1274 1337

	
1275 1338
      // Update _pred, _forward, _last_succ and _succ_num for the
1276 1339
      // stem nodes from u_out to u_in
1277 1340
      int tmp_sc = 0, tmp_ls = _last_succ[u_out];
1278 1341
      u = u_out;
1279 1342
      while (u != u_in) {
1280 1343
        w = _parent[u];
1281 1344
        _pred[u] = _pred[w];
1282 1345
        _forward[u] = !_forward[w];
1283 1346
        tmp_sc += _succ_num[u] - _succ_num[w];
1284 1347
        _succ_num[u] = tmp_sc;
1285 1348
        _last_succ[w] = tmp_ls;
1286 1349
        u = w;
1287 1350
      }
1288 1351
      _pred[u_in] = in_arc;
1289 1352
      _forward[u_in] = (u_in == _source[in_arc]);
1290 1353
      _succ_num[u_in] = old_succ_num;
1291 1354

	
1292 1355
      // Set limits for updating _last_succ form v_in and v_out
1293 1356
      // towards the root
1294 1357
      int up_limit_in = -1;
1295 1358
      int up_limit_out = -1;
1296 1359
      if (_last_succ[join] == v_in) {
1297 1360
        up_limit_out = join;
1298 1361
      } else {
1299 1362
        up_limit_in = join;
1300 1363
      }
1301 1364

	
1302 1365
      // Update _last_succ from v_in towards the root
1303 1366
      for (u = v_in; u != up_limit_in && _last_succ[u] == v_in;
1304 1367
           u = _parent[u]) {
1305 1368
        _last_succ[u] = _last_succ[u_out];
1306 1369
      }
1307 1370
      // Update _last_succ from v_out towards the root
1308 1371
      if (join != old_rev_thread && v_in != old_rev_thread) {
1309 1372
        for (u = v_out; u != up_limit_out && _last_succ[u] == old_last_succ;
1310 1373
             u = _parent[u]) {
1311 1374
          _last_succ[u] = old_rev_thread;
1312 1375
        }
1313 1376
      } else {
1314 1377
        for (u = v_out; u != up_limit_out && _last_succ[u] == old_last_succ;
1315 1378
             u = _parent[u]) {
1316 1379
          _last_succ[u] = _last_succ[u_out];
1317 1380
        }
1318 1381
      }
1319 1382

	
1320 1383
      // Update _succ_num from v_in to join
1321 1384
      for (u = v_in; u != join; u = _parent[u]) {
1322 1385
        _succ_num[u] += old_succ_num;
1323 1386
      }
1324 1387
      // Update _succ_num from v_out to join
1325 1388
      for (u = v_out; u != join; u = _parent[u]) {
1326 1389
        _succ_num[u] -= old_succ_num;
1327 1390
      }
1328 1391
    }
1329 1392

	
1330 1393
    // Update potentials
1331 1394
    void updatePotential() {
1332 1395
      Cost sigma = _forward[u_in] ?
1333 1396
        _pi[v_in] - _pi[u_in] - _cost[_pred[u_in]] :
1334 1397
        _pi[v_in] - _pi[u_in] + _cost[_pred[u_in]];
1335 1398
      // Update potentials in the subtree, which has been moved
1336 1399
      int end = _thread[_last_succ[u_in]];
1337 1400
      for (int u = u_in; u != end; u = _thread[u]) {
1338 1401
        _pi[u] += sigma;
1339 1402
      }
1340 1403
    }
1341 1404

	
1342 1405
    // Execute the algorithm
1343 1406
    ProblemType start(PivotRule pivot_rule) {
1344 1407
      // Select the pivot rule implementation
1345 1408
      switch (pivot_rule) {
1346 1409
        case FIRST_ELIGIBLE:
1347 1410
          return start<FirstEligiblePivotRule>();
1348 1411
        case BEST_ELIGIBLE:
1349 1412
          return start<BestEligiblePivotRule>();
1350 1413
        case BLOCK_SEARCH:
1351 1414
          return start<BlockSearchPivotRule>();
1352 1415
        case CANDIDATE_LIST:
1353 1416
          return start<CandidateListPivotRule>();
1354 1417
        case ALTERING_LIST:
1355 1418
          return start<AlteringListPivotRule>();
1356 1419
      }
1357 1420
      return INFEASIBLE; // avoid warning
1358 1421
    }
1359 1422

	
1360 1423
    template <typename PivotRuleImpl>
1361 1424
    ProblemType start() {
1362 1425
      PivotRuleImpl pivot(*this);
1363 1426

	
1364 1427
      // Execute the Network Simplex algorithm
1365 1428
      while (pivot.findEnteringArc()) {
1366 1429
        findJoinNode();
1367 1430
        bool change = findLeavingArc();
1368 1431
        if (delta >= INF) return UNBOUNDED;
1369 1432
        changeFlow(change);
1370 1433
        if (change) {
1371 1434
          updateTreeStructure();
1372 1435
          updatePotential();
1373 1436
        }
1374 1437
      }
1375 1438
      
1376 1439
      // Check feasibility
1377
      if (_sum_supply < 0) {
1378
        for (int u = 0, e = _arc_num; u != _node_num; ++u, ++e) {
1379
          if (_supply[u] >= 0 && _flow[e] != 0) return INFEASIBLE;
1380
        }
1381
      }
1382
      else if (_sum_supply > 0) {
1383
        for (int u = 0, e = _arc_num; u != _node_num; ++u, ++e) {
1384
          if (_supply[u] <= 0 && _flow[e] != 0) return INFEASIBLE;
1385
        }
1386
      }
1387
      else {
1388
        for (int u = 0, e = _arc_num; u != _node_num; ++u, ++e) {
1389
          if (_flow[e] != 0) return INFEASIBLE;
1390
        }
1440
      for (int e = _search_arc_num; e != _all_arc_num; ++e) {
1441
        if (_flow[e] != 0) return INFEASIBLE;
1391 1442
      }
1392 1443

	
1393 1444
      // Transform the solution and the supply map to the original form
1394 1445
      if (_have_lower) {
1395 1446
        for (int i = 0; i != _arc_num; ++i) {
1396 1447
          Value c = _lower[i];
1397 1448
          if (c != 0) {
1398 1449
            _flow[i] += c;
1399 1450
            _supply[_source[i]] += c;
1400 1451
            _supply[_target[i]] -= c;
1401 1452
          }
1402 1453
        }
1403 1454
      }
1455
      
1456
      // Shift potentials to meet the requirements of the GEQ/LEQ type
1457
      // optimality conditions
1458
      if (_sum_supply == 0) {
1459
        if (_stype == GEQ) {
1460
          Cost max_pot = std::numeric_limits<Cost>::min();
1461
          for (int i = 0; i != _node_num; ++i) {
1462
            if (_pi[i] > max_pot) max_pot = _pi[i];
1463
          }
1464
          if (max_pot > 0) {
1465
            for (int i = 0; i != _node_num; ++i)
1466
              _pi[i] -= max_pot;
1467
          }
1468
        } else {
1469
          Cost min_pot = std::numeric_limits<Cost>::max();
1470
          for (int i = 0; i != _node_num; ++i) {
1471
            if (_pi[i] < min_pot) min_pot = _pi[i];
1472
          }
1473
          if (min_pot < 0) {
1474
            for (int i = 0; i != _node_num; ++i)
1475
              _pi[i] -= min_pot;
1476
          }
1477
        }
1478
      }
1404 1479

	
1405 1480
      return OPTIMAL;
1406 1481
    }
1407 1482

	
1408 1483
  }; //class NetworkSimplex
1409 1484

	
1410 1485
  ///@}
1411 1486

	
1412 1487
} //namespace lemon
1413 1488

	
1414 1489
#endif //LEMON_NETWORK_SIMPLEX_H
0 comments (0 inline)