... | ... |
@@ -124,10 +124,13 @@ |
124 | 124 |
/// |
125 |
/// \ref NetworkSimplex provides five different pivot rule |
|
126 |
/// implementations that significantly affect the running time |
|
125 |
/// \ref NetworkSimplex provides five different implementations for |
|
126 |
/// the pivot strategy that significantly affects the running time |
|
127 | 127 |
/// of the algorithm. |
128 |
/// By default, \ref BLOCK_SEARCH "Block Search" is used, which |
|
129 |
/// turend out to be the most efficient and the most robust on various |
|
130 |
/// test inputs. |
|
131 |
/// However, another pivot rule can be selected using the \ref run() |
|
132 |
/// |
|
128 |
/// According to experimental tests conducted on various problem |
|
129 |
/// instances, \ref BLOCK_SEARCH "Block Search" and |
|
130 |
/// \ref ALTERING_LIST "Altering Candidate List" rules turned out |
|
131 |
/// to be the most efficient. |
|
132 |
/// Since \ref BLOCK_SEARCH "Block Search" is a simpler strategy that |
|
133 |
/// seemed to be slightly more robust, it is used by default. |
|
134 |
/// However, another pivot rule can easily be selected using the |
|
135 |
/// \ref run() function with the proper parameter. |
|
133 | 136 |
enum PivotRule { |
... | ... |
@@ -157,3 +160,3 @@ |
157 | 160 |
/// It is a modified version of the Candidate List method. |
158 |
/// It keeps only |
|
161 |
/// It keeps only a few of the best eligible arcs from the former |
|
159 | 162 |
/// candidate list and extends this list in every iteration. |
... | ... |
@@ -540,3 +543,3 @@ |
540 | 543 |
bool operator()(int left, int right) { |
541 |
return _map[left] |
|
544 |
return _map[left] < _map[right]; |
|
542 | 545 |
} |
... | ... |
@@ -558,3 +561,3 @@ |
558 | 561 |
const int MIN_BLOCK_SIZE = 10; |
559 |
const double HEAD_LENGTH_FACTOR = 0. |
|
562 |
const double HEAD_LENGTH_FACTOR = 0.01; |
|
560 | 563 |
const int MIN_HEAD_LENGTH = 3; |
... | ... |
@@ -602,5 +605,5 @@ |
602 | 605 |
for (e = 0; e != _next_arc; ++e) { |
603 |
_cand_cost[e] = _state[e] * |
|
604 |
(_cost[e] + _pi[_source[e]] - _pi[_target[e]]); |
|
605 |
|
|
606 |
c = _state[e] * (_cost[e] + _pi[_source[e]] - _pi[_target[e]]); |
|
607 |
if (c < 0) { |
|
608 |
_cand_cost[e] = c; |
|
606 | 609 |
_candidates[_curr_length++] = e; |
... | ... |
@@ -617,12 +620,12 @@ |
617 | 620 |
|
618 |
// Make heap of the candidate list (approximating a partial sort) |
|
619 |
make_heap( _candidates.begin(), _candidates.begin() + _curr_length, |
|
620 |
|
|
621 |
// Perform partial sort operation on the candidate list |
|
622 |
int new_length = std::min(_head_length + 1, _curr_length); |
|
623 |
std::partial_sort(_candidates.begin(), _candidates.begin() + new_length, |
|
624 |
_candidates.begin() + _curr_length, _sort_func); |
|
621 | 625 |
|
622 |
// |
|
626 |
// Select the entering arc and remove it from the list |
|
623 | 627 |
_in_arc = _candidates[0]; |
624 | 628 |
_next_arc = e; |
625 |
pop_heap( _candidates.begin(), _candidates.begin() + _curr_length, |
|
626 |
_sort_func ); |
|
627 |
|
|
629 |
_candidates[0] = _candidates[new_length - 1]; |
|
630 |
_curr_length = new_length - 1; |
|
628 | 631 |
return true; |
0 comments (0 inline)