1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
|
/* $Id: CbcHeuristicDivePseudoCost.cpp 2093 2014-11-06 16:17:38Z forrest $ */
// Copyright (C) 2008, International Business Machines
// Corporation and others. All Rights Reserved.
// This code is licensed under the terms of the Eclipse Public License (EPL).
#if defined(_MSC_VER)
// Turn off compiler warning about long names
# pragma warning(disable:4786)
#endif
#include "CbcHeuristicDivePseudoCost.hpp"
#include "CbcStrategy.hpp"
#include "CbcBranchDynamic.hpp"
// Default Constructor
CbcHeuristicDivePseudoCost::CbcHeuristicDivePseudoCost()
: CbcHeuristicDive()
{
}
// Constructor from model
CbcHeuristicDivePseudoCost::CbcHeuristicDivePseudoCost(CbcModel & model)
: CbcHeuristicDive(model)
{
}
// Destructor
CbcHeuristicDivePseudoCost::~CbcHeuristicDivePseudoCost ()
{
}
// Clone
CbcHeuristicDivePseudoCost *
CbcHeuristicDivePseudoCost::clone() const
{
return new CbcHeuristicDivePseudoCost(*this);
}
// Create C++ lines to get to current state
void
CbcHeuristicDivePseudoCost::generateCpp( FILE * fp)
{
CbcHeuristicDivePseudoCost other;
fprintf(fp, "0#include \"CbcHeuristicDivePseudoCost.hpp\"\n");
fprintf(fp, "3 CbcHeuristicDivePseudoCost heuristicDivePseudoCost(*cbcModel);\n");
CbcHeuristic::generateCpp(fp, "heuristicDivePseudoCost");
fprintf(fp, "3 cbcModel->addHeuristic(&heuristicDivePseudoCost);\n");
}
// Copy constructor
CbcHeuristicDivePseudoCost::CbcHeuristicDivePseudoCost(const CbcHeuristicDivePseudoCost & rhs)
:
CbcHeuristicDive(rhs)
{
}
// Assignment operator
CbcHeuristicDivePseudoCost &
CbcHeuristicDivePseudoCost::operator=( const CbcHeuristicDivePseudoCost & rhs)
{
if (this != &rhs) {
CbcHeuristicDive::operator=(rhs);
}
return *this;
}
bool
CbcHeuristicDivePseudoCost::selectVariableToBranch(OsiSolverInterface* solver,
const double* newSolution,
int& bestColumn,
int& bestRound)
{
int numberIntegers = model_->numberIntegers();
const int * integerVariable = model_->integerVariable();
double integerTolerance = model_->getDblParam(CbcModel::CbcIntegerTolerance);
// get the LP relaxation solution at the root node
double * rootNodeLPSol = model_->continuousSolution();
// get pseudo costs
double * pseudoCostDown = downArray_;
double * pseudoCostUp = upArray_;
bestColumn = -1;
bestRound = -1; // -1 rounds down, +1 rounds up
double bestScore = -1.0;
bool allTriviallyRoundableSoFar = true;
int bestPriority = COIN_INT_MAX;
for (int i = 0; i < numberIntegers; i++) {
int iColumn = integerVariable[i];
double rootValue = rootNodeLPSol[iColumn];
double value = newSolution[iColumn];
double fraction = value - floor(value);
int round = 0;
if (fabs(floor(value + 0.5) - value) > integerTolerance) {
if (allTriviallyRoundableSoFar || (downLocks_[i] > 0 && upLocks_[i] > 0)) {
if (allTriviallyRoundableSoFar && downLocks_[i] > 0 && upLocks_[i] > 0) {
allTriviallyRoundableSoFar = false;
bestScore = -1.0;
}
double pCostDown = pseudoCostDown[i];
double pCostUp = pseudoCostUp[i];
assert(pCostDown >= 0.0 && pCostUp >= 0.0);
if (allTriviallyRoundableSoFar && downLocks_[i] == 0 && upLocks_[i] > 0)
round = 1;
else if (allTriviallyRoundableSoFar && downLocks_[i] > 0 && upLocks_[i] == 0)
round = -1;
else if (value - rootValue < -0.4)
round = -1;
else if (value - rootValue > 0.4)
round = 1;
else if (fraction < 0.3)
round = -1;
else if (fraction > 0.7)
round = 1;
else if (pCostDown < pCostUp)
round = -1;
else
round = 1;
// calculate score
double score;
if (round == 1)
score = fraction * (pCostDown + 1.0) / (pCostUp + 1.0);
else
score = (1.0 - fraction) * (pCostUp + 1.0) / (pCostDown + 1.0);
// if variable is binary, increase its chance of being selected
if (solver->isBinary(iColumn))
score *= 1000.0;
// if priorities then use
if (priority_) {
int thisRound=static_cast<int>(priority_[i].direction);
if ((thisRound&1)!=0)
round = ((thisRound&2)==0) ? -1 : +1;
if (priority_[i].priority>bestPriority) {
score=COIN_DBL_MAX;
} else if (priority_[i].priority<bestPriority) {
bestPriority=static_cast<int>(priority_[i].priority);
bestScore=COIN_DBL_MAX;
}
}
if (score > bestScore) {
bestColumn = iColumn;
bestScore = score;
bestRound = round;
}
}
}
}
return allTriviallyRoundableSoFar;
}
void
CbcHeuristicDivePseudoCost::initializeData()
{
int numberIntegers = model_->numberIntegers();
if (!downArray_) {
downArray_ = new double [numberIntegers];
upArray_ = new double [numberIntegers];
}
// get pseudo costs
model_->fillPseudoCosts(downArray_, upArray_);
// allow for -999 -> force to run
int diveOptions = (when_>0) ? when_ / 100 : 0;
if (diveOptions) {
// pseudo shadow prices
int k = diveOptions % 100;
if (diveOptions >= 100)
k += 32;
model_->pseudoShadow(k - 1);
int numberInts = CoinMin(model_->numberObjects(), numberIntegers);
OsiObject ** objects = model_->objects();
for (int i = 0; i < numberInts; i++) {
CbcSimpleIntegerDynamicPseudoCost * obj1 =
dynamic_cast <CbcSimpleIntegerDynamicPseudoCost *>(objects[i]) ;
if (obj1) {
//int iColumn = obj1->columnNumber();
double downPseudoCost = 1.0e-2 * obj1->downDynamicPseudoCost();
double downShadow = obj1->downShadowPrice();
double upPseudoCost = 1.0e-2 * obj1->upDynamicPseudoCost();
double upShadow = obj1->upShadowPrice();
downPseudoCost = CoinMax(downPseudoCost, downShadow);
downPseudoCost = CoinMax(downPseudoCost, 0.001 * upShadow);
downArray_[i] = downPseudoCost;
upPseudoCost = CoinMax(upPseudoCost, upShadow);
upPseudoCost = CoinMax(upPseudoCost, 0.001 * downShadow);
upArray_[i] = upPseudoCost;
}
}
}
}
// Fix other variables at bounds
int
CbcHeuristicDivePseudoCost::fixOtherVariables(OsiSolverInterface * solver,
const double * solution,
PseudoReducedCost * candidate,
const double * random)
{
const double * lower = solver->getColLower();
const double * upper = solver->getColUpper();
double integerTolerance = model_->getDblParam(CbcModel::CbcIntegerTolerance);
double primalTolerance;
solver->getDblParam(OsiPrimalTolerance, primalTolerance);
int numberIntegers = model_->numberIntegers();
const int * integerVariable = model_->integerVariable();
const double* reducedCost = solver->getReducedCost();
bool fixGeneralIntegers = (switches_&65536)!=0;
// fix other integer variables that are at their bounds
int cnt = 0;
int numberFree = 0;
int numberFixedAlready = 0;
for (int i = 0; i < numberIntegers; i++) {
int iColumn = integerVariable[i];
if (upper[iColumn] > lower[iColumn]) {
numberFree++;
double value = solution[iColumn];
if (value - lower[iColumn] <= integerTolerance) {
candidate[cnt].var = iColumn;
candidate[cnt++].pseudoRedCost = CoinMax(1.0e-2 * reducedCost[iColumn],
downArray_[i]) * random[i];
} else if (upper[iColumn] - value <= integerTolerance) {
candidate[cnt].var = iColumn;
candidate[cnt++].pseudoRedCost = CoinMax(-1.0e-2 * reducedCost[iColumn],
downArray_[i]) * random[i];
} else if (fixGeneralIntegers &&
fabs(floor(value + 0.5) - value) <= integerTolerance) {
candidate[cnt].var = iColumn;
candidate[cnt++].pseudoRedCost = CoinMax(-1.0e-6 * reducedCost[iColumn],
1.0e-4*downArray_[i]) * random[i];
}
} else {
numberFixedAlready++;
}
}
#ifdef CLP_INVESTIGATE
printf("cutoff %g obj %g - %d free, %d fixed\n",
model_->getCutoff(), solver->getObjValue(), numberFree,
numberFixedAlready);
#endif
return cnt;
//return CbcHeuristicDive::fixOtherVariables(solver, solution,
// candidate, random);
}
|
