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
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
|
// $Id: CbcHeuristicDW.hpp 1899 2013-04-09 18:12:08Z stefan $
// Copyright (C) 2006, International Business Machines
// Corporation and others. All Rights Reserved.
// This code is licensed under the terms of the Eclipse Public License (EPL).
#ifndef CbcHeuristicDW_H
#define CbcHeuristicDW_H
#include "CbcHeuristic.hpp"
/**
This is unlike the other heuristics in that it is very very compute intensive.
It tries to find a DW structure and use that
*/
class CbcHeuristicDW : public CbcHeuristic {
public:
// Default Constructor
CbcHeuristicDW ();
/* Constructor with model - assumed before cuts
*/
CbcHeuristicDW (CbcModel & model, int keepContinuous=0);
/* Constructor with model - assumed before cuts
*/
CbcHeuristicDW (CbcModel & model,
int callBack(CbcHeuristicDW * currentHeuristic,
CbcModel * thisModel,
int whereFrom),
int keepContinuous=0);
// Copy constructor
CbcHeuristicDW ( const CbcHeuristicDW &);
// Destructor
~CbcHeuristicDW ();
/// Clone
virtual CbcHeuristic * clone() const;
/// Assignment operator
CbcHeuristicDW & operator=(const CbcHeuristicDW& rhs);
/// Create C++ lines to get to current state
virtual void generateCpp( FILE * fp) ;
/// Resets stuff if model changes
virtual void resetModel(CbcModel * model);
/// update model (This is needed if cliques update matrix etc)
virtual void setModel(CbcModel * model);
using CbcHeuristic::solution ;
/** returns 0 if no solution, 1 if valid solution.
Sets solution values if good, sets objective value (only if good)
This does Relaxation Induced Neighborhood Search
*/
virtual int solution(double & objectiveValue,
double * newSolution);
/** Return number of blocks
<=0 - no usable structure */
inline int numberBlocks() const
{ return numberBlocks_;}
/// Pass in a solution
void passInSolution(const double * solution);
/// Pass in continuous solution
void passInContinuousSolution(const double * solution);
/** DW Proposal actions
fullDWEverySoOften -
0 - off
k - every k times solution gets better
*/
void setProposalActions(int fullDWEverySoOften);
/// Objective value when whichDw created
double objectiveValueWhen(int whichDW) const;
/// Number of columns in DW
int numberColumnsDW(int whichDW) const;
/// Solver
inline OsiSolverInterface * solver() const
{ return solver_;}
/// DW model (user must delete)
OsiSolverInterface * DWModel(int whichDW) const;
/// Best objective value
inline double bestObjective() const
{ return bestObjective_;}
/// Best solution found so far
inline const double * bestSolution() const
{ return bestSolution_;}
/// Continuous solution
inline const double * continuousSolution() const
{ return continuousSolution_;}
/// Reduced costs of fixed solution
inline const double * fixedDj() const
{ return fixedDj_;}
/// Objective at which DW updated
inline const double * objectiveDW() const
{ return objectiveDW_;}
/// Number of times we have added to DW model
inline int numberDWTimes() const
{ return numberDWTimes_;}
/// Number of columns in DW
inline const int * numberColumnsDW() const
{ return numberColumnsDW_;}
/// Set number of passes
inline void setNumberPasses(int value)
{ numberPasses_ = value;}
/// Set number of passes without better solution
inline void setNumberBadPasses(int value)
{ numberBadPasses_ = value;}
/// Set number free integers needed (Base value)
inline void setNumberNeeded(int value)
{ nNeededBase_ = value;}
/// Get number free integers needed (Base value)
inline int getNumberNeeded() const
{return nNeededBase_;}
/// Set number free integers needed (Current value)
inline void setCurrentNumberNeeded(int value)
{ nNeeded_ = value;}
/// Get number free integers needed (Current value)
inline int getCurrentNumberNeeded() const
{return nNeeded_;}
/// Set number nodes (could be done in callback) (Base value)
inline void setNumberNodes(int value)
{ nNodesBase_ = value;}
/// Get number nodes (could be done in callback) (Base value)
inline int getNumberNodes() const
{return nNodesBase_;}
/// Set number nodes (could be done in callback) (Current value)
inline void setCurrentNumberNodes(int value)
{ nNodes_ = value;}
/// Get number nodes (could be done in callback) (Current value)
inline int getCurrentNumberNodes() const
{return nNodes_;}
/// Set target objective
inline void setTargetObjective(double value)
{ targetObjective_ = value;}
/// Sets how often to do it
inline void setHowOften(int value) {
howOften_ = value;
}
/// Block for every row
inline const int * whichRowBlock() const
{ return whichRowBlock_;}
/// Block for every column
inline const int * whichColumnBlock() const
{ return whichColumnBlock_;}
/// Initial Lower bounds
inline double * initialLower() const
{ return saveLower_;}
/// Initial Upper bounds
inline double * initialUpper() const
{ return saveUpper_;}
/// Local integer arrays (each numberBlocks_ long)
inline int * intArrays() const
{ return intArray_;}
/// Local double arrays (each numberBlocks_ long)
inline double * doubleArrays() const
{ return doubleArray_;}
/// Phase of solution
inline int phase() const
{ return phase_;}
/// Pass number
inline int pass() const
{ return pass_;}
/// Which columns are in block
inline const int * columnsInBlock() const
{ return columnsInBlock_;}
/// Starts for columnsInBlock
inline const int * startColumnBlock() const
{ return startColumnBlock_;}
/// Number of integer variables in each block
inline const int * intsInBlock() const
{ return intsInBlock_;}
/// Objective value (could also check validity)
double objectiveValue(const double * solution);
private:
/// Guts of copy
void gutsOfCopy(const CbcHeuristicDW & rhs);
/// Guts of delete
void gutsOfDelete();
/// Set default values
void setDefaults();
/// Find structure
void findStructure();
/// Set up DW structure
void setupDWStructures();
/// Add DW proposals
int addDW(const double * solution,int numberBlocksUsed,
const int * whichBlocks);
protected:
typedef int (*heuristicCallBack) (CbcHeuristicDW * ,CbcModel *, int) ;
// Data
/// Target objective
double targetObjective_;
/// Best objective value
double bestObjective_;
/// Objective value last time
double lastObjective_;
/** Call back
whereFrom -
0 - after blocks found but before data setup
1 - after blocks sorted but before used
2 - just before normal branch and bound
3 - after DW has been updated
4 - if better solution found
5 - every time a block might be used
next few for adjustment of nNeeded etc
6 - complete search done - no solution
7 - stopped on nodes - no improvement
8 - improving (same as 4 but after nNeeded changed
Pointers to local data given by following pointers
*/
heuristicCallBack functionPointer_;
/// Local integer arrays (each numberBlocks_ long)
int * intArray_;
/// Local double arrays (each numberBlocks_ long)
double * doubleArray_;
/// Base solver
OsiSolverInterface * solver_;
/// DW solver
OsiSolverInterface * dwSolver_;
/// Best solution found so far
double * bestSolution_;
/// Continuous solution
double * continuousSolution_;
/// Reduced costs of fixed solution
double * fixedDj_;
/// Original lower bounds
double * saveLower_;
/// Original Upper bounds
double * saveUpper_;
/// random numbers for master rows
double * random_;
/// Weights for each proposal
double * weights_;
/// Objective at which DW updated
double * objectiveDW_;
/// Number of columns in each DW
int * numberColumnsDW_;
/// Block for every row
int * whichRowBlock_;
/// Block for every column
int * whichColumnBlock_;
/// Block number for each proposal
int * dwBlock_;
/// Points back to master rows
int * backwardRow_;
/// Which rows are in blocke
int * rowsInBlock_;
/// Which columns are in block
int * columnsInBlock_;
/// Starts for rowsInBlock
int * startRowBlock_;
/// Starts for columnsInBlock
int * startColumnBlock_;
/// Number of integer variables in each block
int * intsInBlock_;
/// Bits set for 1 integers in each block
unsigned int * fingerPrint_;
/// Affinity each block has for other (will be triangular?)
unsigned short * affinity_;
/** DW Proposal actions
fullDWEverySoOften -
0 - off
k - every k times solution gets better
*/
int fullDWEverySoOften_;
/// Number of passes
int numberPasses_;
/// How often to do (code can change)
int howOften_;
/// Current maximum number of DW proposals
int maximumDW_;
/// Number of DW proposals
int numberDW_;
/// Number of times we have added to DW model
int numberDWTimes_;
/// Number of unsigned ints needed for each block of fingerPrint
int sizeFingerPrint_;
/// Number of columns in master
int numberMasterColumns_;
/// Number of rows in master
int numberMasterRows_;
/// Number of blocks
int numberBlocks_;
/// Action on decomposition - 1 keep continuous, 0 don't
int keepContinuous_;
/// Phase of solution
int phase_;
/// Pass number
int pass_;
/// Base number of integers needed
int nNeededBase_;
/// Base number of nodes needed
int nNodesBase_;
/// Base number of integers needed
int nNeeded_;
/// Base number of nodes needed
int nNodes_;
/// Number of passes without better solution
int numberBadPasses_;
// 0 - fine, 1 can't be better, 2 max node
int solveState_;
};
#endif
|