gravatar
kpeter (Peter Kovacs)
kpeter@inf.elte.hu
Various search limits for the max clique alg (#405)
0 2 0
default
2 files changed with 225 insertions and 53 deletions:
↑ Collapse diff ↑
Show white space 12 line context
... ...
@@ -43,14 +43,18 @@
43 43
  /// \e clique \e problem \ref grosso08maxclique.
44 44
  /// It is to find the largest complete subgraph (\e clique) in an
45 45
  /// undirected graph, i.e., the largest set of nodes where each
46 46
  /// pair of nodes is connected.
47 47
  ///
48 48
  /// This class provides a simple but highly efficient and robust heuristic
49
  /// method that quickly finds a large clique, but not necessarily the
49
  /// method that quickly finds a quite large clique, but not necessarily the
50 50
  /// largest one.
51
  /// The algorithm performs a certain number of iterations to find several
52
  /// cliques and selects the largest one among them. Various limits can be
53
  /// specified to control the running time and the effectiveness of the
54
  /// search process.
51 55
  ///
52 56
  /// \tparam GR The undirected graph type the algorithm runs on.
53 57
  ///
54 58
  /// \note %GrossoLocatelliPullanMc provides three different node selection
55 59
  /// rules, from which the most powerful one is used by default.
56 60
  /// For more information, see \ref SelectionRule.
... ...
@@ -81,28 +85,54 @@
81 85
      /// A node of minimum penalty is selected randomly at each step.
82 86
      /// The node penalties are updated adaptively after each stage of the
83 87
      /// search process.
84 88
      PENALTY_BASED
85 89
    };
86 90

	
91
    /// \brief Constants for the causes of search termination.
92
    ///
93
    /// Enum type containing constants for the different causes of search
94
    /// termination. The \ref run() function returns one of these values.
95
    enum TerminationCause {
96

	
97
      /// The iteration count limit is reached.
98
      ITERATION_LIMIT,
99

	
100
      /// The step count limit is reached.
101
      STEP_LIMIT,
102

	
103
      /// The clique size limit is reached.
104
      SIZE_LIMIT
105
    };
106

	
87 107
  private:
88 108

	
89 109
    TEMPLATE_GRAPH_TYPEDEFS(GR);
90 110

	
91 111
    typedef std::vector<int> IntVector;
92 112
    typedef std::vector<char> BoolVector;
93 113
    typedef std::vector<BoolVector> BoolMatrix;
94 114
    // Note: vector<char> is used instead of vector<bool> for efficiency reasons
95 115

	
116
    // The underlying graph
96 117
    const GR &_graph;
97 118
    IntNodeMap _id;
98 119

	
99 120
    // Internal matrix representation of the graph
100 121
    BoolMatrix _gr;
101 122
    int _n;
102 123

	
124
    // Search options
125
    bool _delta_based_restart;
126
    int _restart_delta_limit;
127
 
128
    // Search limits
129
    int _iteration_limit;
130
    int _step_limit;
131
    int _size_limit;
132

	
103 133
    // The current clique
104 134
    BoolVector _clique;
105 135
    int _size;
106 136

	
107 137
    // The best clique found so far
108 138
    BoolVector _best_clique;
... ...
@@ -377,71 +407,187 @@
377 407
    /// The global \ref rnd "random number generator instance" is used
378 408
    /// during the algorithm.
379 409
    ///
380 410
    /// \param graph The undirected graph the algorithm runs on.
381 411
    GrossoLocatelliPullanMc(const GR& graph) :
382 412
      _graph(graph), _id(_graph), _rnd(rnd)
383
    {}
413
    {
414
      initOptions();
415
    }
384 416

	
385 417
    /// \brief Constructor with random seed.
386 418
    ///
387 419
    /// Constructor with random seed.
388 420
    ///
389 421
    /// \param graph The undirected graph the algorithm runs on.
390 422
    /// \param seed Seed value for the internal random number generator
391 423
    /// that is used during the algorithm.
392 424
    GrossoLocatelliPullanMc(const GR& graph, int seed) :
393 425
      _graph(graph), _id(_graph), _rnd(seed)
394
    {}
426
    {
427
      initOptions();
428
    }
395 429

	
396 430
    /// \brief Constructor with random number generator.
397 431
    ///
398 432
    /// Constructor with random number generator.
399 433
    ///
400 434
    /// \param graph The undirected graph the algorithm runs on.
401 435
    /// \param random A random number generator that is used during the
402 436
    /// algorithm.
403 437
    GrossoLocatelliPullanMc(const GR& graph, const Random& random) :
404 438
      _graph(graph), _id(_graph), _rnd(random)
405
    {}
439
    {
440
      initOptions();
441
    }
406 442

	
407 443
    /// \name Execution Control
444
    /// The \ref run() function can be used to execute the algorithm.\n
445
    /// The functions \ref iterationLimit(int), \ref stepLimit(int), and 
446
    /// \ref sizeLimit(int) can be used to specify various limits for the
447
    /// search process.
448
    
408 449
    /// @{
409 450

	
451
    /// \brief Sets the maximum number of iterations.
452
    ///
453
    /// This function sets the maximum number of iterations.
454
    /// Each iteration of the algorithm finds a maximal clique (but not
455
    /// necessarily the largest one) by performing several search steps
456
    /// (node selections).
457
    ///
458
    /// This limit controls the running time and the success of the
459
    /// algorithm. For larger values, the algorithm runs slower, but it more
460
    /// likely finds larger cliques. For smaller values, the algorithm is
461
    /// faster but probably gives worse results.
462
    /// 
463
    /// The default value is \c 1000.
464
    /// \c -1 means that number of iterations is not limited.
465
    ///
466
    /// \warning You should specify a reasonable limit for the number of
467
    /// iterations and/or the number of search steps.
468
    ///
469
    /// \return <tt>(*this)</tt>
470
    ///
471
    /// \sa stepLimit(int)
472
    /// \sa sizeLimit(int)
473
    GrossoLocatelliPullanMc& iterationLimit(int limit) {
474
      _iteration_limit = limit;
475
      return *this;
476
    }
477
    
478
    /// \brief Sets the maximum number of search steps.
479
    ///
480
    /// This function sets the maximum number of elementary search steps.
481
    /// Each iteration of the algorithm finds a maximal clique (but not
482
    /// necessarily the largest one) by performing several search steps
483
    /// (node selections).
484
    ///
485
    /// This limit controls the running time and the success of the
486
    /// algorithm. For larger values, the algorithm runs slower, but it more
487
    /// likely finds larger cliques. For smaller values, the algorithm is
488
    /// faster but probably gives worse results.
489
    /// 
490
    /// The default value is \c -1, which means that number of steps
491
    /// is not limited explicitly. However, the number of iterations is
492
    /// limited and each iteration performs a finite number of search steps.
493
    ///
494
    /// \warning You should specify a reasonable limit for the number of
495
    /// iterations and/or the number of search steps.
496
    ///
497
    /// \return <tt>(*this)</tt>
498
    ///
499
    /// \sa iterationLimit(int)
500
    /// \sa sizeLimit(int)
501
    GrossoLocatelliPullanMc& stepLimit(int limit) {
502
      _step_limit = limit;
503
      return *this;
504
    }
505
    
506
    /// \brief Sets the desired clique size.
507
    ///
508
    /// This function sets the desired clique size that serves as a search
509
    /// limit. If a clique of this size (or a larger one) is found, then the
510
    /// algorithm terminates.
511
    /// 
512
    /// This function is especially useful if you know an exact upper bound
513
    /// for the size of the cliques in the graph or if any clique above 
514
    /// a certain size limit is sufficient for your application.
515
    /// 
516
    /// The default value is \c -1, which means that the size limit is set to
517
    /// the number of nodes in the graph.
518
    ///
519
    /// \return <tt>(*this)</tt>
520
    ///
521
    /// \sa iterationLimit(int)
522
    /// \sa stepLimit(int)
523
    GrossoLocatelliPullanMc& sizeLimit(int limit) {
524
      _size_limit = limit;
525
      return *this;
526
    }
527
    
528
    /// \brief The maximum number of iterations.
529
    ///
530
    /// This function gives back the maximum number of iterations.
531
    /// \c -1 means that no limit is specified.
532
    ///
533
    /// \sa iterationLimit(int)
534
    int iterationLimit() const {
535
      return _iteration_limit;
536
    }
537
    
538
    /// \brief The maximum number of search steps.
539
    ///
540
    /// This function gives back the maximum number of search steps.
541
    /// \c -1 means that no limit is specified.
542
    ///
543
    /// \sa stepLimit(int)
544
    int stepLimit() const {
545
      return _step_limit;
546
    }
547
    
548
    /// \brief The desired clique size.
549
    ///
550
    /// This function gives back the desired clique size that serves as a
551
    /// search limit. \c -1 means that this limit is set to the number of
552
    /// nodes in the graph.
553
    ///
554
    /// \sa sizeLimit(int)
555
    int sizeLimit() const {
556
      return _size_limit;
557
    }
558

	
410 559
    /// \brief Runs the algorithm.
411 560
    ///
412
    /// This function runs the algorithm.
561
    /// This function runs the algorithm. If one of the specified limits
562
    /// is reached, the search process terminates.
413 563
    ///
414
    /// \param step_num The maximum number of node selections (steps)
415
    /// during the search process.
416
    /// This parameter controls the running time and the success of the
417
    /// algorithm. For larger values, the algorithm runs slower but it more
418
    /// likely finds larger cliques. For smaller values, the algorithm is
419
    /// faster but probably gives worse results.
420 564
    /// \param rule The node selection rule. For more information, see
421 565
    /// \ref SelectionRule.
422 566
    ///
423
    /// \return The size of the found clique.
424
    int run(int step_num = 100000,
425
            SelectionRule rule = PENALTY_BASED)
567
    /// \return The termination cause of the search. For more information,
568
    /// see \ref TerminationCause.
569
    TerminationCause run(SelectionRule rule = PENALTY_BASED)
426 570
    {
427 571
      init();
428 572
      switch (rule) {
429 573
        case RANDOM:
430
          return start<RandomSelectionRule>(step_num);
574
          return start<RandomSelectionRule>();
431 575
        case DEGREE_BASED:
432
          return start<DegreeBasedSelectionRule>(step_num);
433
        case PENALTY_BASED:
434
          return start<PenaltyBasedSelectionRule>(step_num);
576
          return start<DegreeBasedSelectionRule>();
577
        default:
578
          return start<PenaltyBasedSelectionRule>();
435 579
      }
436
      return 0; // avoid warning
437 580
    }
438 581

	
439 582
    /// @}
440 583

	
441 584
    /// \name Query Functions
585
    /// The results of the algorithm can be obtained using these functions.\n
586
    /// The run() function must be called before using them. 
587

	
442 588
    /// @{
443 589

	
444 590
    /// \brief The size of the found clique
445 591
    ///
446 592
    /// This function returns the size of the found clique.
447 593
    ///
... ...
@@ -528,12 +674,24 @@
528 674
    };
529 675

	
530 676
    /// @}
531 677

	
532 678
  private:
533 679

	
680
    // Initialize search options and limits
681
    void initOptions() {
682
      // Search options
683
      _delta_based_restart = true;
684
      _restart_delta_limit = 4;
685
     
686
      // Search limits
687
      _iteration_limit = 1000;
688
      _step_limit = -1;             // this is disabled by default
689
      _size_limit = -1;             // this is disabled by default
690
    }
691

	
534 692
    // Adds a node to the current clique
535 693
    void addCliqueNode(int u) {
536 694
      if (_clique[u]) return;
537 695
      _clique[u] = true;
538 696
      _size++;
539 697
      BoolVector &row = _gr[u];
... ...
@@ -583,36 +741,38 @@
583 741
      _tabu.clear();
584 742
      _tabu.resize(_n, false);
585 743
    }
586 744

	
587 745
    // Executes the algorithm
588 746
    template <typename SelectionRuleImpl>
589
    int start(int max_select) {
590
      // Options for the restart rule
591
      const bool delta_based_restart = true;
592
      const int restart_delta_limit = 4;
593

	
594
      if (_n == 0) return 0;
747
    TerminationCause start() {
748
      if (_n == 0) return SIZE_LIMIT;
595 749
      if (_n == 1) {
596 750
        _best_clique[0] = true;
597 751
        _best_size = 1;
598
        return _best_size;
752
        return SIZE_LIMIT;
599 753
      }
600 754

	
601
      // Iterated local search
755
      // Iterated local search algorithm
756
      const int max_size = _size_limit >= 0 ? _size_limit : _n;
757
      const int max_restart = _iteration_limit >= 0 ?
758
        _iteration_limit : std::numeric_limits<int>::max();
759
      const int max_select = _step_limit >= 0 ?
760
        _step_limit : std::numeric_limits<int>::max();
761

	
602 762
      SelectionRuleImpl sel_method(*this);
603
      int select = 0;
763
      int select = 0, restart = 0;
604 764
      IntVector restart_nodes;
605

	
606
      while (select < max_select) {
765
      while (select < max_select && restart < max_restart) {
607 766

	
608 767
        // Perturbation/restart
609
        if (delta_based_restart) {
768
        restart++;
769
        if (_delta_based_restart) {
610 770
          restart_nodes.clear();
611 771
          for (int i = 0; i != _n; i++) {
612
            if (_delta[i] >= restart_delta_limit)
772
            if (_delta[i] >= _restart_delta_limit)
613 773
              restart_nodes.push_back(i);
614 774
          }
615 775
        }
616 776
        int rs_node = -1;
617 777
        if (restart_nodes.size() > 0) {
618 778
          rs_node = restart_nodes[_rnd[restart_nodes.size()]];
... ...
@@ -660,18 +820,18 @@
660 820
          }
661 821
          else break;
662 822
        }
663 823
        if (_size > _best_size) {
664 824
          _best_clique = _clique;
665 825
          _best_size = _size;
666
          if (_best_size == _n) return _best_size;
826
          if (_best_size >= max_size) return SIZE_LIMIT;
667 827
        }
668 828
        sel_method.update();
669 829
      }
670 830

	
671
      return _best_size;
831
      return (restart >= max_restart ? ITERATION_LIMIT : STEP_LIMIT);
672 832
    }
673 833

	
674 834
  }; //class GrossoLocatelliPullanMc
675 835

	
676 836
  ///@}
677 837

	
Show white space 12 line context
... ...
@@ -55,45 +55,46 @@
55 55
  "5 7    14\n"
56 56
  "6 7    15\n";
57 57
      
58 58

	
59 59
// Check with general graphs
60 60
template <typename Param>
61
void checkMaxCliqueGeneral(int max_sel, Param rule) {
61
void checkMaxCliqueGeneral(Param rule) {
62 62
  typedef ListGraph GR;
63 63
  typedef GrossoLocatelliPullanMc<GR> McAlg;
64 64
  typedef McAlg::CliqueNodeIt CliqueIt;
65 65
  
66 66
  // Basic tests
67 67
  {
68 68
    GR g;
69 69
    GR::NodeMap<bool> map(g);
70 70
    McAlg mc(g);
71
    check(mc.run(max_sel, rule) == 0, "Wrong clique size");
71
    mc.iterationLimit(50);
72
    check(mc.run(rule) == McAlg::SIZE_LIMIT, "Wrong termination cause");
72 73
    check(mc.cliqueSize() == 0, "Wrong clique size");
73 74
    check(CliqueIt(mc) == INVALID, "Wrong CliqueNodeIt");
74 75

	
75 76
    GR::Node u = g.addNode();
76
    check(mc.run(max_sel, rule) == 1, "Wrong clique size");
77
    check(mc.run(rule) == McAlg::SIZE_LIMIT, "Wrong termination cause");
77 78
    check(mc.cliqueSize() == 1, "Wrong clique size");
78 79
    mc.cliqueMap(map);
79 80
    check(map[u], "Wrong clique map");
80 81
    CliqueIt it1(mc);
81 82
    check(static_cast<GR::Node>(it1) == u && ++it1 == INVALID,
82 83
          "Wrong CliqueNodeIt");
83 84
    
84 85
    GR::Node v = g.addNode();
85
    check(mc.run(max_sel, rule) == 1, "Wrong clique size");
86
    check(mc.run(rule) == McAlg::ITERATION_LIMIT, "Wrong termination cause");
86 87
    check(mc.cliqueSize() == 1, "Wrong clique size");
87 88
    mc.cliqueMap(map);
88 89
    check((map[u] && !map[v]) || (map[v] && !map[u]), "Wrong clique map");
89 90
    CliqueIt it2(mc);
90 91
    check(it2 != INVALID && ++it2 == INVALID, "Wrong CliqueNodeIt");
91 92

	
92 93
    g.addEdge(u, v);
93
    check(mc.run(max_sel, rule) == 2, "Wrong clique size");
94
    check(mc.run(rule) == McAlg::SIZE_LIMIT, "Wrong termination cause");
94 95
    check(mc.cliqueSize() == 2, "Wrong clique size");
95 96
    mc.cliqueMap(map);
96 97
    check(map[u] && map[v], "Wrong clique map");
97 98
    CliqueIt it3(mc);
98 99
    check(it3 != INVALID && ++it3 != INVALID && ++it3 == INVALID,
99 100
          "Wrong CliqueNodeIt");
... ...
@@ -107,13 +108,14 @@
107 108
    std::istringstream input(test_lgf);
108 109
    graphReader(g, input)
109 110
      .nodeMap("max_clique", max_clique)
110 111
      .run();
111 112
    
112 113
    McAlg mc(g);
113
    check(mc.run(max_sel, rule) == 4, "Wrong clique size");
114
    mc.iterationLimit(50);
115
    check(mc.run(rule) == McAlg::ITERATION_LIMIT, "Wrong termination cause");
114 116
    check(mc.cliqueSize() == 4, "Wrong clique size");
115 117
    mc.cliqueMap(map);
116 118
    for (GR::NodeIt n(g); n != INVALID; ++n) {
117 119
      check(map[n] == max_clique[n], "Wrong clique map");
118 120
    }
119 121
    int cnt = 0;
... ...
@@ -124,22 +126,22 @@
124 126
    check(cnt == 4, "Wrong CliqueNodeIt");
125 127
  }
126 128
}
127 129

	
128 130
// Check with full graphs
129 131
template <typename Param>
130
void checkMaxCliqueFullGraph(int max_sel, Param rule) {
132
void checkMaxCliqueFullGraph(Param rule) {
131 133
  typedef FullGraph GR;
132 134
  typedef GrossoLocatelliPullanMc<FullGraph> McAlg;
133 135
  typedef McAlg::CliqueNodeIt CliqueIt;
134 136
  
135 137
  for (int size = 0; size <= 40; size = size * 3 + 1) {
136 138
    GR g(size);
137 139
    GR::NodeMap<bool> map(g);
138 140
    McAlg mc(g);
139
    check(mc.run(max_sel, rule) == size, "Wrong clique size");
141
    check(mc.run(rule) == McAlg::SIZE_LIMIT, "Wrong termination cause");
140 142
    check(mc.cliqueSize() == size, "Wrong clique size");
141 143
    mc.cliqueMap(map);
142 144
    for (GR::NodeIt n(g); n != INVALID; ++n) {
143 145
      check(map[n], "Wrong clique map");
144 146
    }
145 147
    int cnt = 0;
... ...
@@ -147,30 +149,40 @@
147 149
    check(cnt == size, "Wrong CliqueNodeIt");
148 150
  }
149 151
}
150 152

	
151 153
// Check with grid graphs
152 154
template <typename Param>
153
void checkMaxCliqueGridGraph(int max_sel, Param rule) {
155
void checkMaxCliqueGridGraph(Param rule) {
154 156
  GridGraph g(5, 7);
155 157
  GridGraph::NodeMap<char> map(g);
156 158
  GrossoLocatelliPullanMc<GridGraph> mc(g);
157
  check(mc.run(max_sel, rule) == 2, "Wrong clique size");
159
  
160
  mc.iterationLimit(100);
161
  check(mc.run(rule) == mc.ITERATION_LIMIT, "Wrong termination cause");
162
  check(mc.cliqueSize() == 2, "Wrong clique size");
163

	
164
  mc.stepLimit(100);
165
  check(mc.run(rule) == mc.STEP_LIMIT, "Wrong termination cause");
166
  check(mc.cliqueSize() == 2, "Wrong clique size");
167

	
168
  mc.sizeLimit(2);
169
  check(mc.run(rule) == mc.SIZE_LIMIT, "Wrong termination cause");
158 170
  check(mc.cliqueSize() == 2, "Wrong clique size");
159 171
}
160 172

	
161 173

	
162 174
int main() {
163
  checkMaxCliqueGeneral(50, GrossoLocatelliPullanMc<ListGraph>::RANDOM);
164
  checkMaxCliqueGeneral(50, GrossoLocatelliPullanMc<ListGraph>::DEGREE_BASED);
165
  checkMaxCliqueGeneral(50, GrossoLocatelliPullanMc<ListGraph>::PENALTY_BASED);
175
  checkMaxCliqueGeneral(GrossoLocatelliPullanMc<ListGraph>::RANDOM);
176
  checkMaxCliqueGeneral(GrossoLocatelliPullanMc<ListGraph>::DEGREE_BASED);
177
  checkMaxCliqueGeneral(GrossoLocatelliPullanMc<ListGraph>::PENALTY_BASED);
166 178

	
167
  checkMaxCliqueFullGraph(50, GrossoLocatelliPullanMc<FullGraph>::RANDOM);
168
  checkMaxCliqueFullGraph(50, GrossoLocatelliPullanMc<FullGraph>::DEGREE_BASED);
169
  checkMaxCliqueFullGraph(50, GrossoLocatelliPullanMc<FullGraph>::PENALTY_BASED);
179
  checkMaxCliqueFullGraph(GrossoLocatelliPullanMc<FullGraph>::RANDOM);
180
  checkMaxCliqueFullGraph(GrossoLocatelliPullanMc<FullGraph>::DEGREE_BASED);
181
  checkMaxCliqueFullGraph(GrossoLocatelliPullanMc<FullGraph>::PENALTY_BASED);
170 182
                       
171
  checkMaxCliqueGridGraph(50, GrossoLocatelliPullanMc<GridGraph>::RANDOM);
172
  checkMaxCliqueGridGraph(50, GrossoLocatelliPullanMc<GridGraph>::DEGREE_BASED);
173
  checkMaxCliqueGridGraph(50, GrossoLocatelliPullanMc<GridGraph>::PENALTY_BASED);
183
  checkMaxCliqueGridGraph(GrossoLocatelliPullanMc<GridGraph>::RANDOM);
184
  checkMaxCliqueGridGraph(GrossoLocatelliPullanMc<GridGraph>::DEGREE_BASED);
185
  checkMaxCliqueGridGraph(GrossoLocatelliPullanMc<GridGraph>::PENALTY_BASED);
174 186
                       
175 187
  return 0;
176 188
}
0 comments (0 inline)