highs-sys 1.14.2

Rust binding for the HiGHS linear programming solver. See http://highs.dev.
Documentation 
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

#include "Highs.h"
#include "catch.hpp"

const bool dev_run = false;

bool smallDoubleDifference(double v0, double v1) {
  double difference = std::fabs(v0 - v1);
  //  printf("smallDoubleDifference = %g\n", difference);
  return difference < 1e-4;
}

TEST_CASE("multi-objective", "[util]") {
  HighsLp lp;
  lp.num_col_ = 2;
  lp.num_row_ = 3;
  lp.col_cost_ = {0, 0};
  lp.col_lower_ = {0, 0};
  lp.col_upper_ = {kHighsInf, kHighsInf};
  lp.row_lower_ = {-kHighsInf, -kHighsInf, -kHighsInf};
  lp.row_upper_ = {18, 8, 14};
  lp.a_matrix_.start_ = {0, 3, 6};
  lp.a_matrix_.index_ = {0, 1, 2, 0, 1, 2};
  lp.a_matrix_.value_ = {3, 1, 1, 1, 1, 2};
  Highs h;
  h.setOptionValue("output_flag", dev_run);

  for (HighsInt k = 0; k < 2; k++) {
    // Pass 0 is continuous; pass 1 integer
    if (dev_run)
      printf(
          "\n******************\nPass %d: var type is %s\n******************\n",
          int(k), k == 0 ? "continuous" : "integer");
    for (HighsInt l = 0; l < 2; l++) {
      // Pass 0 is with unsigned weights and coefficients
      double obj_mu = l == 0 ? 1 : -1;
      if (dev_run)
        printf(
            "\n******************\nPass %d: objective multiplier is "
            "%g\n******************\n",
            int(l), obj_mu);

      if (k == 0) {
        lp.integrality_.clear();
      } else if (k == 1) {
        lp.integrality_ = {HighsVarType::kInteger, HighsVarType::kInteger};
      }
      h.passModel(lp);

      h.setOptionValue("blend_multi_objectives", true);

      HighsLinearObjective linear_objective;
      std::vector<HighsLinearObjective> linear_objectives;
      REQUIRE(h.clearLinearObjectives() == HighsStatus::kOk);

      // Begin with an illegal linear objective
      if (dev_run) printf("\nPass illegal linear objective\n");
      linear_objective.weight = -obj_mu;
      linear_objective.offset = -obj_mu;
      linear_objective.coefficients = {obj_mu * 1, obj_mu * 1, obj_mu * 0};
      linear_objective.abs_tolerance = 0.0;
      linear_objective.rel_tolerance = 0.0;
      REQUIRE(h.addLinearObjective(linear_objective) == HighsStatus::kError);
      // Now legalise the linear objective so LP has nonunique optimal
      // solutions on the line joining (2, 6) and (5, 3)
      if (dev_run) printf("\nPass legal linear objective\n");
      linear_objective.coefficients = {obj_mu * 1, obj_mu * 1};
      REQUIRE(h.addLinearObjective(linear_objective) == HighsStatus::kOk);

      REQUIRE(h.run() == HighsStatus::kOk);
      h.writeSolution("", kSolutionStylePretty);
      REQUIRE(smallDoubleDifference(h.getInfo().objective_function_value, -7));
      // Save the linear objective for the next
      linear_objectives.push_back(linear_objective);

      // Add a second linear objective with a very small minimization
      // weight that should push the optimal solution to (2, 6)
      if (dev_run) printf("\nPass second linear objective\n");
      linear_objective.weight = obj_mu * 1e-4;
      linear_objective.offset = 0;
      linear_objective.coefficients = {obj_mu * 1, obj_mu * 0};
      REQUIRE(h.addLinearObjective(linear_objective) == HighsStatus::kOk);

      REQUIRE(h.run() == HighsStatus::kOk);
      h.writeSolution("", kSolutionStylePretty);
      REQUIRE(smallDoubleDifference(h.getSolution().col_value[0], 2));
      REQUIRE(smallDoubleDifference(h.getSolution().col_value[1], 6));
      linear_objectives.push_back(linear_objective);

      if (dev_run) printf("\nClear and pass two linear objectives\n");
      REQUIRE(h.clearLinearObjectives() == HighsStatus::kOk);
      REQUIRE(h.passLinearObjectives(2, linear_objectives.data()) ==
              HighsStatus::kOk);
      REQUIRE(h.run() == HighsStatus::kOk);
      h.writeSolution("", kSolutionStylePretty);
      REQUIRE(smallDoubleDifference(h.getSolution().col_value[0], 2));
      REQUIRE(smallDoubleDifference(h.getSolution().col_value[1], 6));

      // Set illegal priorities - that can be passed OK since
      // blend_multi_objectives = true
      if (dev_run)
        printf(
            "\nSetting priorities that will be illegal when using "
            "lexicographic "
            "optimization\n");
      linear_objectives[0].priority = 0;
      linear_objectives[1].priority = 0;
      REQUIRE(h.passLinearObjectives(2, linear_objectives.data()) ==
              HighsStatus::kOk);

      // Now test lexicographic optimization
      h.setOptionValue("blend_multi_objectives", false);

      if (dev_run) printf("\nLexicographic using illegal priorities\n");
      REQUIRE(h.run() == HighsStatus::kError);

      if (dev_run)
        printf(
            "\nSetting priorities that are illegal now blend_multi_objectives "
            "= "
            "false\n");
      REQUIRE(h.passLinearObjectives(2, linear_objectives.data()) ==
              HighsStatus::kError);

      if (dev_run)
        printf(
            "\nSetting legal priorities for blend_multi_objectives = false\n");
      linear_objectives[0].priority = 10;
      REQUIRE(h.passLinearObjectives(2, linear_objectives.data()) ==
              HighsStatus::kOk);

      if (dev_run)
        printf("\nLexicographic using existing multi objective data\n");
      REQUIRE(h.run() == HighsStatus::kOk);
      h.writeSolution("", kSolutionStylePretty);
      REQUIRE(smallDoubleDifference(h.getSolution().col_value[0], 2));
      REQUIRE(smallDoubleDifference(h.getSolution().col_value[1], 6));

      // Back to blending
      h.setOptionValue("blend_multi_objectives", true);
      //  h.setOptionValue("output_flag", true);
      REQUIRE(h.clearLinearObjectives() == HighsStatus::kOk);
      linear_objectives[0].coefficients = {obj_mu * 1.0001, obj_mu * 1};
      linear_objectives[0].abs_tolerance = 1e-5;
      linear_objectives[0].rel_tolerance = 0.05;
      linear_objectives[1].weight = obj_mu * 1e-3;
      if (dev_run)
        printf(
            "\nBlending: first solve objective just giving unique optimal "
            "solution\n");
      REQUIRE(h.passLinearObjectives(1, linear_objectives.data()) ==
              HighsStatus::kOk);

      REQUIRE(h.run() == HighsStatus::kOk);
      h.writeSolution("", kSolutionStylePretty);

      REQUIRE(h.passLinearObjectives(2, linear_objectives.data()) ==
              HighsStatus::kOk);

      REQUIRE(h.run() == HighsStatus::kOk);
      h.writeSolution("", kSolutionStylePretty);

      // Back to lexicographic optimization
      h.setOptionValue("blend_multi_objectives", false);

      if (dev_run) printf("\nLexicographic using non-trivial tolerances\n");
      REQUIRE(h.run() == HighsStatus::kOk);
      h.writeSolution("", kSolutionStylePretty);

      if (k == 0) {
        REQUIRE(smallDoubleDifference(h.getSolution().col_value[0], 4.9));
        REQUIRE(smallDoubleDifference(h.getSolution().col_value[1], 3.1));
      } else {
        REQUIRE(smallDoubleDifference(h.getSolution().col_value[0], 5));
        REQUIRE(smallDoubleDifference(h.getSolution().col_value[1], 3));
      }

      linear_objectives[0].abs_tolerance = kHighsInf;

      REQUIRE(h.passLinearObjectives(2, linear_objectives.data()) ==
              HighsStatus::kOk);

      REQUIRE(h.run() == HighsStatus::kOk);
      h.writeSolution("", kSolutionStylePretty);

      //  printf("Solution = [%23.18g, %23.18g]\n",
      //  h.getSolution().col_value[0], h.getSolution().col_value[1]);
      if (k == 0) {
        REQUIRE(smallDoubleDifference(h.getSolution().col_value[0], 1.30069));
        REQUIRE(smallDoubleDifference(h.getSolution().col_value[1], 6.34966));
      } else {
        REQUIRE(smallDoubleDifference(h.getSolution().col_value[0], 2));
        REQUIRE(smallDoubleDifference(h.getSolution().col_value[1], 6));
      }
    }
  }

  h.resetGlobalScheduler(true);
}