or-tools求解整数规划

前言

OR-Tools是一个用于优化的开源软件套件，用于解决车辆路径、流程、整数和线性规划以及约束编程等世界上最棘手的问题。同时OR-Tools提供了C++,Python,Java,.NET的接口，同时提供统一接口封装来调用商业求解器如Gurobi, CPLEX等，也包括开源求解器如SCIP, GLPK, ortools等。在python中安装OR-Tools使用pip命令即可。

问题

代码

from ortools.linear_solver import pywraplp

# 数据准备
data = dict()
data['objective_coefficient'] = [7, 8, 2, 9, 6]
data['constraint_coefficient'] = [[5, 7, 9, 2, 1], [18, 4, -9, 10, 12], [4, 7, 3, 8, 5], [5, 13, 16, 3, -7]]
data['constraint_bound'] = [250, 285, 211, 315]
data['variable_number'] = 5
data['constraint_number'] = 4

# 求解器（SCIP整数规划求解器）
solver = pywraplp.Solver.CreateSolver('SCIP')

# 决策变量
x = []
for i in range(data['variable_number']):
    x.append(solver.IntVar(0, solver.infinity(), 'x[%i]' % i))
print('number of variables=', solver.NumVariables())

# 目标函数
expression = 0
for i in range(data['variable_number']):
    expression += data['objective_coefficient'][i] * x[i]
solver.Maximize(expression)

# 约束
for j in range(data['constraint_number']):
    expression = 0
    for i in range(data['variable_number']):
        expression += data['constraint_coefficient'][j][i] * x[i]
    solver.Add(expression <= data['constraint_bound'][j])
print('number of constraints=', solver.NumConstraints())

# 求解
status = solver.Solve()

# 结果输出
if status == solver.OPTIMAL:
    for i in range(data['variable_number']):
        print(x[i].name(), '=', x[i].solution_value())
    print('objective=', solver.Objective().Value())
    print('problem solved in %f ms' % solver.wall_time())
    print('problem solved in %d iterations' % solver.iterations())
    print('problem solved in %d branch-and-bound node' % solver.nodes())
else:
    print('problem have no optimal solution')

结果