mipsolver 1.0.3

高性能混合整数规划求解器，现代Python API

MIPSolver

高性能混合整数规划求解器，提供现代化Python API。

安装

pip install mipsolver

快速开始

import mipsolver as mp

# 创建优化模型
model = mp.Model("example")

# 添加变量
x = model.add_var(vtype=mp.BINARY, name="x")
y = model.add_var(vtype=mp.BINARY, name="y")

# 设置目标函数
model.set_objective(5*x + 8*y, mp.MAXIMIZE)

# 添加约束
model.add_constr(2*x + 4*y <= 10, "capacity")

# 求解
model.optimize()

# 获取结果
print(f"最优值: {model.obj_val}")
print(f"x = {x.value}, y = {y.value}")

特性

• 高性能C++求解器核心
• 现代化Python API
• 多种C++ API接口（直接API、C API）
• 支持二进制、整数和连续变量
• MPS文件格式支持
• 跨平台兼容性
• 完整类型提示支持

C++ API

MIPSolver也提供了C++ API用于高性能应用：

C++ 直接API

#include "src/core.h"
#include "src/branch_bound_solver.h"

MIPSolver::Problem problem("example", MIPSolver::ObjectiveType::MAXIMIZE);
int x = problem.addVariable("x", MIPSolver::VariableType::BINARY);
int y = problem.addVariable("y", MIPSolver::VariableType::BINARY);

problem.setObjectiveCoefficient(x, 5.0);
problem.setObjectiveCoefficient(y, 8.0);

int c = problem.addConstraint("capacity", MIPSolver::ConstraintType::LESS_EQUAL, 10.0);
problem.getConstraint(c).addVariable(x, 2.0);
problem.getConstraint(c).addVariable(y, 4.0);

MIPSolver::BranchBoundSolver solver;
MIPSolver::Solution solution = solver.solve(problem);

C API

#include "api/mipsolver_c_api.h"

MIPSolver_ProblemHandle problem = MIPSolver_CreateProblem("example", MIPSOLVER_OBJ_MAXIMIZE);
int x = MIPSolver_AddVariable(problem, "x", MIPSOLVER_VAR_BINARY);
int y = MIPSolver_AddVariable(problem, "y", MIPSOLVER_VAR_BINARY);
// ... 设置目标函数和约束
MIPSolver_SolutionHandle solution = MIPSolver_Solve(problem);

更多C++示例请参考 examples/README.md

构建C++示例

cd examples
./build_examples.sh
./build/test_cpp_direct

许可证

MIT License