Replit平台Python环境下PuLP库solve()方法报错求助

问题背景 #

近期在Replit平台创建Python模板Repl，使用PuLP 2.6版本配合Python 3.8求解简单线性规划问题时，调用默认CBC求解器的solve()方法触发PulpSolverError异常。相同代码在旧版Repl（使用PuLP 1.6.10）中可正常运行，新旧Repl的poetry.lock和pyproject.toml配置文件存在差异。尝试降级PuLP到旧版本时出现导入错误，添加msg=True参数也无法获取更多报错日志。

完整代码 #

import pulp

x = pulp.LpVariable.dicts('x', range(2), cat=pulp.LpContinuous)

model = pulp.LpProblem('test', pulp.LpMaximize)

model += 4*x[0] + 3*x[1]

model += 2*x[0] + x[1] <= 10
model += 3*x[0] - 2*x[1] <= 20

status = model.solve()

报错信息 #

Traceback (most recent call last):
  File "main.py", line 12, in <module>
    status=model.solve()
  File "/(...)/lib/python3.8/site-packages/pulp/pulp.py", line 1913, in solve
    status=solver.actualSolve(self,**kwargs)
  File "/(...)/lib/python3.8/site-packages/pulp/apis/coin_api.py", line 137, in actualSolve
    return self.solve_CBC(lp,**kwargs)
  File "/(...)/lib/python3.8/site-packages/pulp/apis/coin_api.py", line 198, in solve_CBC
    raise PulpSolverError(pulp.apis.core.PulpSolverError: Pulp: Error while trying to execute, use msg=True for more details (...)

修复方案 #

1. 手动指定CBC求解器路径 #

Replit环境中默认CBC求解器路径可能异常，手动指定路径并验证：

import pulp
# 先在终端执行`which cbc`获取实际路径，替换下方路径
solver = pulp.COIN_CMD(path='/usr/bin/cbc', msg=True)

x = pulp.LpVariable.dicts('x', range(2), cat=pulp.LpContinuous)
model = pulp.LpProblem('test', pulp.LpMaximize)
model += 4*x[0] + 3*x[1]
model += 2*x[0] + x[1] <= 10
model += 3*x[0] - 2*x[1] <= 20

status = model.solve(solver)

2. 锁定兼容版本的PuLP依赖 #

修改Repl根目录下的pyproject.toml，指定兼容的PuLP版本：

[tool.poetry.dependencies]
python = "^3.8"
pulp = "1.6.10"

执行以下终端命令更新依赖：

poetry lock --no-update
poetry install

若仍有导入错误，可删除poetry.lock文件后重新执行上述命令。

3. 切换到旧版Python模板 #

创建Repl时选择Python 3.7模板，再安装PuLP 1.6.10，该组合已验证可正常运行线性规划求解。

4. 替换为GLPK求解器 #

若CBC求解器始终无法正常工作，可切换到GLPK求解器：

1. 在终端安装GLPK：

apt-get install glpk-utils

2. 修改代码指定求解器：

import pulp
solver = pulp.GLPK_CMD(msg=True)

x = pulp.LpVariable.dicts('x', range(2), cat=pulp.LpContinuous)
model = pulp.LpProblem('test', pulp.LpMaximize)
model += 4*x[0] + 3*x[1]
model += 2*x[0] + x[1] <= 10
model += 3*x[0] - 2*x[1] <= 20

status = model.solve(solver)

内容的提问来源于stack exchange，提问作者Beatriz Oliveira