添加许多约束时，PuLP非常慢 [英] PuLP very slow when adding many constraints

问题描述

我正在尝试使用PuLP，但是要花 50秒来添加4000个约束(带有67个变量).解决问题只需要几分之一秒.

I'm trying to use PuLP, but it is taking 50 seconds to add 4000 constraints (with 67 variables). Solving the problem only takes a fraction of a second.

我们希望使用PuLP轻松测试大量问题上的多个求解器.

We want to use PuLP to easily test several solvers on a large set of problems.

要花这么长时间的PuLP吗?直接使用PyGLPK只需几分之一秒，包括设置和求解，所以我希望不要.我该怎么做才能提高PuLP中此步骤的效率?

Should it be taking PuLP this long? Using PyGLPK directly takes only a fraction of second including both setup and solving, so I hope not. What can I do to improve the efficiency of this step in PuLP?

更新

我的约束矩阵非常稀疏，并且通过包含非零系数，我能够将此特定问题的建立时间减少到4或5秒.我仍然能够编写自己的.lp或.mps格式的文件，解决cbc或glpsol子进程的问题，并且比PuLP更有效地解析解决方案，这仅仅是因为我可以在输入Pulp时在几毫秒内写入输入文件需要几秒钟.我仍然不确定为什么会这样.

My constraints matrix is very sparse, and I was able to reduce the setup time to 4 or 5 seconds for this particular problem by only including nonzero coefficients. I am still able to write my own .lp or .mps formatted file, solve the problem with a cbc or glpsol subprocess, and parse the solution much more efficiently than PuLP, simply because I can write the input file in a few ms when PuLP takes several seconds. I'm still not sure why this would be.

推荐答案

我的代表不足以发表评论.

I don't have reps enough to comment.

但是您看过这个吗?

https://groups.google.com/forum/#！topic/pulp-or-discuss/p1N2fkVtYyM

问的问题:

"The problem is solved in less than 0.5 second.
However setting it up with PULP takes more than 10 seconds. "

这似乎也是您报告的内容.解决方案是不使用+ =或sum(...)运算符，而是按照链接中的说明进行操作:

This seems to be what you report as well. And the solution is to not use += or sum(...) operators, but instead, as explained in the link as well:

yeah using the += operator with pulp is really slow, as is using sum() 
instead of lpSum()

那么也许您一次将约束1添加到PuLP，而不是先构建约束列表，然后最后将约束添加到PuLP?

So maybe you are adding the constraints 1 at a time to PuLP instead of building the list of constraints first and then, at the end, add the constraints to PuLP?

这篇关于添加许多约束时，PuLP非常慢的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！