Parallel Optimization: Loop or Finite-Difference Gradient

Question

David Leather 2017-1-20

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/321301-parallel-optimization-loop-or-finite-difference-gradient

回答： Alan Weiss 2017-1-23

According to the documentation on parallel optimization, one cannot use the 'UseParallel' option inside a 'parfor' loop.

I am optimizing a very messy likelihood function over 32 parameters w/ fmincon. Unfortunately I run the optimization around 1,000 times each at a different initial point.

Would it be more computationally efficient to parallelize the finite-difference gradient calculation, or parallelize the loop of 1,000 calls to fmincon.

Thanks!

PS I have 25 cores available to me.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Walter Roberson 2017-1-20

编辑：Walter Roberson 2017-1-20

Is what you are doing really a global optimization? If so then the tools such as MultiStart in Global Optimization Toolbox would seem appropriate. Or patternsearch.

请先登录，再进行评论。

请先登录，再回答此问题。