On the Selection of Decomposition Methods for Large Scale Fully Non-separable Problems

Yuan Sun, Michael Kirley, Saman Kumara Halgamuge

Abstract

Cooperative co-evolution is a framework that can be used to effectively solve large scale optimization problems. This approach employs a divide and conquer strategy, which decomposes the problem into sub-components that are optimized separately. However, solution quality relies heavily on the decomposition method used. In recent years, a number of decomposition methods have been proposed, which raises another research question: Which decomposition method is best for a given large scale optimization problem? In this paper, we focus on the selection of the best decomposition method for large scale fully non-separable problems. Four decomposition methods are compared on a suite of benchmark functions. We observe that the random grouping method obtains the best solution quality on the benchmark large scale fully non-separable problems.
BibTeX
@inproceedings{sun2015selection,
  author = {Yuan Sun and Michael Kirley and Saman Kumara Halgamuge},
  title = {{On the Selection of Decomposition Methods for Large Scale Fully Non-separable Problems}},
  booktitle = {GECCO (Companion)},
  year = {2015},
  doi = {10.1145/2739482.2768483},
}

Loading related content...

Figure