Fast Greedy Subset Selection From Large Candidate Solution Sets in Evolutionary Multiobjective Optimization

Subset selection plays an important role in the field of evolutionary multiobjective optimization (EMO). Especially, EMO algorithm with unbounded external archive (UEA), subset is essential post-processing procedure to select a prespecified number solutions as final result. In this article, we discuss efficiency greedy for hypervolume, inverted generational distance (IGD), and IGD plus (IGD+) indicators. Greedy algorithms usually efficiently handle selection. However, when large are given (e.g., from tens thousands UEA), they often become time consuming. Our idea use submodular property, which known hypervolume indicator, improve their efficiency. First, prove that IGD+ indicators also submodular. Next, based on propose efficient inclusion each indicator. We demonstrate through computational experiments proposed much faster than standard algorithms. The help research performance