Paper Info
Title
Distributed NSGA-II sharing extreme non-dominated solutions.
Abstract
A recent trend in multiobjective evolutionary algorithms is to increase the population size to approximate the Pareto front with high accuracy. On the other hand, the NSGA-II algorithm widely used in multiobjective optimization performs non-dominated sorting in solution ranking, which means an increase in computational complexity proportional to the square of the population. This execution time becomes a problem in engineering applications. It is also difficult to achieve high speeds while maintaining the accuracy of solution searching by simply applying fast, parallel processing to standard genetic operations. In this paper, we propose NSGA-II distributed processing in a many-core environment and a migration method that shares extreme Pareto solutions of the current generation among all cores after performing compensation of the non-dominated solution set obtained by distributed processing. Using a two-objective and three-objective constrained knapsack problem for evaluation, we show that the proposed method is effective in improving diversity in solution searching while shortening execution time and increasing the accuracy of solution searching.
Year
Venue
Field
2018
GECCO (Companion)
Population,Mathematical optimization,Evolutionary algorithm,Computer science,Multi-objective optimization,Sorting,Solution set,Knapsack problem,Pareto principle,Computational complexity theory
DocType
ISBN
Citations 
Conference
978-1-4503-5764-7
0
PageRank 
References 
Authors
0.34
3
3
Name
Order
Citations
PageRank
Yuji Sato14818.14
Mikiko Sato22211.53
Minami Miyakawa3156.41