Paper Info
Title
Evaluating Migration Strategies for an Evolutionary Algorithm Based on the Constraint-Graph that Solves CSP
Abstract
Constraint satisfaction problems (CSPs) occur widely in artificial intelligence. In the last twenty years, many algorithms and heuristics were developed to solve CSP. Recently, a constraint-graph based evolutionary algorithm was proposed to solve CSP, [17]. It shown that it is advantageous to take into account the knowledge of the constraint network to design genetic operators. On the other hand, recent publications indicate that parallel genetic algorithms (PGA's) with isolated evolving subpopulations (that exchange individuals from time to time) may offer advantages over sequential approaches, [1]. In this paper we examine the gain of the performance obtained using multiple populations - that evolve in parallel - of the constraint-graph based evolutionary algorithm with a migration policy. We show that a multiple populations approach outperforms a single population implementation when applying it to the 3-coloring problem.
Year
DOI
Venue
2000
10.1007/3-540-39963-1_21
ISMIS
Keywords
Field
DocType
constraint network,constraint satisfaction problem,multiple population,parallel genetic algorithm,evolutionary algorithm,migration strategies,3-coloring problem,artificial intelligence,genetic operator,multiple populations approach,solves csp,exchange individual
Population,Constraint satisfaction,Mathematical optimization,Evolutionary algorithm,Parallel algorithm,Computer science,Constraint graph,Constraint satisfaction problem,Heuristics,Genetic algorithm
Conference
ISBN
Citations 
PageRank 
3-540-41094-5
1
0.36
References 
Authors
11
2
Name
Order
Citations
PageRank
Arturo Núñez131.85
María Cristina Riff220023.91