Multiagent Search Strategy for Combinatorial Optimization Problems in Ant Model

  • Authors:
  • Seokmi Hong;Seunggwan Lee

  • Affiliations:
  • School of Computer Information and Communication Engineering, Sangji University, 660 USan-Dong, WonJu-Si, KangWon-Do, 220-702, Korea;School of General Education, Kyunghee University, 1 SoChon-Dong, GiHung-Gu, YongIn-Si, GyongGi-Do, 446-701, Korea

  • Venue:
  • Computational Intelligence and Security
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Ant Colony System (ACS) is a meta heuristic approach based on biology in order to solve combinatorial optimization problem. It is based on the tracing action of real ants that accumulate pheromone on the passed path and uses as communication medium. In order to search the optimal path, it is necessary to make a search for various edges. In existing ACS, the local updating rule assigns the fixed pheromone value to visited edge in all process. In this paper, modified local updating rule gives the pheromone value according to the number of visiting and the edge's distance between visited nodes. Our approach can have less local optima than existing ACS and can find better solution by taking advantage of more information during searching.