Multiagent Elite Search Strategy for Combinatorial Optimization Problems

摘要

Ant Colony System is a new meta heuristics algorithms to solve hard combinatorial optimization problems. It is a population based approach that uses exploitation of positive feedback as well as greedy search. In this paper, we propose a multi colony interaction ant model that achieves positive-negative interaction through an elite strategy divided by intensification strategy and diversification strategy to improve the performance of original ACS. Positive interaction makes agents belonging to other colony to select the high frequency of the visit of edge, and negative interaction makes to escape the selection of relevant edge. And, we compares with original ACS method for the performance. This multi colony interaction ant model can be applied effectively in occasion that problem regions are big and complex, parallel processing is available, and can improve the performance ACS model.