A pseudo parallel ant algorithm with an adaptive migration controller

Abstract

This paper presents a pseudo parallel ant algorithm in the continuous domain. The solution vector of the optimized problem is divided into two independent sub-vectors. The proposed algorithm comprises two cooperative colonies, each of which contains tunable and untunable sub-vectors. The colonies optimize their own tunable sub-vectors while their untunable sub-vectors remain unchanged before migration. The excellent tunable sub-vectors in one colony are transferred to the other colony as untunable sub-vectors through a stagnation-based asynchronous migration controller, which adaptively determines the timing and the contents of the migration according to the degree of the stagnation. The proposed parallel ant algorithm with the controller can improve the solution quality and the convergence speed of the ant algorithm in high dimensional space. Numerical experiments are conducted to verify the efficiency of the proposed method. (C) 2008 Elsevier Inc. All rights reserved.