Clonal Selection Algorithm for Solving Permutation Optimisation Problems: A Case Study of Travelling Salesman Problem

Abstract

As an attempt to solve Permutation Optimisation Problems (POP) by using CLONALG (Clonal Selection Algorithm), a well-established artificial immune system, Traveling Salesman Problem (TSP) is studied as an example in this paper. Operators of CLONALG, especially the hyper-mutation operators are analyzed and modified to make CLONALG adapt to POP. Furthermore, the local search technique is employed to speed up the mature of the repertoire system, and receptor-editing operator is also employed to avoid the premature of the antibody population. Finally, several benchmark problems in TSPLIB are tested to evaluate the best and average performance of the proposed algorithm. Experimental results show the proposed competitive algorithm performs better than both the standard CLONALG and a genetic algorithm.