An Artificial Immune Algorithm for Travelling Salesman Problem
旅行商问题的人工免疫算法

This paper presents an Artificial Immune Algorithm (AIA)simulating the biological immune systems, and offers its basic principle and approach. Comparing AIA with Genetic Algorithm (GAs)simulating the biological evolution process, the paper points out that the method producing new antibodies in AIA is more versatile than the one producing new individuals in GAs. AIA reflects mechanism of natural selection better than GAs does, as AIA selects effective antibodies from all antibodies by the appetency between an antibody and an antigen and by the repulsion between an antibody and another, while GAs selects new individuals of next colony by the proportion of individual fitness. For Travel Salesman Problem (TSP), this paper brings forward how to describe antibodies artificially, how to produce original antibodies, how to compute the appetency between an antibody and an antigen and the repulsion between an antibody and another, and works out several artificial immune operators producing new antibodies. Simulating examples show that AIA is a very effective method for TSP.