Guiding genetic operators with immunology principle: a case study in TSP

Traditional genetic operators are often semi-blind due to their randomness and lack of knowledge about problem and lessons from past generations. In this paper, an immunology-guided genetic algorithm, which aims to guide the search direction of genetic operators, is developed to conquer these disadvantages. By simulating memory mechanism and vaccination in immunology, a gene library is established and a vaccination process is integrated into genetic operators. We apply this clever genetic operators to solving TSP, a representative combinatorial optimization problem, and detail the construction of gene library and inver-over operator with vaccination mechanism. Experiment results show that the performance of the original inver-over operator has been significantly improved both in convergent speed and the ability of finding optimum for TSPs.