Applying Principles of Memetics to Solving the Traveling Salesman Problem
Anton V. Eremeev, Yu.V. Kovalenko · Matematičeskie struktury i modelirovanie · 2019
We consider a population-based algorithm for the asymmetric travelling salesman problem, based on the principles of the memetics. The initial population of tentative solutions is built by means of greedy constructive heuristics. The 3-opt local search is used to improve the initial and the final populations. A balance between the intensity of population-based search and the intensity of the local search methods is chosen by the means of adaptive restart rules. The results of computational experiment on the benchmark instances indicate the viability and effectiveness of the proposed algorithm.