The Impact of the Mutation Strategy on the Quality of Solution of Parallel Genetic Algorithms

摘要

The paper is aimed at investigating the influence of the mutation strategy on the quality of solution of parallel evolutionary algorithms. Parallel computational model based on independent subpopulation evolutions on multicomputer platform is suggested. Several parallel strategies for variable mutation rate at subpopulation and individual levels are investigated and their impact on the quality of the solution is evaluated and analyzed for the case study of the traveling salesman problem. Hybrid programming model utilizing both message passing (MPI) and multithreading (OpenMP) is applied. Parallelism profiling and solution quality analysis are made for the purpose of estimating the efficiency of several parallel mutation strategies.