Directed Self-Organising Dynamic Topology Hybrid Swarm Intelligence Algorithm And Applications

Optimizing problems exist widely in scientific research and engineering practice.As a kind of optimization algorithm,swarm intelligence algorithm solves many optimization problems in engineering practice,improves the quality and efficiency of production.For this reason,the research focuses on particle swarm optimization,bat algorithm,hybrid swarm intelligence algorithm and its applications.Firstly,in order to solve the problem of single force rule and simple population structure in PSO,a multi force multi subpopulations structure is constructed,and a multi force multi subpopulations PSO algorithm is proposed.Three subpopulations are constructed in the particle swarm.The strategy of searching before and after two stages is adopted in subpopulation 1 and 2.The equilibrium attraction rule is constructed in the early stage of subpopulation 1 and 2,the double acceleration gravitation rule is constructed in the later stage of subpopulation 1,the adaptive attraction rule is constructed in the later stage of subpopulation 2,and the single attraction rule is constructed in subpopulation 3.The standard optimization test function is used to test the algorithm,and the test results are compared with other improved particle swarm optimization algorithms to verify the search ability of multi force multi subpopulations particle swarm optimization algorithm.Secondly,to solve the problem of low precision and poor diversity of bat algorithm,a multi force multi subpopulations bat algorithm is proposed based on multi force multi subpopulations structure.Three subpopulations are constructed in the bat population.In subpopulations 1 and 2,the strategy of searching in the first and second stages is adopted.The equilibrium repulsion rule is constructed in the early stage of subpopulation 1 and 2,the double acceleration gravitation rule is constructed in the later stage of subpopulation 1,the adaptive repulsion rule is constructed in the later stage of subpopulation 2,and the single gravitation rule is constructed in subpopulation 3.The algorithm is tested by standard optimization test function,and the test results are compared with PSO algorithm,EPSO algorithm,BA algorithm and multi force multi subpopulations particle swarm optimization algorithm to verify the search ability of multi force multi subpopulations bat algorithm.Moreover,aiming at the problem that single swarm intelligence algorithm is easy to premature convergence and fall into local optimal solution,a directed self-organizing dynamic topology hybrid swarm intelligence algorithm is proposed.A hybrid swarm intelligence algorithm is formed by combining the multi force multi subpopulations particle swarm optimization algorithm with the multi force multi subpopulations bat algorithm.To improve the interaction between individuals in the hybrid swarm intelligence algorithm,a directed self-organizing dynamic topology is studied,which is combined with the hybrid swarm intelligence algorithm.The performance test of the directed self-organizing dynamic topology hybrid swarm intelligence algorithm verifies the search ability of the directed self-organizing dynamic topology hybrid swarm intelligence algorithm.Finally,the directed self-organizing dynamic topology hybrid swarm intelligence algorithm is applied to the optimization of the structure of the four-roller mill frame and the optimization of the process parameters and gear transmission optimization design.and the practical application of the algorithm is realized.