Hybrid Intelligent Optimization Algorithm Based On Particle Awarm And Its Application Research

Optimization problems are generally existed in engineering design,information technology,industrial production and other scientific fields;thus solving optimization problems has a certain theoretical significance and practical value.Intelligent optimization algorithm with the characteristics of simple,efficient and robustness is one of research methods which derived from natural biological phenomena.It has attracted increasing attention and been applied to solve various optimization problems.However,simple intelligent optimization algorithm does not solve the complex high-dimensional,discrete and dynamic optimization problems,so designing hybrid intelligent optimization algorithms to saving complex optimization problems has a certain degree of significance.This paper not only summarizes the research background,significance and status,but also analyzed the drawbacks of many intelligent optimization methods.In this paper,we do some introduction to high-dimensional optimization,virtual machine placement problems and intelligent optimization algorithms,and design two improved hybrid method,PS-ABC and HGPSO-EC.The main research work of this paper as follows:(1)The complexity of high-dimensional optimization problems will grow exponentially with the increasing dimension in the search space,so we propose a hybrid algorithm called PS-ABC.The main idea in this algorithm is that each individual use the exploration ability of the ABC based on PSO in the algorithm process,thus our method has fast convergence and excellent computing performance.The simulation results show that the proposed algorithm is efficient to deal with high-dimensional optimization problems.(2)In order to reduce the waste of server resources in cloud data center,this paper proposes a virtual machine placement hybrid algorithm(HGPSO-EC)based on cluster theory,PSO and GA.The hybrid optimization algorithm mainly includes two parts:ensemble clustering(EC)and HGPSO algorithm.The mind of ensemble clustering can achieve the minimum cost of communication,while HGPSO algorithm which combines the exploration ability of GA with the exploitation ability of PSO can reduce the waste of server resources.The experiment simulation results show that the HGPSO-EC algorithm is feasibility and validity to solve the virtual machine placement problems.