Font Size: a A A

Multi-objective Hybrid Optimization Algorithm And Applications Based On PSO And DE

Posted on:2015-02-25Degree:MasterType:Thesis
Country:ChinaCandidate:Y H LiFull Text:PDF
GTID:2298330467470238Subject:Computer technology
Abstract/Summary:PDF Full Text Request
Many decisions in our real-life can be regarded as multi-objective optimization problems,so the study of theories and methods about multi-objective optimization has a very broad scopeof application. Given the limitations of the traditional solution and the advantages ofevolutionary algorithm solving multi-objective optimization problems, it has become a hotresearch field to study the multi-objective evolutionary optimization algorithm. Particle swarmoptimization and differential evolution algorithms are efficient stochastic optimizationalgorithms. After combining them, according to a certain hybrid strategy, then it will get goodresults for solving optimization problems. The main work of the paper is as follows:Initially, the paper outlines the related conception of issues about multi-objectiveoptimization, sums up the traditional and common solution of multi-objective and analyses oftheir limitations. Then it introduces the development process and the latest state of researchabout multi-objective evolutionary algorithm and classifies multi-objective evolutionaryalgorithm classification with introducing in detail the three typical multi-objective evolutionaryalgorithm: PAES, SPEA, NSGA II. Finally, it summarizes the advantages of multi-objectiveoptimization algorithm for solving the problem of evolution.Being based on Particle swarm optimization and differential evolution algorithm, the paperproposes the mixing of the two algorithms by introducing an adaptive algorithm to determinemixing of the two factors. The individual in each circular process can use differential evolutionparticle swarm algorithm or the laws of probability to be update by using adaptive judgmentfactors. The process of particle swarm optimization uses a new velocity update method, and theprocess of differential evolution also adopts improved mutation operation. The paper defines theindividual intensity values of Pareto and improves the crowding distance and definition of theindividual in order to sort and select of the population. Finally, it uses four test functions to testthe performance of the proposed algorithm, and compares with NSGA-II algorithm. Thesimulation results show that the algorithm in terms of convergence and distribution of thesolution set of slightly is better than NSGA-II algorithm.Finally, After analysis of the meaning of urban emergency medical location and the variousfactors affecting location, we built a multi-objective urban emergency medical point location optimization model based on the existing categories Optimization Model summary. Emergencymedical point location problem was as an example in economic development zone of Hanzhong to build corresponding multi-objective location model. We obtained preferred results withusing the particle swarm optimization-based and multi-objective hybrid differential evolutionalgorithm to solve.
Keywords/Search Tags:Multi-objective optimization, Particle swarm optimization, Differentialevolution, Emergency point location
PDF Full Text Request
Related items