Font Size: a A A

Study Of Particle Swarm Optimization Based On Particle Swarm Optimization

Posted on:2013-10-12Degree:MasterType:Thesis
Country:ChinaCandidate:N XiongFull Text:PDF
GTID:2248330374974904Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Known as the “Third Profit Source”, the logistics has been attached more and moreimportance. With the rapid development of the e-commerce based on the internet, logistics hasplayed more and more important supporting role in e-commerce. While the Vehicle RoutingProblem (VRP) is the key link in the process of logistics distribution optimization as well asthe indispensable content in the e-commerce activity. Optimizing the vehicle routing canimprove the logistics economic benefit and achieve the scientific logistics. Therefore, study onVRP has not only great theoretical significance, but also huge practical value. Particle SwarmAlgorithm is an optimization algorithm based on the Swarm Intelligence, formed by a group ofparticles. Particle swarm makes the cooperative search in the problem space. It is a parallelalgorithm with relatively fast search speed and high search efficiency.The works done in this article include the following aspects:(1). Made a systematic research on VRP and the Particle Swarm Optimization Algorithm,on the basis of which set up a VRP model adopted to solve the mathematical model with theimproved Particle Swarm Algorithm.(2). Proposed the improved Particle Swarm Algorithm on the basis of the feature of theParticle Swarm Algorithm. Due to the Particle Swarm Optimization Algorithm has not onlythe advantages like simplicity, accessibility, few parameters and high speed of convergence,but also some problems. The overriding problems lie in the premature convergence and poorlocal optimization ability. It is prone to get into local optimization and thereby makes theproblem deviate from the optimal solution. Considering that simulate anneal algorithm hasrelatively strong local searching ability, proposed the Particle Swarm Optimization Algorithmcombining the idea of mixed simulate anneal algorithm. At first, made an improvement on thebasic Particle Swarm Algorithm, in which the speed updating formula of the particle adoptedthe updating method with compression factor. As simulate anneal algorithm is highlydependent on the initial temperature, this article will set up a relationship between thedefinition and the feature of the initial temperature. While selected the next generation of theparticles in simulate anneal algorithm, adopted the roulette strategy in the Genetic Algorithm.(3). Solved ordinary vehicle routing problem and the vehicle routing problem with timewindows with the algorithm proposed in this article and verified the efficiency of thisalgorithm through the simulation experiment.
Keywords/Search Tags:Particle Swarm Optimization, Vehicle Routing Problem, mixed strategy, Simulate Anneal Arithmetic
PDF Full Text Request
Related items