Improved Pso Algorithm Applied Research, Ground Transportation Optimization

Along with the rapid development of societal economy and the durative augment of the city motor vehicles, the traffic of city increases quickly.The problems of the exacerbation of traffic jam,the high frequency of accident and the labefaction of traffic environment become more and more worse .The optimization of ground traffic is one of the importmant approaches to improve the efficiency of ground traffic,and also is the important sign of traffic modernization and intelligentizing.Doing some research on the pivotal problems of ground traffic: Travelling Salesman Problem, Vehicle Routing Problem and traffic signal timing optimization in this dissertation.The primary research is as follows:Particle Swarm Optimization algorithm bases on the theory of community intelligence and has been successfully used in function optimization, restriction optimization and multi-object optimization, but the algorithm has limitation on partial searching capability and premature astringency in some problems. Doing some research on the astringency and complexity of PSO algorithm and obtainning the area of constringency and complexity of PSO algorithm. Improving the PSO algorithm and using the improved PSO algorithm to resolve the Travelling Salesman Problem, Vehicle Routing Problem and traffic signal timing optimization.For Travelling Salesman Problem, using basic PSO algorithm, Particle Swarm Optimization Mutation Operator and Ant Colony Optimization to resolve it about ten and thirty cities respectively.Analysing the result from the average outcome,better outcome,the time of searching and the astringency of improved PSO algorithm.Owing to the particularity of the VRP,designing particular swarm code and using the Simulated Annealing Particle Swarm Optimization, Second-Order Oscillating Particle Swarm Optimization, Particle Swarm Optimization Mutation Operator and PSO-Randomly Varying inertia weight and analysing the result.For single cross signal timing optimization, owing to the particularity of our national traffic flow at present, not using delay time and the degree of parking, but also using the capability of transit for the total goal.Using w-k-PSO algorithm and improved PSO algorithm to resolve the problem aiming at the traffic flow of traffic normal period and traffic fastigium respectively.Study on the multi-cross is numbered at present, using the system that is composed of multi-single cross as the object to optimize signal timing.Using the traffic, capability of transit, saturation and average delay time in fact of two near single cross in Xi'an as the restriction, designing Particle Swarm Optimization to attain effective green time and minimal delay time of trunk line.