Novel Particle Swarm Optimization Algorithm In Traffic Optimization

Whether it is in the air or on the ground, Vehicle Scheduling Problem is the focusing research of the Intelligent Transportation System (ITS). The research on optimization model and the optimization algorithms of Vehicle Scheduling Problem also has important significance and practical value to enhance the intelligence of Transportation System.Traffic image processing is an important component of the intelligent traffic control. The research on traffic image processing is significance for the control system of the ITS. Applying intelligent optimization algorithm to the optimization problem of image processing can improve the result of image processing, so it is important and valuable for improving veracity and real-timing of the control system of the ITS.As a new and effective parallel optimization algorithm, PSO has attracted the attention of the scholars of our country in the field of intelligent calculation etc and soon is widely applied to function optimization, neural network design, industrial system optimization and fuzzy system control etc. a large number of application research about PSO emerged just a few years time and caused a research focus.This thesis makes research on application of improved PSO in bus scheduling, aircraft departure scheduling and traffic image segmentation and compression. The main work is as follows:Firstly, several new improved PSO algorithms are introduced including the genetic PSO proposed by this thesis. Furthermore, the impact of study factor on the convergence of second-order oscillating PSO is discussed; a new annealing method which making temperature is linear decreased with the increase of the iteration steps is proposed to simulated annealing PSO; the impact of boundary condition on the optimization performance of PSO algorithm is discussed. Finally, the numerical simulation is done to six typical test functions and the performance of each algorithm is analyzed and evaluated according to the simulation results.Secondly, the design idea of using the PSO with shrinkage factor and linear descend inertia weight namely W-K-PSO to solve the bus scheduling problem is given and the simulation experiment is done towards a model of bus schedule.Thirdly, the design idea of using second-order oscillating PSO(SOPSO) and simulated annealing PSO(SAPSO) to solve the aircraft departure scheduling problems is given. The encoding method for particle is designed towards a certain departure scheduling model. Finally, the simulation experiment is done and the result is analyzed.Fourthly, the design ideal of applying PSO inspired by geese warm(GeesePSO) and PSO combined with genetic algorithm(GAPSO) to the threshold image segmentation method based on 2-dimensional histogram is given. Two methods of computing 2-dimensional histogram of an image are proposed and programmed; the rationale of maximal entropy and Otsu thresholding image segmentation method based on 2-dimensional is described. Finally, the simulation experiments are done towards certain images and the results was evaluated.Fifthly, the design ideal of using quantum-behaved PSO(QPSO) to solve the optimal codebook to Vector Quantization(VQ) image compression method is given. The ideal value to the parameter of QPSO was determined according to the theory of VQ technology. The encoding and decoding program to compression is designed and the performance of the codebook solved is evaluated by decoding the image.