Research On Improved Quantum-Behaved Particle Swarm Optimization And Its Application In Optimization Problems

One of the most important topics in control research is optimization,which refers to the selection of a superior solution from a series of feasible solutions based on certain evaluation criteria.With the development of intelligent algorithms and changes to production requirements,the application of intelligent control to solve increasingly complex optimization problems has become a research hotspot.Taking the quantumbehaved particle swarm optimization(QPSO)algorithm as the main object of study.First,the basic principles and the QPSO algorithm model are introduced,and its performance advantages and deficiencies are summarized.Then,an improved algorithm with a clear theoretical basis is proposed.This paper focuses on studying the potential well adjustment mode of QPSO algorithm,mainly improving the update mode of potential well center and the length adjustment mode of potential well.The potential well center plays a guiding role in the whole population and determines the direction of particle search,while the length of potential well reflects the ability of particle search and affects the search step length.Finally,the parameter optimization problem for the fuzzy controller of an electric vehicle is taken as an application case to verify the performance advantages of the improved algorithm.The specific research work is summarized as follows:First,with the update mode of potential well center as the research focus,it is pointed out that the traditional update mode of potential well center only considers the optimal position of the particle itself and the global optimal position,and the algorithm is easy to converge prematurely.To solve this problem,social learning and Levy flight are introduced to optimize the update mode of potential well center,so as to enhance population diversity.Simulation results showed that the convergence accuracy and speed of the algorithm were better than QPSO algorithm,especially in highdimensional and multi-modal optimization problems where the improved algorithm's performance advantages were obvious.Second,the particle optimization direction was optimized by improving the updating mode of the well center;however,the adjustment mode of the well length also restricted the performance of the algorithm.The traditional well length adjustment method lacks self-adaptability;and there is room for further improvement of the algorithm.Therefore,another research focus of this paper was to improve the adjustment mode of the potential well length.The key control parameter,contraction expansion(CE),was taken as the breakthrough point and the concept of particle activity was proposed.The relative rate of change of the potential well length was taken as a measure of particle activity and particle activity was taken as the feedback quantity to achieve self-adaptation.The CE coefficient was adjusted to reasonably control the length of the potential well.The simulation results showed that the improved algorithm not only reduced the convergence speed and improved the convergence precision,but also gave a new potential well length control strategy.Third,the improved QPSO algorithm was applied to parameter optimization for a fuzzy controller of an electric vehicle,focusing on improving the adjustment mode of quantitative factor and scale factor in the fuzzy adaptive controller.The simulation results showed that the adjustment mode of the quantization factor and proportion factor in the fuzzy adaptive controller were improved.Further,the improved controller was superior to the traditional fuzzy adaptive PID controller;the fuzzy adaptive PID controller was optimized by the improved QPSO algorithm with respect to both speed and precision.