基于贝叶斯粒子群算法的控制权重矩阵优化

研究了线性二次型调节控制权重矩阵的优化问题。研究了粒子群算法的基本原理及其特点。在传统粒子群算法中引入了贝叶斯思想,利用贝叶斯判别法的基本思想对参加优化的粒子进行预处理,对粒子群算法中的粒子移动速度与方向的损失函数依概率加权平均,并将贝叶斯粒子群算法应用于二次型控制权重矩阵的优化中,提升了权重矩阵优化的收敛速度。通过仿真实验验证了贝叶斯粒子群算法的有效性。

Control Weight Matrix Based on Bayes Particle Swarm Optimization

This paper makes a study of the optimization of weight matrix of linear quadratic regulation(LQR)and the basic principle and characteristics of the particle swarm optimization(PSO)algorithm.Bayes is introduced into the PSO algorithm,and the Bayes discriminant method is used to pretreat the particles participating in the optimization.The loss function of particle moving speed and direction in particle group algorithm is weighted by probability,and Bayesian particle group algorithm is applied to the optimization of quadratic control weight Matrix,thus improving the convergence speed of weight matrix optimization.The validity of Bayesian particle swarm algorithm is verified by the simulation.