Hybrid Optimization Algorithm Based On Chaos Optimization

As the use of optimal algorithm goes deeper and deeper, a lot of practical control projects can be summed up into optimal problems. But practical problems of control project generally have the characters of non-linear, strong-constraint, random, large scale etc. Because of the limitation of unitary algorithm, more and more difficulties that can't be solved come forth in solving optimal problems of control project. For example, optimal parameters get into local infinitesimal easily, optimal time is too long and optimal effect can't reach the technical parameters of the control project. So optimal problem is always taken seriously in the fields of theory and project. A hybrid optimal algorithm is proposed based on steepest descent method and chaos optimization in the dissertation. The hybrid optimal algorithm can not only solve the local-infinitesimal of steepest descent method and conquer the slow convergence rate of chaos optimization, but also has the rapid convergence rate of steepest descent method and the global-convergence of chaos optimization.The main research achievement of this dissertation is as follows:1. In order to provide basic theories of hybrid optimal algorithm, the basic content of optimal algorithm are briefly introduced.2. In allusion to the merits and the defects of steepest descent method and chaos optimization, the dissertation puts forward a hybrid optimal algorithm based on two methods. Through validating the hybrid optimal algorithm by numerical experiments, the good optimal impact is gained.3. A new method of the PID controller's parameters-determination based on the hybrid optimal algorithm is proposed. In different PID control systems, many simulations based on different transfer-functions are come off. By the contrast of two results of the parameters-determination based on hybrid optimal and chaos optimization, the superiority of hybrid optimal is proved.4. A PID controller method of invert-pendulum based on the hybridoptimal algorithm is proposed. Good effect is gained in the simulation of the ZWP-II invert-pendulum system.