Design of Adaptive Chaotic Gradient Optimization Algorithm and Its Application in Electromagnetic Disturbance Testing

With the influence of electrical field value, electrical field frequency and electrical field direction in electrical equipment radial disturbance, the system always appears several local extremum points. In order to solve this problem, a novel adaptive chaos grads optimization method is presented. Firstly, by using grads descend searching method, the optimal points can be figured out, and these points are regarded as initial points. Secondly, search in big pace with Logistic mapping as chaos serial generator and then search in small pace with Ulam-Von Neumann mapping as chaos serial generator. Finally, use this method to reduce the chaos searching space, and exit until another better point in several local extremum points is found. The new initial points are those points figured out with chaos research method. In order to calculate fast, grads descend method with variable pace length is also been used in our method. Simulation results validate the proposed adaptive chaos grads optimization method can approaches limit very fast.