Abstract
In this paper, we will propose optimal control algorithm that using Radial Basis Function Networks (RBFN) as a type of the Neural Network. Approximation of using RBFN, which is good at nonlinear system, multimodal problem, and optimize locally-detail and globally-rough at once. But, when we use the RBFN at optimal control, the response speed will be important point. Then, we propose Experimental Learning Algorithm that set the basis optimal number and its point to be learning quickly. The basic rule of this algorithm was shown in this paper.
