A New Combination Method for Fuzzy Optimal Control

Abstract

In this paper, we present a new hybrid method for solving fuzzy optimal control problems (FOCP). This hybrid method consists of a polynomial and an improved multi-layer perceptron (IMLP) network neural network. An improved neural network is a two-layer neural network. The first layer consists of inputs, weights, and six non-linear sigmoid transfer functions per α-cut and every training point. The second layer, which is the same output layer, includes the weights of the output layer and the neural network outputs, and six linear transfer functions per α-cut and each educational point. Artificial Neural Network training is based on the optimization technique on the target function. This objective function is the error function, and is equal to the sum of the squared errors that are based on the Pontryagin minimization principle.KeywordsFuzzy optimal control problemsIMLP networksPontryagin minimum principleBFGs method