最小時間制御における最適量子化制御信号について

Abstract

This paper is concerned with the problem of the optimum quantization of control signals in the minimum-time control of the multistage deterministic and discrete processes. For the purpose of minimizing the controlling time needful to bring the values of the state variables of the plant to within the allowable error of the given reference value, being given that the initial state variables of the plant are distributed with the gaussian probability density founction and the reference value of the system takes different values, we obtained the results of computation using the method of dynamic programming from the foundation of the definition of optimum quantization. Some figures are presented which compare the curves of the minimum-time and the timeresponse in case of using the optimum quantized signal with those of the other case and visualize the effect of the optimum quantization for minimizing the expected value of the controlling time.