Improved Optimal Control Methods Based upon the Adjoining Cell Mapping Technique

摘要

In this paper, cell mapping methods are studied and refined for the optimal control of autonomous dynamical systems. First, the method proposed by Hsu (Ref.1) is analyzed and some improvements are presented. Second, adjoining cell mapping (ACM), based on an adaptive time of integration (Refs. 2-3), is formulated as an alternative technique for computing optimal control laws of nonlinear systems, employing the cellular state-space approximation. This technique over-comes the problem of determining an appropriate duration of the integration time for the simple cell mapping method and provides a suitable mapping for the search procedures. Artificial intelligence techniques, together with some improvements on the original formulation lead to a very efficient algorithm for computing optimal control laws with ACM (CACM). Several examples illustrate the performance of the CACM algorithm.