The authors present a novel methodology for developing knowledge-based control systems. A knowledge-based optimal control system is described. It consists of several expert systems, such as IDSOC (Intelligent Decision-making System for Optimal Control), SMP (Symbolic Mathematical manipulation and Processing system), and a metasystem, as well as a number of numerical computation routines. The authors illustrate the idea of using AI (artificial intelligence) techniques in solving optimal control problems. The philosophy of problem solving is presented. Knowledge representation and implementation issues are discussed in detail. The control of the intelligent system is addressed. The authors discuss the acquisition and representation of expertise knowledge and point out that building an expert system is not just a translation from the existing knowledge into a computer program, but a process in which the new expertise knowledge may be generated and acquired.
