| Artificial Intelligence (AI) tools and techniques were used to construct KISMET, a discrete manufacturing simulation system. Because of the high degree of isolation between the system's scheduling knowledge and mechanisms used to apply this knowledge, KISMET allows quick and easy examination, modification, and replacement of all or part of the system's knowledge by people or LISP-based computer programs.;This dissertation discusses the fundamental AI concepts germane to KISMET; examines the nature and source of scheduling knowledge; explains the theory, structure, and operation of the system; presents simulations of various manufacturing shops; and explores the use of KISMET as the basis of more advanced manufacturing AI systems.;KISMET incorporates several novel manufacturing concepts, including a clear, efficient language for the representation of multiple alternative production sequences, and a distributed manufacturing control philosophy that is a superset of current manufacturing philosophies. |
