A knowledge-based intelligent crane scheduling (INCS) system is developed for controlling a stacker crane in a CIM environment where PCBs are made. INCS involves two types of rules, one for cyclic scheduling, which is shown to be near optimal if there is no machine breakdown, and another for handling machine breakdowns, which adjusts the cyclic schedule to real-time disturbances. After briefly reviewing the literature on knowledge-based systems for manufacturing, the authors describe the problem environment. A framework is used to develop not only INCS but also several other knowledge-based systems that use dispatching rules in place of cyclic schedule. The knowledge-based systems are tested by using data from an existing CIM environment, the feasibility and superiority of INCS are demonstrated, and its advantages and disadvantages are discussed. >