PRODUCTION CONTROL POLICIES FOR A FLEXIBLE ASSEMBLY SYSTEM

Production control policies for an automated printed circuit board (PCB) assembly system are developed which fulfill a number of characteristics. First, they satisfy demand requirements for a realistic environment, i.e. an environment which allows machine failures and repairs as well as demand changes to occur. Second, they constitute solutions of intermediate generality, i.e. they are not specific to the system under study. Third, they are easily implementable in real time for a computerized system. Policies seek to satisfy both total production volume and balance among part types and to minimize work-in-process. The system's main controlling variables are the part entry decisions; that is, the time and type of part or parts to enter the system. Entry decisions are based on cumulative performance measures and current system capacity constraints like machine failures, bottleneck cell status, and maximum allowed work-in-process. Dispatching in front of production cells is done with the use of simple priority rules. Policies are tested on a simulation model of the PCB assembly system implemented on an AI environment, the KEE-SIMKIT package on an Explorer Lisp machine.