FEEDER ASSIGNMENT MODELS FOR CONCURRENT PLACEMENT MACHINES

Abstract

Consider computer numerically controlled placement machines which populate printed circuit boards with surface mounted components. These highly concurrent systems require a hierarchy of decision models to determine efficient machine setups and operational sequences so as to achieve effective throughput rates. The Feeder Assignment models developed here can be used to determine the particular stations in which each component fixture is to be attached to the carrier. Noninterference and space constraints may also need to be addressed. Several nonlinear objectives are developed to minimize "excess delay" due to carrier movements. These nonlinear formulations are then transformed to achieve equivalent Linear Mixed Integer Programming formulations. Carrier movements are often the slowest of the concurrent processes and therefore can cause significant deterioration in the overall system performance, if not adequately addressed and properly integrated into a hierarchical decision process. The methodology is illustrated by developing detailed assignment models for a dual delivery system. This system is highly illustrative of the difficulties and advantages arising in the modelling of these complex, concurrent systems. Adapting the general methodology to other machine designs and different manufacturing environments is also discussed.