Throughput matching algorithm and stochastic models for analysis of open and closed manufacturing systems

Abstract

Order release policies such as card-controlled CONWIP policy aim at improving system responsiveness and minimizing system-wide inventory levels. It is not clear if order release policies (without card control) can be equally effective under certain settings of the production system design parameters. In this research, we study the performance of alternate order release (material control policies) under a variety of design parameter settings such as number of stations, station service time characteristics, and the location of bottleneck station using queuing network models. To compare CONWIP policy (closed system) with open control policies (such as deterministic start times), we develop a throughput matching algorithm. The throughput times in CONWIP system is about 2%-50% less in comparison to open control systems; however, we show that the number of stations and the location of the bottleneck station affects the choice of the order release policy, and the benefit of card-controlled policies diminishes for a system with large number of stations (the throughput time percentage benefit using CONWIP in comparison to open control is 1/K, where K is the number of stations). We also analyze the system for a variety of demand inter-arrival times and check its effect on the expected number of backlogs and system wide expected waiting time.