Modelling of a pull production system: a simulation approach

Abstract

In recent years the success of Japanese management techniques has led to their increasing popularity among both practitioners and researchers. One of the techniques that has attracted much attention is the pull production system. It has been defined as a system for producing one unit in a process to be incorporated just in time into a subsequent process. Much attention has been paid to the design of pull systems and related implementation issues. Of the several research methodologies used, one is simulation modelling. Particularly, simulation studies have helped researchers assess the impact of various factors that affect a pull system. However, most of the studies have limited themselves to an investigation of a few factors at a time. Also most of the simulation models have made simplistic assumptions. Additionally most studies have limited themselves to a flow-shop context. Hence, it was felt that there is a need to model a pull production system in the context of a job shop which is more complex than the studies undertaken so far and to identify the issues that may be important for the implementation of such a system. Hence, the problem addressed by this study is: To investigate the effect of various factors on a pull production system, in the context of a job shop, using a simulation model. The job shop being studied has eleven products whose total numbers of operations show a wide variation. Also the shop-floor is characterized by wide variations in total operation time of products and machines and operation times of products on various machines. A simulation model of the shop-floor under a pull system was constructed. The experimental factors and the measures of performance have been decided based on literature as well as discussions with the shop—floor management. The factors of experimentation used in the study are scheduling policies, capacity planning, defect rate, inter-arrival time variations, flexible workforce requirement and due date. The measures of performance used in the study are percentage of jobs in time, delivery time and average work-in-process (WIP). The findings that emerge from the study are: 1. For the system under consideration, the First Come First Served (FCFS) scheduling rule gives a better delivery performance than the Shortest Processing Time (SPT) scheduling rule. 2. If the system has no dominant bottlenecks then, SPT performs as well as FCFS. 3. In case of a flexible workforce the number of workers can be reduced to 50% of the maximum number necessary without significantly affecting the perforrnance. Also, the system needs a slack over and above the theoretical minimum number necessary and this slack is higher than the slack in machine utilization. 4. If the due dates are set tightly then, SPT performs better than FCFS.

In this study an attempt has been made to assess the impact of various factors on a pull production system in the context of a job shop. This study thus has implications for managers who would like to adopt a similar system on their shop-floors.