Principles of Sequencing and Scheduling, 2nd Edition

15Simulation Models for the Dynamic Job Shop

15.1 Introduction

One of the most thoroughly studied and widely applied areas of scheduling research involves the dynamic version of the job shop model. When we refer to the “dynamic” version, we mean that jobs are released and arrive at the shop over time. In the dynamic version of simpler models, we have assumed that information about all arrivals is known in advance and that the list of arrivals is finite – no larger than, say, 100 jobs. The dynamic job shop model usually connotes a different setting: Information about arriving jobs is not known in advance – even the timing of arrivals is unknown – and the arrivals are ongoing. Some studies involve performance measures for thousands of jobs. Because different studies involve different numbers of jobs, it is common to use mean values (of flowtime, tardiness, etc.) instead of totals as performance measures.

Because the timing of arrivals is uncertain, we assume that jobs arrive randomly, so that the shop itself behaves like a network of queues. In this context, scheduling is typically carried out by means of dispatching decisions: Each time a machine becomes free, we must decide what it should do next. These scheduling decisions are unavoidable in the operation of such a system. Furthermore, research has demonstrated substantial differences among dispatching procedures, so it makes sense to seek out the decision rules that promote good performance.

The effects of dispatching procedures ...