The influence of shop characteristics on workload control

https://doi.org/10.1016/S0925-5273(99)00141-3Get rights and content

Abstract

Several order release methods have been developed for workload control in job shop production. The release methods of the traditional workload control concepts differ in how they deal with the flow of work to each station. Previous research has pointed at strengths and weaknesses of each method. Till now the choice of the appropriate method for a particular situation has hardly received attention. This research shows that shop characteristics are an important factor to this choice. A simulation study indicates that the relative performance of the release methods changes completely with for instance the presence or absence of a dominant flow direction in the shop. Adjustments to the traditional release methods are suggested which prove to make these methods more robust.

Introduction

An important category of production control approaches for job shop production is based on workload control (WLC) principles. The WLC concepts buffer the shop floor against the dynamics of arriving orders by means of input/output control. The order release decision is the main instrument for the input control. Once released, a job remains on the shop floor until all its operations have been completed. WLC concepts set norms for the workload allowed on the floor. If a job does not fit in these norms, the release decision will hold it back. This results in a pool of unreleased jobs.

WLC has received a lot of attention from both practitioners and researchers. Practitioners appreciate the concepts, because they correspond with their intuitive ways of controlling shops. Moreover, they expect practical support in taking decisions. Researchers developed several concepts and workload controlling release methods. Pure job shop models have been used for evaluation, as the concepts were mainly developed for job shop environments. Few researchers have stressed the importance to test the concepts in more realistic situations that deviate from the pure job shop [1], [2].

In a pure job shop model the flows of jobs are undirected, routing sequences are completely random. However, in most real life shops we generally distinguish a dominant flow direction, with workstations having different positions in this flow. As each of the WLC concepts deals differently with the flow of work to these stations, one may expect these deviating characteristics to have influence on the performance of WLC concepts. In this paper we analyse this influence. By means of a simulation study, different shop configurations are examined. The tested WLC concepts are three traditional concepts and two recently developed alternatives [3]. Preliminary simulation results have been discussed in [4].

The results of this study should contribute to the choice of an appropriate workload control in a practical situation. As there still is a lot of confusion on the gap between theoretical and practical results of WLC concepts, the simulation of more realistic shops may also contribute to our understanding of this gap.

The paper is organised as follows. First, we describe the basic principles of workload control and give a detailed analysis of differences between the release methods of WLC concepts. Next, we formulate our expectations with respect to the influence of shop characteristics on each of these methods. Section 5 discusses the experimental design of the simulation study to verify these expectations. 6 Simulation results, 7 Discussion of results deal with the results and the implications of the study.

Section snippets

The workload control (WLC) concept and job release

An important decision within WLC concepts is job release (see [5], [6] for thorough reviews of job release research. Job release determines when each job should enter the shop floor. Once released, a job remains on the floor until all its operations have been completed. The progress of jobs on the shop floor is controlled by priority dispatching in the queues at work stations. The principle of WLC concepts is to control these queues. Norms are set for the workload allowed on the shop floor. If

Three approaches of workload control

The workload released for a station can be subdivided into a direct part (work from jobs queuing at the considered station) and an indirect or upstream part (from jobs queuing at a station upstream). One aim of WLC concepts is to keep the direct load at a low and stable level. Job release cannot completely control the direct load of a work station. Only a part of the jobs in the pool are released directly to the work station. Other jobs arrive from the other work stations after their upstream

Expected influences of shop characteristics

This research started from the perspective that pure job shops do not exist. In every real life job shop there will be more or less dominant flow direction. The operations performed by some stations have a preparative character (gateways or upstream stations), other stations perform typical finishing operations (downstream or finishing stations). Finishing stations will have most of their load upstream, while typical gateways have a lot of completed work downstream on the floor. As we observed

Experimental design

The previous section stated our expectations with respect to the influence of the shop configuration on each of the workload control approaches. We will analyse these influences by means of a simulation study. This section details the release methods and the shop configurations to be simulated.

Simulation results

Fig. 6 shows the lead time performance for each of the four simulated shop configurations. The average total lead time (see Fig. 1) is plotted against the average shop floor time. For each release method a curve is constructed. A mark on the curves is the result of simulating a release method with a specific norm level. As we simulated nine norm levels per method, each curve contains nine marks. A mark is the result of 50 simulation runs of 6000 days, including a start-up period of 2000 days.

Discussion of results

The different results found in each of the shop configurations give rise to further analysis. We will subdivide our discussion into two parts. First, we discuss the influence of variable station positions and routing lengths. Next, we assess the influence of a directed flow.

Conclusions

This research started from the perspective that pure job shops do not exist. In every real life job shop there will be a more or less dominant flow direction. The operations performed by some stations will have a preparative character, other stations will perform typical finishing operations. Finishing stations have most of their load upstream, while typical gateways have a lot of completed work downstream on the floor.

Three workload control approaches have been analysed and it has been

References (19)

There are more references available in the full text version of this article.

Cited by (181)

  • Non-hierarchical work-in-progress control in manufacturing

    2023, International Journal of Production Economics
View all citing articles on Scopus
View full text