A three-phase simulation methodology for generating accurate and precise cycle time–throughput curves. (2nd June 2005)

• Record Type: Journal Article
• Authors: Mackulak, Gerald T.; Fowler, John W.; Park, Sungmin; McNeill, Jennifer E.
• Is Part Of: International journal of simulation and process modelling. Volume 1:Number 1/2(2005)
• Page Start: 35
• Page End: 48
• Publication Date: 2005-06-02
• Subjects: simulation methodology -- manufacturing modelling -- cycle time–throughput curves -- fixed sample size procedures -- stopping rules -- cycle time -- throughput -- capacity management -- production management -- discrete event simulation; Management -- Computer simulation -- Periodicals; Mathematical models -- Periodicals; Operations research -- Periodicals; Simulation methods -- Periodicals; 003.05
• Languages: English
• ISSNs: 1740-2123

Abstract

Cycle time–throughput (CT–TH) curves allow a visual examination of the relationship between projected average cycle time and throughput. They permit one to appraise sensitivity of cycle time to start rate through curvature or steepness. This paper presents an overall framework for efficiently generating simulation-based CT–TH curves that are both accurate and precise. The framework presented in this paper consists of three phases: fixed sample allocation methods; prioritisation of design point inclusion; run length determination through sequential stopping rules. The first phase proposes how to efficiently generate a precise and accurate CT–TH curve given fixed sample size limitations. The second phase provides a set of sequentially ranked design points for systematic simulation experimentation. Finally, the third phase provides a sequential stopping rule that is developed to determine the length of a simulation run based on a time series forecasting procedure. A case study is presented which illustrates the methodology and highlights the fact that the technique provides workable guidelines for both experienced and novice simulation practitioners faced with the task of generating an accurate and precise cycle time–throughput curve. The case study illustrates that this framework is able to produce an equivalently accurate and precise CT–TH curve, for a five station M/M/1 queuing model at a cost of 25% of the samples required using the naïve approach.