Published In

Portland International Conference on Management of Engineering & Technology (PICMET)

Document Type


Publication Date



Business logistics -- Technological innovations, Business logistics -- Management, Business logistics -- Computer simulation


The emergence of flexible supply chain systems (FSC) has sparked increased interest in real-time planning, scheduling, and logistics--with particular consideration for strategic implications and overall cost control. Important aspects of an FSC include forecasting, production, materials handling, transportation, and distribution center inventory. There exists a variety of software applications for addressing tactical issues, such as adaptive scheduling, and short term forecasting. However, these programs typically do not permit the user to assess the strategic implications of different policies for flexing capacity and making alternative commitments to manufacturing plants. Recently there has been increased interest in the use of simulation models for strategic policy analysis. Simulation compares favorably to purely analytical methods that often fail to capture the complex interactions of a particular FSC. The challenge is to create an FSC simulation model that is general enough and flexible enough to allow the user to analyze the overall costs and benefits of different policies. This paper presents just such an FSC model, implemented in a system dynamics modeling language. The model is specifically designed to help the user evaluate different policies for scheduling production in the factories and policies that govern factory capacity…in terms of their impact on overall production cost and inventory turns. Preliminary results include new insights regarding the conventional wisdom that minimizing the incremental amount that a factory can fluctuate at any given point in tam will reduce cost. The model suggests the contrary, that requiring larger fluctuations actually reduces the frequency and overall magnitude of the changes without adverse impact on factory utilization.


This paper was presented as part of the proceedings of PICMET '03 (Portland International Conference on Management of Engineering and Technology), Portland, OR, July 2003.

Persistent Identifier