Hybrid SD/DES Simulation for Supply Chain Analysis

Background

Complexity of the supply chain is one of the difficulties associated to supply chain analysis. A supply chain consists of a series of linked activities that intersect departmental and organizational boundaries. When combined with elements of dynamic complexity (Sterman, 2000), such as feedback, time delays, nonlinearity, accumulation and dispersal of resources (stock and flows), the analysis becomes even more difficult. The existence of these elements compels the use of a systemic and dynamic approach to tackle such complexity (Fowler, 1998). System dynamics (SD) can be used in supply chain analysis to generate insight into dynamic behaviour and causal relationships (Towill, 1996; Akkermans & Dellaert, 2005). In SD modelling, a supply chain is represented as a series of stocks and flows where state changes occur continuously over time; individual entities are not specifically modelled, models are generally deterministic and variables usually represent average values (Tako & Robinson, 2012). Although SD is well suited to model complex systems, it has limitations in providing a detailed representation of a supply chain with discrete changes such as high variability, stochastic uncertainties and discrete disturbances (Pereira, 2009). Failure to grasp the impact of uncertainties is considered to be a major pitfall in understanding supply chains (Tannock, Cao, Farr, & Byrne, 2007).