Application of computer simulation for the study of supply chain managerial actions of medium-ized enterprises in the apparel cluster of Pernambuco

Abstract

This work represents an effort to model the internal dynamics of a manufacturing cluster’s supply chain strategy by computational means, through an agent-based modelling (ABM). As a case study, we have selected an apparel-manufacturing cluster in Pernambuco, Brazil. The cluster’s supply chain strategy was first expressed by means of the Conceptual System Assessment and Reformulation (CSAR) theoretical framework, and then key elements of it were represented as agents, each one embodying a supply chain manager from one of the medium-sized enterprises (ME) in the cluster. Machine Learning (ML) was used to guide the development of these agents, informed and constrained by real data from the cluster. Therefore, the resulting system allows insights on how elements relate to each other, such as the difference between the conceptual and the implemented model, which variables affect the decision-making process of the agents, whether the perceived current market share position matters more or less in budget allocation, which actions would be performed under specific agent temperaments according to demand forecasting and so on. This approach is relevant by virtue of granting not only stress test and sensitivity analysis of impactful factors regarding supply chain management of the ME, but it also avows ad hoc confirmation of theoretical propositions, such as CSAR itself, onto simulated environments.