Disruption Management Policy Assessment Through Discrete-Event Agent-Based Simulation

Abstract

Aiming towards the goals of Industry 4.0, companies are increasingly using simulation as a tool to improve performance of manufacturing and logistics systems. The ultimate goal of this line of research is to implement a Digital Twin of a full system, being able to automate a rapid response to disruptions when they occur making the best decisions. Moving towards this goal, this research leverages a high-fidelity large-scale discrete-event agent-based (DEABS) simulator platform of an assembly plant to assess the performance of different disruption management policies offline. This study analyzes the overall performance of individual response strategies, and studies static state-based policies, evaluating the conditions under which one policy dominates another depending on the state on the system when a disruption occurs.The automation of decision making through a Digital Twin will be economically viable when the best decisions are state dependent, meaning the selection of the best policy to implement will depend on the state of the system. Once the existence of state-based policies is validated and all policies are embedded in the model, a live connection to the physical system data enables the automation of disruption management policy selection under uncertainty.