Track: Simulation Competition

Abstract

Abstract

Queueing is a common phenomenon in manufacturing and service environments. Practitioners often use queueing network models to analyse these systems for decision-making processes. However, finite queueing systems with blocking are used to a lesser extent due to the complexity of solving the systems. Generally, closed queueing networks (CQN) with finite capacities do not lend themselves to product-form solutions and must be solved using approximation algorithms. This study considers the problem of determining the optimal buffers allocation for an inter-facility material transportation system in a manufacturing facility considering the blocking after service (BAS) phenomenon. Under BAS, a job can only reach the next node until there is space available in the queue of the next node. If the queue capacity of the destination node is full, the job in the preceding node, even after being serviced, will block that server. In a typical material flow, trucks transfer materials between the facility and undergo several sub-processes such as gate entry, weighing, loading, and unloading. This material flow process is modelled using a CQN, and each sub-process is designed as a node with a finite queueing capacity. An analytical model is developed within this context to determine the optimal buffer sizes to maximise the system's throughput. More specifically, a simulation model is developed using the Anylogic software, and the optimisation problem is incorporated into the model to determine the optimal buffer allocations. The Anylogic simulation-optimisation experiment engine uses various metaheuristics through the OptQuest tool to determine the non-dominated solutions of the corresponding CQN.