Multi-attribute and predictive cascaded fuzzy system for the AGV dispatching in a flexible manufacturing system

Vinicius Fernandes Caridá

Abstract: In recent years, manufacturers have increasingly applied automation techniques to enhance efficiency and remain competitive. Material handling is an essential activity in any production process. Its effectiveness has a significant impact on production costs. Automated guided vehicle (AGV) systems have become a strategic tool for factories and automated warehouses. They can reduce production costs and improve delivery times in a competitive business scenario. One of the main problems encountered in managing AGVs is the dispatching decision. The present paper proposes a vehicle dispatching system based on a multi-attribute cascaded fuzzy system with a state-space-based Petri net model to predict the future states of the factory. This work aims to reduce the production system’s makespan and/or tardiness values. Additionally, the paper presents one simulated factory scenario to evaluate the proposed fuzzy system against five other dispatching methods. The statistical validation results of the simulations show a 97% confidence level in the hypotheses of this work.