Intelligent Simulation Modeling of a Flexible Manufacturing System with Automated Guided Vehicles

Abstract

Although simulation is a very flexible and cost effective problem solving technique, it has been traditionally limited to building models which are merely descriptive of the system under study. Relatively new approaches combine improvement heuristics and artificial intelligence with simulation to provide prescriptive power in simulation modeling. This study demonstrates the synergy obtained by bringing together the "learning automata theory" and simulation analysis. Intelligent objects are embedded in the simulation model of a Flexible Manufacturing System (FMS), in which Automated Guided Vehicles (AGVs) serve as the material handling system between four unique workcenters. The objective of the study is to find satisfactory AGV routing patterns along available paths to minimize the mean time spent by different kinds of parts in the system. System parameters such as different part routing and processing time requirements, arrivals distribution, number of palettes, available paths between workcenters, number and speed of AGVs can be defined by the user. The network of learning automata acts as the decision maker driving the simulation, and the FMS model acts as the training environment for the automata network; providing realistic, yet cost-effective and risk-free feedback. Object oriented design and implementation of the simulation model with a process oriented world view, graphical animation and visually interactive simulation (using GUI objects such as windows, menus, dialog boxes; mouse sensitive dynamic automaton trace charts and dynamic graphical statistical monitoring) are other issues dealt with in the study.