Designing a Zoned Automated Guided Vehicle System with Multiple Vehicles and Multiple Load Capacity

• Published in: Operations research Vol. 45; no. 6; pp. 857 - 873
• Main Authors: Thonemann, Ulrich W, Brandeau, Margaret L
• Format: Journal Article
• Language: English
• Published: Linthicum, MD INFORMS 01.11.1997; Operations Research Society of America; Institute for Operations Research and the Management Sciences
• Subjects: AGVS; Applied sciences; Approximation; automated systems; Automation; Dispatching; Exact sciences and technology; Incumbents; Industrial trucks. Pallet, palletization; Integer programming; manufacturing; Material handling, hoisting. Storage. Packaging; Materials handling; Motor vehicle traffic; Objective functions; Operational research and scientific management; Operational research. Management science; Operations research; Queuing; Queuing systems; Scheduling, sequencing; Studies; Travel time; Vehicles; Zoning; Sensitivity analysis; Branch and bound method; Traffic; Automated guided vehicle; Poisson process; Waiting time; Dispatching rule; Service time; Multiserver queue; Simulation; Loading; Multiple channel; Handling
• ISSN: 0030-364X; 1526-5463
• DOI: 10.1287/opre.45.6.857

We introduce an analytical model for the design of a multiple-vehicle automated guided vehicle system (AGVS) with multiple-load capacity operating under a "go-when-filled" dispatching rule. The AGVS delivers containers of material from a central depot to workcenters throughout the factory floor. The workcenters are partitioned into delivery zones. They are served by a common pool of automated guided vehicles (AGVs), each of which can carry multiple orders per delivery. The demand of the workcenters and the time until delivery are stochastic. We develop a nonlinear binary integer program to determine the optimal partition of workcenters into zones, the optimal number of AGVs to purchase, and the set of workcenters that warrant AGV delivery, subject to constraints on maximum allowable mean waiting time for material delivery. We develop an analytical expression for the mean waiting time until material delivery and present an efficient branch-and-bound algorithm that solves the AGVS design model optimally. Without the analytical solution method, one would have to simulate the system for all zoning options, all combinations of open and closed workcenters, and all possible numbers of AGVs, in order to determine the optimal AGVS design—an approach likely to be infeasible for most problems.