Planning of flow path to minimize expected waiting time for free-ranging automated guided vehicle systems

Recent advances in the technology permit automated guided vehicles (AGVs) to travel without any physical guide path: free-ranging AGVs. These virtual flow paths provide more flexibility since vehicles can travel in both directions on the path. This paper presents a two-phase model, which regards the AGV system network as a graph consisting of nodes connected by a set of arcs, and the AGV system as an M/G/N queueing system with the vehicles representing servers and the materials representing customers. The nodes represent points such as pickup and delivery stations. The decisions are to find the direction on the arcs, such that the materials will have the least expected waiting time when the number of vehicles is given. Phase I, which is an integer programming-based approach, determines the minimum number of vehicles required to satisfy the system designer´s specification on waiting time. Phase II, a heuristic rule-based approach, is to find a flow path which gives the least overall expected waiting time and service time