Allocation of queuing facilities using a minimax criterion

This paper presents a decision model for allocating demands generated at a set of fixed points to a set of queuing-type facilities at specified locations. The objective of the model is to minimize the maximum expected response time (travel time plus waiting time at the facilities). The case of two M/G/1 facilities is analyzed in depth to provide important insights into the general problem, and an efficient solution algorithm is derived for this special case. Extensions to the algorithm are outlined to handle multiple servers and more than two facilities. Applications of the model may include any type of service where there is a central planner assigning demands to facilities, as found, for example, in many public sector systems.