A decomposition approach for the probabilistic maximal covering location-allocation problem

The maximal covering location problem (MCLP) maximizes the population that has a facility within a maximum travel distance or time. Numerous extensions have been proposed to enhance its applicability, like the probabilistic model for the maximum covering location-allocation with a constraint in waiting time or queue length for congested systems, with one or more servers per service center. This paper presents a solution procedure for that probabilistic model, considering one server per center, using a column generation and covering graph approaches. The computational tests report interesting results for network instances up to 818 vertices. The column generation results are competitive solving the instances in reasonable computational times, reaching optimality for some and providing good bounds for the difficult instances.