Track: Facilities Planning and Management

Abstract

Capacitated facility location problem involves choosing a set of locations from a given set of potential locations for establishing facilities, where each potential location has a capacity constraint and fixed cost of establishing the facility. The locations are so chosen that they satisfy the demands of the demand points in a way that minimizes the total cost of establishing the facilities and of transporting goods from facilities to demand points.

 This conventional facility location problem assumes that once the facilities are established, they are always functional and hence can always meet the demand. In real life this assumption does not always hold. To model this reality, with each potential location we associate a probability representing its chances of being functional. We model this problem as a two-stage stochastic programming problem with recourse.

We use sample average approximation method to solve this stochastic integer linear program. Our solution method uses branch and cut method to solve the first stage integer program, and uses Bender’s decomposition method to solve the second stage problems associated with sample points.