A Kernel search matheuristic to solve the discrete leader-follower location problem

Abstract

In the leader-follower, (r|p)-centroid or Stackelberg location problem, two players sequentially enter the market and compete to provide goods or services. This paper considers this competitive facility location problem in a discrete space. To solve it, the linear programming formulations for the leader and the follower are integrated into an algorithm which, in an iterative process, finds a solution by solving a sequence of these linear problems. We propose a matheuristic procedure that provides solutions for the leader via a kernel search algorithm. Taking into account binary and S-shaped customer choice rules, we present the computational results obtained and compare the exact algorithm with two versions of the kernel search procedure.