A combined genetic algorithm and simulated annealing approach for solving competitive hub location and pricing problem

Abstract

The competitive hub location and pricing problem (CHLPP) describes a situation in which the incumbent firm has already established an optimal hub network with existing hubs for cost minimisation to satisfy all demands. The entrant designs a network to maximise its profit and applies optimal pricing, considering that the existing firm applies mill pricing. Customer's choice factor is solely price modelled using logit function. According to the literature, CHLPP is a NP-hard problem and genetic algorithm (GA) has been previously applied for solving it. In this paper, we propose a more efficient algorithm through combining GA and simulated annealing (SA) algorithm (GA-SA) to solve the mentioned problem. We test the algorithm on the Australia post (AP) data set. Comparing GA-SA and GA computational results indicates that the hybrid GA-SA method outperforms the GA approach in terms of both solution quality (on average 10%) and run time (on average 9%).