A column generation algorithm for the choice-based congested location-pricing problem. (April 2019)

• Record Type: Journal Article
• Authors: Etebari, Farhad
• Is Part Of: Computers & industrial engineering. Volume 130(2019)
• Page Start: 687
• Page End: 698
• Publication Date: 2019-04
• Subjects: Queuing -- Choice-based congested location-pricing problem -- Discrete choice model -- Column generation -- Hybrid heuristic; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2019.03.023
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: Integrating location and pricing decisions in the congested facilities during consecutive time periods. Incorporating Multinomial Logit choice model for capturing customer behavior. Considering dynamic regional pricing approach for demand management. Alternative mathematical model is developed and column generation algorithm is proposed for solving the problem. Developing hybrid heuristic for solving the subproblem. Abstract: This paper presents the congested location-pricing problem under customers' choice behavior. In the choice-based congested location-pricing problem (CCLPP), the appropriate facilities are selected for offering service by deciding about their fare classes during consecutive time periods. The customer choosing behavior is captured by the discrete choice models. The dominant factors in the customers' decision are proximity and service fare classes offered by the facilities. In the congested systems with immobile facilities, customer waiting time for getting service is important and the M / M / 1 / k queuing system is used for capturing facilities queue in this paper. Customers can choose the competitors for getting service according to the no-service utility. The obtained mathematical model for solving this problem is a nonlinear mixed integer model which is extremely hard to solve and therefore, the alternative modeling approach is developed for solving the problem by the column generation algorithm. The subproblem in this approach determines configurations including open facilities and their fare classes and the master problem specifies the number of time periods for offering each configuration. We also develop a hybrid heuristic including greedy and metaheuristic algorithms for solving the subproblem. The numerical experiments and sensitivity analysis are carried out for demonstrating the efficiency of the proposed algorithm. This analysis indicates that the total profit is the most sensitive to the offering fare classes and available resources.