Mathematical models for the periodic vehicle routing problem with time windows and time spread constraints

• Main Authors: Hande Öztop, Damla Kizilay, Zeynel Abidin Çil
• Format: Article
• Language: English
• Published: Balikesir University 2020-09-01
• Series: An International Journal of Optimization and Control: Theories & Applications
• Subjects: Constraint Programming; Mixed Integer Programming; Periodic Vehicle Routing Problem; Time Windows; Capacitated Vehicles
• Online Access: http://www.ijocta.org/index.php/files/article/view/899
• ISSN: 2146-0957; 2146-5703

The periodic vehicle routing problem (PVRP) is an extension of the well-known vehicle routing problem. In this paper, the PVRP with time windows and time spread constraints (PVRP-TWTS) is addressed, which arises in the high-value shipment transportation area. In the PVRP-TWTS, period-specific demands of the customers must be delivered by a fleet of heterogeneous capacitated vehicles over the several planning periods. Additionally, the arrival times to a customer should be irregular within its time window over the planning periods, and the waiting time is not allowed for the vehicles due to the security concerns. This study, proposes novel mixed-integer linear programming (MILP) and constraint programming (CP) models for the PVRP-TWTS. Furthermore, we develop several valid inequalities to strengthen the proposed MILP and CP models as well as a lower bound. Even though CP has successful applications for various optimization problems, it is still not as well-known as MILP in the operations research field. This study aims to utilize the effectiveness of CP in solving the PVRP-TWTS. This study presents a CP model for PVRP-TWTS for the first time in the literature to the best of our knowledge. Having a comparison of the CP and MILP models can help in providing a baseline for the problem. We evaluate the performance of the proposed MILP and CP models by modifying the well-known benchmark set from the literature. The extensive computational results show that the CP model performs much better than the MILP model in terms of the solution quality.