The 4th of November, 2019, Wenjuan Gu succesfully defended his thesis titled
Formulations and Algorithms for General and Security Stackelberg Games.
The jury was composed by:
- Yasemin Arda, Professor, Université de Liège
- Luce Brotcorne, Inria Lille Nord Europe
- Diego Cattaruzza, Associate Professor, Centrale Lille
- Fabien Lehuédé, Professor, IMT Atlantique
- Maxime Ogier, Associate Professor, Centrale Lille
- Frédéric Semet, Professor, Centrale Lille
- Kenneth Sorensen, Professor, University of Antwerpen
- Daniele Vigo, Professor, Università di Bologna
Thesis abstract: In this thesis, we study vehicle routing problems considering multiple commodities, with applications in the local fresh food supply chains. The studied supply chain contains two echelons with three sets of actors: suppliers, distribution centers and customers. Suppliers are farmers that produce some fresh foods. Distribution centers are in charge of consolidation and delivery of the products to customers. Distribution centers collect products from the suppliers that perform direct trips. Products are delivered to the customers with a fleet of vehicles performing routes. Each customer requires several commodities, and the farmers produce a limited quantity of these commodities. For the minimization of the transportation cost, it is beneficial that a single customer is delivered by several vehicles. However, for the convenience of the customer, it is imposed that a single commodity is delivered at once by a single vehicle. Hence, different commodities have been explicitly considered. The complete problem is named Multi-Commodity two-echelon Distribution Problem (MC2DP). The restricted problem that addresses only the delivery from a
single distribution center is named Commodity constrained Split Delivery Vehicle Routing Problem (C-SDVRP). We first propose a heuristic based on the Adaptive
Large Neighborhood Search (ALNS) for the C-SDVRP. Then, we address the whole problem (MC2DP) with collection and delivery operations and multiple distribution centers. In order to tackle this complex problem, we propose to decompose the problem: collection and delivery are sequentially solved. Furthermore, we develop an integrated approach for the MC2DP to improve the solutions obtained by the sequential approach.