Models and algorithms for a class of production routing problems.

Abstract : The production routing problem (PRP) consists of determining an integrated production and distribution planning that aims to optimize overall cost and improve service level. Although the PRP has been attracting academic and practical interests, it has not been well studied in the literature. Food production routing problem (FPRP) that is more complex than the classic PRP due to food perishability, has rarely been studied. This thesis focuses on developing new models and algorithms for the PRP and FPRP.Firstly, a multi-product PRP with outsourcing (MPRPOS) that is a generalization of the classic PRP is addressed. For the problem, a mixed integer linear programming (MILP) model is proposed and a three-level heuristic is designed. Computational experiments on 225 newly generated MPRPOS instances and 1530 PRP benchmark instances demonstrate the effectiveness and efficiency of the proposed heuristic. Especially, 283 new best solutions for PRP benchmark instances are found by the heuristic.Considering food quality and perishability, and based on the study for the PRP, three new FPRPs are then investigated, i.e., 1) a multi-plant FPRP with packaging consideration (MFPRP); 2) a bi-objective FPRP (BFPRP) that minimizes the total supply chain cost and maximizes food quality simultaneously; and 3) a FPRP with delivery time window constraints (FPRPTW). For each of the studied problems, a MILP model is proposed. Moreover, a hybrid matheuristic that combines a two-phase iterative method, a fix-and-optimize procedure, and a route-based optimization is developed for the MFPRP. For the BFPRP, an epsilon-constraint-based heuristic and a fuzzy logic decision method are proposed to generate near-optimal Pareto solutions and to help decision makers select a preferred solution. And the FPRPTW is directly solved by the state-of-the-art solver CPLEX. A case study shows the proposed model and algorithm for BFPRP can improve food supply chain performance. Computational results on randomly generated instances demonstrate the proposed hybrid matheuristic and epsilon-constraint-based heuristic outperform CPLEX.
Document type :
Theses
Complete list of metadatas

Cited literature [20 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/tel-01961017
Contributor : Frédéric Davesne <>
Submitted on : Wednesday, December 19, 2018 - 4:26:20 PM
Last modification on : Friday, May 17, 2019 - 11:40:39 AM
Long-term archiving on : Thursday, March 21, 2019 - 1:56:03 AM

File

2018SACLE036.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01961017, version 1

Citation

Yantong Li. Models and algorithms for a class of production routing problems.. Automatic. Université Paris-Saclay; Université d'Evry-Val-d'Essonne, 2018. English. ⟨NNT : 2018SACLE036⟩. ⟨tel-01961017⟩

Share

Metrics

Record views

78

Files downloads

55