Mathematical models and optimization algorithms for low-carbon Location-Inventory-Routing Problem with uncertainty
This thesis considers the low carbon Location-Inventory-Routing Problem (LIRP) by addressing the challenges of demand uncertainty through the application of stochastic and fuzzy methods. Multi-objective mathematical models are developed to solve the conflict between total supply chain cost, carbo...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2024
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/119118/ http://psasir.upm.edu.my/id/eprint/119118/1/119118.pdf |
| Summary: | This thesis considers the low carbon Location-Inventory-Routing Problem (LIRP) by
addressing the challenges of demand uncertainty through the application of stochastic
and fuzzy methods. Multi-objective mathematical models are developed to solve
the conflict between total supply chain cost, carbon emission cost, and customer satisfaction
in logistics management. This thesis also aims to solve the low-carbon LIRP
model with uncertainty factors such as carbon trading, customer demand, shortages,
and soft time windows using advanced algorithms. Three LIRP models involving multiple
distribution centers and periods are proposed. The first model is a fuzzy chanceconstrained
programming model that considers factors such as cost, out-of-stock inventory,
carbon trading mechanisms, and fuzzy customer demand. The other two models
are bi-objective mixed integer nonlinear programming models with soft time window
constraints developed to minimize costs and maximize customer satisfaction under uncertain
demand, which include stochastic and fuzzy demand, respectively. Given the
NP-Hard nature of the three models proposed in this thesis, two metaheuristic algorithms
have been developed. A hybrid Particle Swarm Optimization-Bacterial Foraging
Algorithm is developed for solving the single objective LIRP model. Further more, an improved non-dominated sorting genetic algorithm with an elite strategy II
(IMNSGA-II) has been developed to solve the two bi-objective models, surpassing
existing literature’s algorithms such as Pareto Envelope-based Selection Algorithm II
(PESA-II) and NSGA-II. Empirical validation using benchmark dataset and real-world
data from three logistics companies in China demonstrates significant improvements in
supply chain efficiency and cost reduction. When compared to the Supply Chain Guru
X (SCGX) software, the proposed algorithms offer higher practical applicability. |
|---|