• Login
• Admin

Özdemir, Seda (2017) An integrated model of supplier selection and inventory planning using fuzzy logic and multi-objective evolutionary algorithms. PhD thesis, University of Nottingham.

PDF (Thesis - as examined) - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader Download (4MB) Request Thesis

Abstract

Supplier selection and inventory planning are critical and challenging tasks in supply chain management. There are many studies on both topics and many solution techniques have been proposed dealing with each problem separately. In this thesis, we present a two-stage integrated approach to the supplier selection and inventory planning. In the first stage, in order to get a risk value of each supplier, suppliers are evaluated based on various criteria derived from cost, quality, service and delivery using Interval Type-2 Fuzzy Sets (IT2FSs). In the second stage, the information of supplier rank is fed into an inventory model built to cover the effect of suppliers on the total cost of a supply chain.

The proposed model is formulated as single, multi and many-objective optimisation problems, respectively. Firstly, we generated a set of new instances based on a real world problem. Twenty four problem instances are provided as a benchmark for the community. Various metaheuristics, including MOSA (Multi-objective Simulated Annealing) as a single point based search algorithm, NSGA-II (Nondominated Sorting Genetic Algorithm-II), SPEA2 (Strength Pareto Evolutionary Algorithm 2), IBEA (Indicator Based Evolutionary Algorithm) as population based multi-objective algorithms and NSGA-III (Non-dominated Sorting Genetic Algorithm-III) as a many objective algorithm are applied to those integrated supply chain management problem instances.

It is a well-known fact that parameter setting is crucial for an improved performance of a metaheuristic. Hence, in order to use each algorithm at its best, we employed the experimental design methodology of Taguchi orthogonal arrays for parameter tuning detecting the best setting for each algorithm. A comparative analysis of multi-objective metaheuristics is provided to find the best performing approach in the second stage. The experimental results show that the proposed two-stage approach is indeed capable of solving the integrated supply chain management problem successfully. NSGA-III as a population based technique outperforms the single point based search approach of Simulated Annealing which aggregates multiple objectives into a single objective and other three population based techniques, NSGA-II, SPEA2 and IBEA. Within the population based approaches, NSGA-II performs the best as a multi-objective algorithm when excluding NSGA-III as a many objective algorithm.

Item Type:	Thesis (University of Nottingham only) (PhD)
Supervisors:	John, Robert Ozcan, Ender
Subjects:	H Social sciences > HD Industries. Land use. Labor Q Science > QA Mathematics > QA150 Algebra T Technology > TS Manufactures
Faculties/Schools:	UK Campuses > Faculty of Science > School of Computer Science
Item ID:	44504
Depositing User:	�zdemir, Seda
Date Deposited:	25 Oct 2017 13:57
Last Modified:	07 May 2020 18:02
URI:	https://eprints.nottingham.ac.uk/id/eprint/44504

Actions (Archive Staff Only)

Edit View