- Title
- Lot-sizing and scheduling optimization in food supply chain
- Creator
- Chen, Shuo
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2022
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- A common assumption in supply chain management is that products have unlimited lifespans, i.e., they can be stored indefinitely. However, in a real-world environment, especially in the food industry, products undergo a process of decay, damage or spoilage during storage and cannot be served for their original purpose. Due to the perishability of food products, production planning in the food industry is more complex than in other environments. The challenge lies not only in the trade-off between minimizing total costs and meeting market demand, but also in considering uncertainties such as limited shelf life limits and deterioration of food products during production and inventory. Moreover, food industry has to provide quality food products while satisfying quality requirements and safety regulations. In this thesis, we first introduce the distinctive characteristic of food products, perisha- bility that stress the food industry to search for efficient production planning. We also present an overview of fundamental characteristics of production planning in terms of lot sizing and scheduling and the generic single-level capacitated lot sizing and scheduling models. After that, the primary attention is centered on the modeling of perishability in production planning, and the solution approaches used to solve production planning problems available in the research literature. We consider the perishability of the food product in terms of fixed shelf life and disposal. Fixed shelf life refers to periods that food production can stay in inventory, while disposal refers to food losses that occur after exceeding fixed shelf life. Two variants of mixed- integer programming formulations are proposed to integrate fixed shelf life constraints with or without disposal into the Capacitated Lot Sizing Problem with Linked Lot Sizes (CLSPL) model. Computational experiments are performed on problem instances of different sizes, showing the added value of explicitly incorporating perishability considerations into production planning problems. Next, we turn our attention to investigating more efficient solution approaches to solve the proposed production planning problem with fixed shelf life constraints. We introduce a heuristic approach that starts with obtaining an initial solution by relaxing the capacity restriction. Three alternative methods, Wagner and Whitin (WW), Lot for Lot (LL) and Dixon–Silver (DS) are introduced to create different starting points. Then, a backward andforward shift method is employed in an attempt to find a feasible solution based on the initial solution. We are using a Tabu Search (TS) structure that prevents cycling moves in the backward and forward shifts. Finally, we try to improve the solution while preserving feasibility. Computational experiments are performed on the same problem instances, showing the advantage of heuristic approaches on finding feasible solutions quickly in less than 1 second. We then incorporate heuristic approaches in a Memetic Algorithm (MA) processes to improve the quality of solutions for the proposed production planning problem with fixed shelf life constraints. Heuristic approaches are embedded in the evolution loop of MA to generate solutions which are represented as individuals in the population. Population re-construction is applied to avoid population diversity converging too quickly if the best solution is not changed in a specific number of generations. Computational experiments are performed on the same problem instances groups, showing the superiority of MA in finding reasonably good quality solutions in a short time. Finally, we comprehensively compare the performance of heuristic approaches, MA and the Gurobi optimization solver on an extensive set of instances based on five parameters of capacity utilization, the ratio of setup time and capacity, the proportion of setup and holding cost, problem size, and shelf life. We also compare the computational efficiency of these methods when implemented using C and Python programming languages. In short, this study makes an important contribution to the integration of perishability into production planning. It also opens up important research pathways to provide an efficient MA for solving production planning problems in the food industry, while taking into account more real-world characteristics.
- Subject
- lot-sizing; scheduling; supply chain; memetic algorithm
- Identifier
- http://hdl.handle.net/1959.13/1458495
- Identifier
- uon:45439
- Rights
- Copyright 2022 Shuo Chen
- Language
- eng
- Full Text
- Hits: 597
- Visitors: 994
- Downloads: 445
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | ATTACHMENT01 | Thesis | 2 MB | Adobe Acrobat PDF | View Details Download | ||
| View Details Download | ATTACHMENT02 | Abstract | 206 KB | Adobe Acrobat PDF | View Details Download |