Open Oral Presentation
Speaker: M. Agnel Xavier Fernando
.
Title:
“Development of Solution Methodologies for Multi-Product, Multi-Component, Production-Inventory Planning Problem in a Component Remanufacturing Environment considering Backorders and Quantity Discounts”
Advisor: Dr. M. Mathirajan
Day & Date: Monday, 31st July 2023
Time: 03:00 PM
Venue: Online M S Team Link:
https://teams.microsoft.com/l/meetup-join/19%3ameeting_ZTFjMmI4ODItZmZiZS00MTJhLThmZGYtNTNlMDAzNTA0Y2I2%40thread.v2/0?context=%7b%22Tid%22%3a%226f15cd97-f6a7-41e3-b2c5-ad4193976476%22%2c%22Oid%22%3a%2244805bd2-0703-4c91-93aa-a976ddb8a045%22%7dms
Abstract:
Closed-loop supply chain has been an area of research which has evolved over the years. This thesis aims to address a problem in this domain, specifically from the perspective of an Original Equipment Manufacturer (OEM) in the discrete product manufacturing industry. The OEM seeks to involve remanufacturing as part of its business process wherein its major concern is to satisfy demands for multiple products from customers. To satisfy the demands of products, the OEM needs to assemble the products using multiple components. These components can be obtained from multiple sources viz: manufacturing, remanufacturing, and purchasing from external suppliers.
Incorporating remanufacturing is beneficial for the OEM as it aids in reducing energy consumption and the overall costs. Specifically, the remanufacturing cost is observed to be generally between 40–60% of the manufacturing cost and consuming only 20% of a firm’s effort. Moreover, due to technological constraints, certain components do exist, which cannot be made in-house. Such components need to be only purchased from external suppliers who offer quantity discounts, to motivate the OEM to buy in bulk. Thus, the OEM needs to consider this aspect too while making the purchase decision. Also, at times, if the complete demand for products cannot be satisfied in a said period due to capacity restrictions, the unsatisfied portion of the demand is backordered to future periods. Furthermore, priority is given to satisfy this backordered demand before satisfying current demand of that period.
The OEM thus needs to plan the production and inventory adequately and optimally in such a way that its overall profit is maximized subject to various constraints. This thesis predominantly intends to aid the OEM in this process by addressing a Multi-Product (MP), Multi-Component (MC), Production-Inventory Planning problem (PIPP) in a Component Remanufacturing Environment (CRE) considering Backorders (BO) and Quantity Discounts (QD), [shortly referred as MP-MC-PIPP in CRE with BO and QD], with an objective of maximizing total profit for the OEM. This research problem is addressed by proposing three research objectives.
As part of the first research objective, an Integer Linear Programming (ILP) model is developed with an objective of maximize the OEM profit. The computational complexity of the developed ILP model is studied by varying values of few parameters, one at a time, and empirically observed that the research problem defined in this study is computationally intractable to obtain optimal solution for large scale problem instances. Due to this computational intractable nature of the ILP model, 20 variants of Greedy Heuristic Algorithm (GHA), are proposed as the second research objective of the study. To arrive at the multiple variants of GHA, the overall research problem is split into four integrated decision problems, namely, the, Product Assembly-Backorder (PAB) Problem, Dismantling Products – Remanufacturing Components (DisP-RMC) Problem, Manufacturing Component (MC) Problem and Purchasing Component with Quantity Discount (PCQD) Problem. Subsequently, certain decision rule(s) are proposed for solving each of these decision problems. The 20 variants of the GHA are thus proposed by having a combination of the decision rules proposed for the various decision problems. Furthermore, the performance evaluation of each of the 20 variants of GHA is analyzed both empirically and statistically, by developing a suitable computational experiment [that is, by (a) defining experimental design, (b) identifying benchmark procedures, and (c) identifying suitable performance measures] and the 3 top better performing variants of GHA are identified.
In the last objective of the study, 12 variants of simulated annealing (SA) algorithm are implemented considering the solution obtained from each of the top 3 variants of the proposed GHA as initial solution to the SA algorithm. Before implementing the SA algorithm, the values of the SA parameters are set by following the Taguchi approach. Subsequently, the performance of the proposed 12 variants of SA is evaluated both, empirically and statistically and it is observed that applying the SA algorithm considering the solution obtained from the top performing proposed GHA is improving the loss of optimality considerably.
The practical implication of this study could be to devise, develop, and provide solution approaches to the OEM for the MP-MC-PIPP in CRE with BO and QD to maximize their optimal profits. Additionally, this will aid the OEM to efficiently (i) plan their inventories, (ii) deal with the aspect of backorders and (iii) choose their component purchase plan including the component suppliers in a wise manner. However, there are certain limitations to this study, such as (a) the problem is addressed from a deterministic problem setting, (b) capacity restrictions for inventory and scrapping are not considered and (c) the input data is generated based on an experimental design and not the real data collected from OEM. In addition to overcoming the limitations mentioned in this thesis, there are many immediate future research directions for interested future researchers such as (a) developing alternate heuristic approaches and/or applying other metaheuristic algorithms for the research problem defined in this study, (b) developing an appropriate lower bound which can act as a better benchmark solution procedure instead of the estimated optimal solution and (c) considering environmental and social objectives along with the economic objectives.
