G5 Doctoral dissertation (article)
Addressing challenges of real-world lot sizing problems with interactive multiobjective optimization (2023)
Ratkaisuja todellisten toimituserän mitoitusongelmien haasteisiin interaktiivisen monitavoiteoptimoinnin avulla

Kania, A. (2023). Addressing challenges of real-world lot sizing problems with interactive multiobjective optimization [Doctoral dissertation]. University of Jyväskylä. JYU dissertations, 657. http://urn.fi/URN:ISBN:978-951-39-9637-6

JYU authors or editors

Publication details

All authors or editorsKania, Adhe


Journal or seriesJYU dissertations


Publication year2023

Number in series657

Number of pages in the book1 verkkoaineisto (70 sivua, 27 sivua useina numerointijaksoina, 19 numeroimatonta sivua)

PublisherUniversity of Jyväskylä

Place of PublicationJyväskylä

Publication countryFinland

Publication languageEnglish

Persistent website addresshttp://urn.fi/URN:ISBN:978-951-39-9637-6

Publication open accessOpenly available

Publication channel open accessOpen Access channel


Many real-world problems involve multiple conflicting objective functions to be optimized simultaneously, including lot sizing problems, where we need to minimize costs while satisfying demand. A multiobjective optimization problem has many so-called Pareto optimal solutions reflecting different trade-offs. A decision maker (DM) is needed to select one of them to be applied in practice representing best his/her preferences. Interactive methods, that iteratively incorporate the DM’s preferences, are beneficial in supporting the DM, and therefore, in this thesis, we focus on solving lot sizing problems with interactive methods. This thesis tackles challenges in modeling and solving lot sizing problems inspired by real challenges. First, we consider a single-item lot sizing problem under demand uncertainty and propose a safety order time concept that can efficiently handle high fluctuations on demand. Second, we focus on a single-item lot sizing problem under demand and lead time uncertainties, and propose a probability of product availability formula to assess the quality of safety lead time. Third, we integrate a lot sizing problem and a minimum order quantity (MOQ) determination and propose a MOQ level formula to measure the quality of MOQ in satisfying demand. Besides, we also propose multiobjective optimization models to solve these problems. Last, we address a challenge in multi-item lot sizing problems by proposing a decision support approach, called DESMILS. DESMILS enables any single-item multiobjective lot sizing models to be applied in solving multi-item problems by accommodating different preferences from the DM. As a proof of concept, we utilized real data from a company to demonstrate the applicability of the proposed models and approaches. We supported the supply chain manager of the company, as the DM, to find his most preferred solutions by solving the proposed single-item lot sizing models, with interactive methods or the hybridization of methods that we propose. We then demonstrate that, with DESMILS, the DM found Pareto optimal lot sizes for 94 items by solving a single-item multiobjective lot sizing problem for only ten representative items. The DM found all concepts, models, interactive decision making processes, and results useful in his daily operations. These successful applications demonstrate the practical value of the research, which can also benefit others in lot sizing

Keywordsinventory controlsupply chainslogisticsoptimisationmulti-objective optimisationdecision makingdoctoral dissertations

Free keywordsinventory management; multi-item; demand uncertainty; lead time uncertainty; minimum order quantity; decision support

Contributing organizations

Ministry reportingYes

Reporting Year2023

Last updated on 2024-15-06 at 01:26