Stochastic transitions of a mixed-integer linear programming model for the construction supply chain: chance-constrained programming and two-stage programming

• Published in: Operational research 2024-09, Vol.24 (3), p.46, Article 46
• Main Authors: Koutsokosta, Aspasia, Katsavounis, Stefanos
• Format: Article
• Language: English
• Subjects: Business and Management; Computational Intelligence; Constraints; Context; Contractors; Decision making; Design optimization; Economies of scale; Integer programming; Linear programming; Logistics; Management Science; Minimum cost; Mixed integer; Network design; Operations Research; Operations Research/Decision Theory; Optimization; Original Paper; Parameter robustness; Parameter uncertainty; Supply chains; Sustainability; Transportation; Uncertainty
• ISSN: 1109-2858; 1866-1505
• DOI: 10.1007/s12351-024-00856-3

This paper addresses the problem of optimal Construction Supply Chain (CSC) design and integration in deterministic and stochastic environments by providing a family of models for the optimization of a dynamic, multi-product, multi-site contractor-led CSC. With the objective of minimizing the total CSC cost, optimal decisions are made on network design, production, inventory holding and transportation, while also considering discounts for bulk purchases, logistics centers, on-site shortages and an inventory-preparation phase. The models integrate the operations of temporal and project-based supply chains into a sustainable network with repetitive flows, large scope contracts and economies of scale to provide the main contractor with a versatile optimization framework which can account for different levels of uncertainty. The novelty of this paper lies in providing a flexible integrative optimization CSC tool that accounts for multiple CSC actors (suppliers and/or logistics centers), projects, products, time periods, operations, and different decision-making environments depending on the nature of the problem and the risk-attitude of the decision maker. This paper contributes to the fast-growing research field of stochastic CSC optimization showcasing stochastic transitions of a mixed-integer linear programming model to chance-constrained programming and two-stage programming and incorporating uncertainties with different types of probability distributions or scenarios, and even interdependent uncertainties—approaches that have not been explored extensively in the CSC context. The results reveal that the stochastic approaches sacrifice the minimum cost of deterministic solutions having average settings to obtain robust well-hedged solutions over the possible parameter variations and that the selection of a suitable method for modeling uncertainty is context-dependent.