Dynamic economic lot size model with perishable inventory and capacity constraints

• Published in: Applied Mathematical Modelling 2017-08, Vol.48, p.806-820
• Main Authors: Sargut, F. Zeynep, Işık, Gül
• Format: Article
• Language: English
• Subjects: Capacitated production; Complexity analysis; Constraint modelling; Demand; Dynamic programming; Economic models; Functions (mathematics); Heuristic; Inventory management; Lot sizing; Perishable goods; Perishable items; Production capacity; Studies
• ISSN: 0307-904X; 1088-8691; 0307-904X
• DOI: 10.1016/j.apm.2017.02.024

•An ELS under perishability and stationary production capacities is introduced.•NP-hardness of the problem under our assumptions is proven.•Structural properties of the optimal solution are shown.•A Dynamic Programing based heuristic for the overall problem is given.•A heuristic is proposed for the subproblem where the production periods are given. In this study, we consider a dynamic economic lot sizing problem for a single perishable item under production capacities. We aim to identify the production, inventory and backlogging decisions over the planning horizon, where (i) the parameters of the problem are deterministic but changing over time, and (ii) producer has a constant production capacity that limits the production amount at each period and is allowed to backorder the unmet demand later on. All cost functions are assumed to be concave. A similar problem without production capacities was studied in the literature and a polynomial time algorithm was suggested (Hsu, 2003 [1]). We assume age-dependent holding cost functions and the deterioration rates, which are more realistic for perishable items. Backordering cost functions are period-pair dependent. We prove the NP-hardness of the problem even with zero inventory holding and backlogging costs under our assumptions. We show the structural properties of the optimal solution and suggest a heuristic that finds a good production and distribution plan when the production periods are given. We discuss the performance of the heuristic. We also give a Dynamic Programing-based heuristic for the solution of the overall problem.