Production Planning in Turkey Breeding Supply Chains: A New Model with Simultaneous Consideration of Harvest Plan and Inventory Decisions

Abstract

Motivated by a real-life turkey meat production problem, this study developed a Mixed Integer Linear Optimization model that addresses flock sizing, flock-coop assignment, mature turkey harvest planning, and carcass meat inventory decisions. Two approaches are proposed to solve the model: a Variable Neighborhood Search (VNS) and a Rolling Horizon Heuristic (RHH). VNS is equipped with problem-specific novel construction and improvement routines. RHH decomposes the problem into smaller subsequent problems. An extensive numerical study is designed to investigate the algorithms' performance and the model's verification. Although VNS outperforms RHH in the solution time, RHH performs better than VNS when solution quality is considered. The overall approach is validated with a real-life case. A significant reduction in total cost compared to the company's existing solution is obtained in a reasonable time. The proposed approach can be generalized to various poultry breeding supply chains through the development of a Decision Support System. Additionally, the study enables parametric and sensitivity analyses to support strategic managerial decisions, such as in-house production ratio.