Solving a Real-World Integrated Production–Distribution Scheduling Problem with a GRASP

Abstract

In make-to-order manufacturing environments, production scheduling and outbound transportation are strongly interdependent, particularly when delivery commitments and vehicle departures are fixed. This paper addresses an integrated production-distribution scheduling problem arising in a real-world food packaging company. The production system consists of unrelated parallel machines with sequence-dependent, machine-specific setup times, while outbound transportation involves a heterogeneous fleet of vehicles with limited capacity, fixed delivery departure times, and customer-specific delivery constraints. The objective function combines total weighted delivery tardiness and total setup times. To address the problem, we propose a GRASP metaheuristic that incorporates several local search procedures. Computational experiments on real-world instances demonstrate the effectiveness of the proposed approach in producing high-quality solutions within limited computational times. The proposed approach has been implemented within a decision support system and validated in collaboration with the industrial partner, confirming its effectiveness in real-world decision-making.