Multi-objective optimization algorithms for freight allocation in a food grain supply chain

• Published in: Applied soft computing Vol. 171; p. 112729
• Main Authors: K. P., Anoop, Panicker, Vinay V., Siby, Jerin, C. S., Aryadutt
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.03.2025
• Subjects: Food grain supply chain; Freight allocation; Multi-objective optimization; Multi-objective simulated annealing; Non-dominated sorting genetic algorithm II; Food grain supply chain; Multi-objective optimization; Non-dominated sorting genetic algorithm II; Multi-objective simulated annealing; Freight allocation
• ISSN: 1568-4946
• DOI: 10.1016/j.asoc.2025.112729

This article addresses multi-objective optimization of a freight allocation problem and presents the case of a food grain organization in India (FOI). The inventory and warehouse parameters that are relevant in the regional level allocation of food grains (using freight trains) are represented using three penalty factors, namely rake penalty factor, capacity utilization penalty factor, and weekly penalty factor. The article formulates a tri-objective optimization model to minimize each of the three penalty factors. Two customized multi-objective optimization algorithms are developed based on Multi-Objective Simulated Annealing (MOSA) and Elitist Non-dominated Sorting Genetic Algorithm II (NSGA II) to solve the formulated model. The algorithms are tested and validated via computational experiments designed using historical data collected from the FOI. The algorithms help the transportation managers at the FOI to generate improved and balanced transportation plans (with respect to the three objectives) in a quick time. Further, the performance of the algorithms is compared based on seven different performance metrics reported in literature. The MOSA-based algorithm performs equally or better than the NSGA II-based algorithm with respect to four performance metrics. •Multi-objective optimization of a freight allocation problem is addressed.•Three penalty factors are defined to represent inventory and warehouse parameters.•Multi-objective optimization model minimizes each penalty factor.•MOSA and NSGA II based optimization algorithms are developed and tested.•Algorithms are compared using seven performance metrics reported in literature.