The production routing problem seeks to simultaneously optimize production, routing, and inventory decisions for the plant and the suppliers. In this article an integrated multi-objective sustainable pricing-production-workforce-routing problem is presented for perishable products. Total profit, workforce planning, and vehicle fuel consumption are considered as objective functions due to the importance of operational performance, social, and environmental concerns. The application of the proposed approach is investigated using real case data from a dairy product supply chain. Furthermore, a new solution approach, called Fuzzy Domination Self-Learning Non-Dominated Sorting Algorithm (FDSL-NSGA-II), is developed to solve the problem. The results show that the Pareto solutions of FDSL-NSGA-II outperform those of the classic NSGA-II. Moreover, the findings show that the proposed model can create a surpassing tradeoff between the various aspects of a supply chain, including production, distribution, and workforce planning. In addition, it concurrently optimizes the selling price and protects the environment from the negative impacts of greenhouse gas emissions (GHGs). A comprehensive analysis of the results reveals several managerial insights for decision makers in the logistics industry.

No abstract

Despite the fact that there is a large body of literature on the Production Routing Problem (PRP), we were struck by the dearth of research on outsource planning and lateral transshipment. This paper presents a mixed-integer linear programming model for incorporating outsourcing, lateral transshipment, back ordering, lost sales, and time windows into production routing problems. Then a robust optimization model will be introduced to overcome the detrimental effects of demand uncertainty. Considering the scale and complexity of the suggested problem, addressing it in a reasonable time was a challenge. Therefore, three matheuristic algorithms, including Genetic Algorithm, Simulated Annealing, and Modified Simulated Annealing, are developed for solving large-scale problems. Eventually, computational experiments on disparate instances are performed, and the results show the effectiveness and efficiency of the proposed algorithms. In other words, our recommended algorithms outperform the CPLEX solver in terms of the quality and time of obtaining the solutions.