fuzzy multi objective mathematical modeling of hub positioning and routing in a competitive environment

Abstract

designing a distribution network for postal shipments is a complex issue in logistics management, involving the determination of hub locations and the allocation of nodes to them. This system must be designed in such a way that shipments are collected and transferred along specific routes utilizing economies of scale. The aim of this dissertation is to model the hub location-routing problem for the specific conditions of Iran. Initially, the problem is modeled based on the geography and roads of the country, and then constraints and real-world conditions are added to the model. These conditions include time windows, delivery methods, various types of vehicles, fuzzy demands, diverse objectives, and competitive conditions. The models are initially nonlinear and are then transformed into mixed-integer programming models. Due to the NP-hard nature of the problems, a two-stage solution method is designed: the first stage determines hub locations and node allocation using the artificial bee colony algorithm, and the second stage optimizes routing using simulated annealing and several local search methods. Finally, the models are evaluated with several test cases and two real-world studies involving 31 provincial centers and 59 border and major cities in Iran. The results indicate high performance of the models and provide valuable managerial insights for establishing or improving postal shipment companies.