Mathematical modeling and solution method of Milkran logistics in Saipa ‎Group's inbound supply chain under consideration of order time windows, ‎return cost of empty pallets and in-vehicle loading restrictions

In the Milkran logistics system, vehicles are sent to collect orders from suppliers and deliver them to ‎assembly lines, according to pre-planned routes. In this way, the vehicle goes to the location of ‎several suppliers to pick up orders and then is sent to one or more destinations for delivery. In this ‎logistics system, cargoes are aggregated within the vehicle and through various nodes in the logistics ‎network. This paper introduces a mixed integer linear programming model for the Milkran logistics ‎problem that takes into account considerations such as three-dimensional loading of the order ‎pallets into vehicles, 50% higher cost for returning empty pallets, order timewindows, and ‎heterogeneous fleets. Given the nature of the problem, an algorithm based on grouping evolutionary ‎strategy is introduced that uses heuristic methods to ensure vehicles’ loading and routing feasibility. ‎The effectiveness of the introduced mathematical model and meta-heuristic algorithm is measured ‎using data collected from Saipa Automotive Group. The computational results show that Milkran ‎Logistics has the ability to reduce costs by 24.5% (on average), compared to the direct shipping ‎strategy pursued by Saipa.‎