An improved adaptive variable neighborhood search algorithm for stochastic order allocation problem

In practical supply chain operations, efficient order allocation significantly enhances the overall efficiency of the supply chain. Real production environments are plagued by numerous uncertainties, such as unpredictable customer orders, which greatly amplify the complexity of solving practical allocation problems. This study focuses on the problem of allocating orders to parallel machines with varying efficiencies under uncertain and high-dimensional conditions. To maximize the expected profit of order processing, a mathematical model for a high-dimensional stochastic optimization problem is developed, considering the uncertainty due to potential customer order cancellations in a real-world production. By integrating an intelligent optimization algorithm for the order assignment problem with a scenario generation approach, a novel framework for intelligent stochastic optimization is proposed. This framework employs an intelligent optimization algorithm suitable for the generalized assignment problem to search for improved solutions and utilizes the scenario generation method to produce the necessary scenarios for evaluating solutions in high-dimension. Experimental results demonstrate that the proposed approach effectively addresses the high-dimensional stochastic order allocation problem, outperforming the compared method in terms of efficiency and capability.