From separation to fusion: Screening-assisted bilevel collaborative evolutionary optimization for railway freight allocation

Efficient freight space allocation and stowage planning are critical for optimizing transportation efficiency and minimizing operational costs in railway transportation systems of large-scale enterprises. Traditional methods typically handle freight space allocation and stowage decisions in isolation or by simply layering these processes, leading to suboptimal results in terms of transportation efficiency and operational costs. To address this issue, this paper proposes a novel screening-assisted bilevel collaborative evolutionary optimization (Sa-BCEO) algorithm to explore and fusion the interdependencies between freight space allocation and stowage problems, thereby improving transportation efficiency. First, a screening-assisted mechanism (SAM) is designed to alleviate the complexity of the nested structure of bilevel optimization. This mechanism narrows the search space by retaining individuals with higher potential in the upper-level optimization, thereby enhancing efficiency in solving the lower-level optimization problem. Then, a bilevel framework is constructed to optimize the freight allocation and stowage. The effectiveness of the Sa-BCEO algorithm is validated through extensive experiments on a real-world enterprise dataset and two random datasets. Extensive results demonstrate significant improvements in transportation efficiency and cost reduction compared to traditional optimization methods.