A multi-objective framework for railway corridor screening and candidate corridor generation using adaptive hybrid optimization

Early railway corridor planning requires efficient identification of feasible alignments from multiple conflicting objectives in large geographical search spaces. This study proposes the Adaptive Hybrid Reference Optimizer (AHRO), a multi-objective optimization framework for railway corridor screening and preliminary alignment design, which combines hybrid population initialization, adaptive exploration and exploitation strategies, reference-guided learning, and Pareto-based environmental selection to produce diverse high-quality corridor alternatives. The proposed framework was tested on five real and three synthetic benchmark instances with the optimization objectives of surrogate construction cost, corridor distance and rainfall exposure. The suggested algorithm was benchmarked against eight representative optimization algorithms through 24 paired runs, employing the fair Hypervolume (HV) indicator, convergence analysis, and non-parametric statistical significance tests. The experimental results demonstrate that AHRO consistently attains the highest or tied highest fair HV values for all real railway corridor screening benchmarks, while remaining competitive on larger synthetic benchmark instances. The statistical analysis also highlighted significant performance improvements over several constructive and evolutionary baseline algorithms on the real screening problems, while the synthetic benchmarks presented instance-dependent optimization behavior, with no single algorithm outperforming all other cases consistently. In summary, the suggested AHRO framework provides an effective and reproducible decision-support methodology for preliminary railway corridor screening that delivers diverse Pareto optimal trade-off solutions for subsequent engineering evaluation. Future work includes validation of the framework on geographically diverse railway corridors, incorporating additional engineering and environmental objectives, and linking the optimization process to detailed railway alignment design and cost estimation models. Not applicable.