The pursuit of cost-efficient and low-carbon logistics has renewed interest in the Capacitated Single Allocation Planar Hub Location Problem (P-CSAHLP), whose combinatorial scale defeats exact optimization once network size moves beyond simplified examples. This study develops an interpretable AI-driven Decision Support System that positions hubs in continuous space, enforces capacity limits, and explores the latent cost trade-off embedded in network design. On benchmark instances, the method remains within 1–2% of mixed-integer optima and solves 1,000-node configurations in minutes. Sensitivity analysis demonstrates how varying first/last-mile versus inter-hub cost ratios simultaneously reshapes topology and carbon footprint, offering managers a defensible parameter dial rather than a black box. By uniting interpretability, scalability, and sustainability, this work leverages Genetic Algorithms as a pragmatic tool for next-generation supply-chain decision support.