Intelligent scheduling in agricultural wholesale markets: an optimisation approach with simulation-driven and GenAI-based interaction support

This study addresses truck congestion at Japanese agricultural wholesale markets by proposing a novel simulation-driven scheduling system that integrates optimisation and Generative AI (GenAI)-based interaction interface. Congestion caused by upstream scheduling constraints and limited unloading space results in excessive waiting, increased CO2 emissions, and deteriorating labour conditions. In the proposed system, a genetic algorithm under fairness constraints is used to minimise total operational and waiting times by utilising real-time data on truck size, shipment content, unloading duration, and spatial constraints. Additionally, a GenAI-based interaction system provides dynamic instructions to drivers, enhancing their compliance. Simulation experiments using data reflecting actual market conditions demonstrated that, compared to the conventional first-come-first-served method, the proposed system can reduce total unloading time by nearly 10% and average waiting time by over 20%. Furthermore, by reducing truck idling during peak hours, the system contributes to measurable reductions in truck idling–related CO2 emissions, reinforcing its practical relevance for sustainable logistics operations. Overall, the proposed system enhances logistics efficiency and labour conditions, providing an empirical digital transformation paradigm for agricultural wholesale markets.