An intelligent transport planning framework using dual modified K-Means and metaheuristic optimisation for capacitated vehicle routing with time windows and road accessibility

Urban logistics operations face constraints from vehicle capacity, delivery windows, and road accessibility. This research develops an intelligent transport planning framework that integrates Dual m-K-means clustering with a metaheuristic routing engine to solve the vehicle routing problem with time windows and access limits in an operational Transport Management System. Delivery data are clustered using Dual m-K-means considering capacity, service time, and road access to create balanced delivery zones. Within each cluster, routing is optimised through metaheuristic algorithms (e.g. SA, GA, PSO, ACO, Firefly). Results from benchmark and real-world data (288 customers) show that Dual m-K-means improves cluster balance, reduces total mileage by 16.5%, and raises vehicle utilisation to 98%. Integrating into the TMS reduces planning time, supports zoning decisions, and enables adaptive route redesign. The research demonstrates how intelligent clustering and routing can improve fleet efficiency and support sustainable and policy-compliant logistics operations within smart-city mobility systems.