A Smart-Selection-Based Genetic Algorithm for Delivery and Pickup Problem With Order Time Windows

In a highly competitive environment, businesses strive to optimize their distribution networks to reduce logistics costs. This study focuses on solving vehicle routing problems involving simultaneous delivery and pickup with time windows, addressing both backhaul and divisible delivery and pickup scenarios. A novel hybrid genetic algorithm incorporating smart selection and harem-based crossover methods is proposed to minimize transportation costs while adhering to capacity and time constraints. The smart selection method expedites the solution process by pre-selecting feasible vehicle-route combinations, significantly reducing the computational complexity. Computational experiments on real-world data from the automotive supply industry demonstrate that the proposed algorithm outperforms traditional approaches, achieving substantial cost reductions and high-quality solutions within shorter computation times.