Optimizing Last-Mile Deliveries : Addressing Customer Absence Through Genetic Algorithm

Share

Abstract

Last-mile delivery logistics face significant challenges, particularly regarding customer absences during scheduled delivery times. This issue not only frustrates customers but also imposes substantial economic costs on delivery companies, estimated at up to 15 euros per failed delivery. This research aims to address this problem by optimizing last-mile delivery processes using a genetic algorithm (GA) designed to minimize rerouting costs while respecting customer time preferences. The study compares the performance of the proposed GA with a Simulated Annealing (SA) algorithm, assessing their efficiency in route optimization. Through detailed simulations, GA reduces operational costs by over 35,000 euros annually by considering customer preferences. It significantly outperforms the SA algorithm in scenarios with high customer variability, highlighting its potential for cost-efficient last-mile delivery solutions. Additionally, the GA consistently respected 4–7 more customer preferences per route compared to traditional methods, leading to enhanced customer satisfaction. This work contributes to the field by providing a robust methodology for balancing cost efficiency and user satisfaction in last-mile deliveries, offering actionable insights for logistics optimization.