Nima Moradi

The importance of Last-Mile Delivery (LMD) in the current economy cannot be overstated, as it is the final and most crucial step in the supply chain between retailers and consumers. In major cities, absent intervention, urban LMD emissions are projected to rise by >30% by 2030 as e-commerce grows (top-100-city “do-nothing” baseline). Sustainable, innovative ground-based solutions for LMD, such as Electric Vehicles, autonomous delivery robots, parcel lockers, pick-up points, crowdsourcing, and freight-on-transit, can revolutionize urban logistics by reducing congestion and pollution while improving efficiency. However, developing these solutions presents challenges in Operations Research (OR), including problem modeling, optimization, and computations. This systematic review aims to provide an OR-centric synthesis of sustainable, ground-based LMD by (i) classifying these innovative solutions across problem types and methods, (ii) linking technique classes to sustainability goals (cost, emissions/energy, service, resilience, and equity), and (iii) identifying research gaps and promising hybrid designs. We support this synthesis by systematically screening 283 records (2010–2025) and analyzing 265 eligible studies. After the gap analysis, the researchers and practitioners are recommended to explore new combinations of innovative solutions for ground-based LMD. While they offer benefits, their complexity requires advanced solution algorithms and decision-making frameworks.

This paper investigates the Heterogeneous Electric Vehicle Routing Problem with Time Windows and Energy Minimization, a critical challenge in sustainable urban logistics. We propose a Mixed-Integer Linear Programming model that explicitly minimizes the total energy consumption of a heterogeneous electric vehicle (EV) fleet while satisfying customer demands within specific time windows. The model accounts for vehicle diversity in terms of battery capacities, consumption rates, and charging characteristics, and incorporates distance- and load-based energy consumption under a full-recharge policy. Computational experiments are conducted on a comprehensive set of benchmark-based instances. Results demonstrate the model’s ability to find optimal or near-optimal solutions for small-scale problems, with acceptable optimality gaps for larger instances. Sensitivity analyses reveal the influence of key parameters—battery capacity ratios, consumption rates, recharging speeds, maximum number of active EVs, and payload sensitivity—on energy usage and fleet configuration. Notably, increasing battery capacity or fleet size significantly reduces energy consumption up to a saturation point, while inefficient or slow-charging vehicles negatively affect performance. The findings have direct implications for logistics operators seeking to enhance energy efficiency and environmental performance in EV-based delivery systems.