New Model Optimizes Electric Vehicle Routes for Multi-Temperature Cold Chain Delivery
In a significant advancement for urban logistics and sustainable transportation, a new study published in the Journal of Jiangsu University (Natural Science Edition) presents a groundbreaking approach to optimizing delivery routes for electric vehicles (EVs) transporting temperature-sensitive goods. The research, led by He Meiling from Jiangsu University’s School of Automotive and Traffic Engineering, introduces a sophisticated model that enables single EVs to deliver multiple types of perishable cargo—ranging from ambient to frozen products—within a single trip, all while minimizing operational costs and environmental impact.
As cities grow denser and consumer demand for fresh, diverse food options increases, traditional cold chain logistics face mounting pressure. Conventional methods often rely on fleets of specialized refrigerated trucks, each dedicated to a specific temperature zone. This segmentation leads to underutilized vehicle capacity, redundant trips, and higher carbon emissions—challenges that are particularly acute in congested urban environments. With the global push toward decarbonization, the integration of electric vehicles into last-mile delivery networks has become a priority. However, EVs bring their own constraints, including limited battery range and the need for strategic charging infrastructure.
The study directly addresses these complexities by proposing a novel variant of the Electric Vehicle Routing Problem (EVRP), tailored specifically for multi-temperature co-distribution (MTCD) under soft time windows (STW). Unlike rigid delivery schedules, soft time windows allow for early or late arrivals, with associated incentives or penalties, reflecting real-world operational flexibility. This nuanced approach acknowledges that while timeliness is critical—especially for perishable goods—absolute punctuality is not always feasible due to traffic, weather, or other disruptions.
What sets this research apart is its practical and cost-effective solution to multi-temperature transport. Instead of relying on complex, energy-intensive mechanical refrigeration systems in each vehicle, the model leverages passive cooling technology. By using insulated boxes equipped with phase-change materials (referred to in the paper as “coolers” ), a standard electric delivery van can maintain separate temperature zones for ambient, chilled, and frozen goods. This innovation dramatically reduces the need for specialized refrigerated EVs, which are more expensive and consume more energy, thereby lowering both capital and operating costs.
The core of the study lies in its mathematical optimization framework, which seeks to minimize total delivery cost. This cost is not limited to fuel or electricity, but encompasses a comprehensive set of factors: vehicle usage, transportation distance, refrigeration (in the form of cooler deployment), charging, incentive or penalty costs for early or late delivery, and crucially, product spoilage. The inclusion of spoilage cost is a major contribution, as it directly ties route efficiency to product quality—a key concern for retailers and consumers alike.
To solve this complex optimization problem, the research team developed an enhanced version of the Ant Colony Optimization (ACO) algorithm, a metaheuristic inspired by the foraging behavior of ants. While ACO is known for its ability to explore vast solution spaces, it can suffer from slow convergence and a tendency to get trapped in suboptimal solutions. The authors addressed these limitations by integrating the 2-opt algorithm, a local search technique that improves solution quality by iteratively refining routes. Furthermore, they introduced three guiding factors into the ant selection process: a savings matrix to prioritize efficient node connections, a time window waiting factor to minimize idle time, and a frozen product impact factor to prioritize deliveries of more sensitive goods. This hybrid approach, referred to as the Improved Ant Colony Algorithm (IACO), significantly boosts both the speed and accuracy of the solution.
The validity and effectiveness of the model and algorithm were rigorously tested using the well-known Solomon benchmark datasets, a standard in vehicle routing research. The first set of experiments focused on validating the IACO algorithm against established methods. Results showed that the proposed algorithm performed competitively, matching or exceeding the known optimal solutions for several benchmark instances. In the RC101 scenario, the IACO achieved a 1.083% improvement in travel distance, demonstrating its superior search capability. When compared to other advanced algorithms from recent literature, the IACO consistently produced solutions with fewer vehicles and shorter routes, underscoring its practical advantages.
The second phase of the analysis provided a more direct and compelling comparison between traditional single-temperature delivery and the proposed multi-temperature co-distribution model. Using a modified RC101 dataset with 30 customers and 20 charging stations, the researchers simulated a real-world urban delivery scenario. The results were striking. The conventional approach, which required separate fleets for ambient, chilled, and frozen goods, necessitated a total of 11 vehicles. In contrast, the multi-temperature co-distribution model achieved the same delivery objectives with only 5 vehicles—a 54.5% reduction in fleet size. This dramatic decrease in vehicle usage translates directly into lower capital investment, reduced labor costs, and a significantly smaller environmental footprint.
Beyond fleet size, the co-distribution model also led to a 78% reduction in charging events—from 6 stops in the traditional model to just 1 in the optimized scenario. This is a critical finding, as frequent charging not only consumes valuable time but also strains urban charging infrastructure. The ability to complete more deliveries on a single charge enhances operational efficiency and reliability. The total optimized delivery cost was calculated at 2,731.53 yuan, a figure that includes all relevant expenses from electricity to potential spoilage.
The study also explored the impact of soft time window flexibility on overall performance. By expanding the acceptable delivery window from 50% to 200% beyond the customer’s ideal timeframe, the researchers observed a clear trend: as the window widened, the number of required vehicles decreased, and the total cost followed a downward trajectory. When the time window was expanded by 100%, the vehicle count reached its minimum of 5. Further expansion did not reduce the fleet size but continued to lower the total cost, primarily by reducing spoilage and shifting the incentive cost from a penalty to a reward. This is because a more flexible schedule allows for better route sequencing, enabling vehicles to arrive earlier and maintain higher product freshness, thus minimizing losses. This finding provides valuable strategic insight for logistics managers: offering customers slightly more flexible delivery windows can lead to substantial operational savings without compromising service quality.
The implications of this research extend far beyond the academic realm. For logistics companies, the model offers a clear pathway to reduce costs and improve efficiency in the rapidly growing cold chain sector. By adopting multi-temperature co-distribution with standard EVs, companies can streamline their fleets, reduce their energy consumption, and enhance their sustainability credentials. For city planners and policymakers, the findings support the case for investing in EV charging infrastructure and incentivizing green last-mile delivery solutions. The reduction in vehicle kilometers traveled (VKT) and associated emissions contributes to cleaner air and less traffic congestion in urban centers.
Moreover, the success of the IACO algorithm demonstrates the power of hybrid optimization techniques in tackling complex real-world problems. The integration of global search (ACO) with local refinement (2-opt) and problem-specific heuristics provides a robust framework that can be adapted to other logistics challenges, such as dynamic routing, multi-echelon distribution, or disaster relief operations.
While the current model represents a significant leap forward, the authors acknowledge its limitations and outline avenues for future research. The study assumes static customer demand and ideal conditions, without accounting for dynamic factors like real-time traffic, weather disruptions, or sudden changes in order volume. Incorporating these elements would make the model even more robust and reflective of actual operating conditions. Additionally, future work could explore more sophisticated local search strategies or machine learning techniques to further enhance solution quality.
The transition to sustainable urban logistics is not a simple task, but this research provides a powerful tool to help navigate the journey. By combining innovative packaging technology with advanced algorithmic optimization, He Meiling, Fu Wenqing, Wu Xiaohui from Jiangsu University, and Han Xun from Sichuan Police College have demonstrated a practical, scalable solution for the future of cold chain delivery. Their work shows that with the right combination of technology and intelligence, it is possible to deliver fresher food, reduce costs, and protect the environment—all with a smaller, smarter fleet of electric vehicles.
This study is a testament to the power of interdisciplinary research, merging insights from transportation engineering, operations research, and environmental science. It addresses a critical need in modern society: the efficient and sustainable delivery of essential goods. As e-commerce continues to grow and urban populations expand, the demand for smart, green logistics solutions will only intensify. The model presented in this paper offers a compelling blueprint for how the logistics industry can meet this challenge head-on.
The research also highlights the importance of soft time windows as a strategic lever for optimization. In an era where customer expectations for delivery speed are constantly rising, this study provides a counterintuitive but valuable insight: sometimes, offering a bit more flexibility can lead to better overall outcomes. By aligning delivery schedules with operational efficiency rather than rigid customer demands, companies can achieve a win-win scenario—lower costs for themselves and fresher, higher-quality products for their customers.
The use of phase-change materials for temperature control is another key takeaway. This low-tech, high-impact solution bypasses the need for expensive and energy-hungry refrigeration units, making it accessible to a wider range of operators, including small and medium-sized enterprises. This democratization of cold chain technology can help expand access to fresh food in underserved communities, contributing to broader social and economic goals.
In conclusion, this research represents a significant contribution to the field of sustainable logistics. It moves beyond theoretical models to provide a practical, data-driven solution that can be implemented in the real world. The combination of multi-temperature co-distribution, electric vehicles, and intelligent route optimization offers a holistic approach to modern delivery challenges. As cities strive to become smarter and greener, studies like this one will be essential in shaping the future of urban mobility and supply chains.
The findings are particularly timely, as governments and corporations around the world set ambitious targets for carbon neutrality. The logistics sector, a major contributor to greenhouse gas emissions, must play a central role in this transition. The model developed by He Meiling and her colleagues provides a clear and actionable roadmap for reducing the environmental impact of last-mile delivery, one of the most carbon-intensive segments of the supply chain.
Ultimately, this research is not just about algorithms and cost functions; it is about creating a more efficient, resilient, and sustainable way to move goods in the 21st century. By rethinking how we deliver temperature-sensitive products, we can build a logistics system that is better for businesses, better for consumers, and better for the planet.
He Meiling, Fu Wenqing, Han Xun, Wu Xiaohui, School of Automotive and Traffic Engineering, Jiangsu University; Intelligent Policing Key Laboratory of Sichuan Province, Sichuan Police College; Department of Transportation Management, Sichuan Police College. Journal of Jiangsu University (Natural Science Edition), DOI: 10.3969/j.issn.1671-7775.2024.06.002