Research Background
The rapid growth of on-demand food-delivery platforms has transformed the way we consume food, but it also presents significant challenges, particularly in densely populated urban areas. This section explores the problem, its importance, the industry context, and the shortcomings of prior approaches.
The Problem and Its Importance
On-demand food-delivery platforms have become increasingly popular, with a 37% increase in takeaway orders facilitated by these platforms in Hong Kong in the first quarter of 2024. Additionally, 34% of restaurants indicated that delivery platforms contribute to at least 30% of their revenue, underscoring the substantial impact these services have on the food industry. However, this growth comes with significant societal challenges. During peak hours, customers often experience delivery delays due to high order volumes, insufficient courier supply, and road congestion. These issues are compounded by the pressure on couriers to meet tight delivery windows, leading to safety concerns and labor rights issues. For instance, in Shanghai, there were 325 road traffic accidents involving the express delivery and food-delivery industries in the first half of 2019, resulting in 5 deaths and 324 injuries. These incidents highlight the urgent need to find lasting solutions that foster a healthy relationship between the platforms and gig workers.
The rapid growth of on-demand food-delivery platforms has not only transformed the way we consume food but also introduced significant societal challenges. In densely populated urban areas, the surge in demand during peak hours often leads to delivery delays, which can be attributed to several factors. First, the high volume of orders can overwhelm the available courier supply, leading to longer wait times for both customers and couriers. Second, road congestion further exacerbates the problem, making it difficult for couriers to navigate through busy streets. Third, the pressure on couriers to meet strict delivery deadlines can compromise their safety, as they may engage in risky behaviors to avoid penalties. These issues have led to increased scrutiny of the working conditions and safety of gig workers in the food-delivery industry. For example, in China, food delivery platforms have faced criticism for their unsympathetic algorithms that impose strict delivery deadlines, placing excessive pressure on drivers and potentially increasing safety risks. The Shanghai Municipal People’s Government reported that in the first half of 2019, there were 325 road traffic accidents involving the express delivery and food-delivery industries in the city, resulting in 5 deaths and 324 injuries. These incidents illustrate the considerable pressure on stakeholders and highlight the urgent need to find lasting solutions that foster a healthy relationship between the platforms and gig workers, which are crucial for the sustainable growth of the food-delivery market.
Industry Context: Supply Chain and AI Decision-Making
The food-delivery industry is a critical part of the modern supply chain, connecting restaurants, couriers, and customers. The introduction of drone technology offers a promising solution to the challenges faced by traditional ground delivery. Drones provide faster delivery speeds, lower operational costs, and independence from traffic conditions. For example, Manna, an Irish drone delivery company, has launched takeaway deliveries in Dublin and plans to expand to 25 more locations. Similarly, Wing, a drone delivery company, launched drone-assisted food-delivery services in Melbourne in 2024, empowering 250,000 residents to order food through DoorDash with deliveries executed by small aircraft. Flytrex, a drone-based food delivery service operating in North Carolina and Texas, has announced that it successfully fulfilled 100,000 food delivery orders in 2024, reaching 70% of households within its service areas. However, integrating drones into the existing infrastructure and coordinating them with human couriers is a complex task that requires sophisticated decision-making algorithms. The coordination of drones and human couriers involves multiple stages, including the initial pickup, transportation to launchpads, loading onto drones, and final delivery, which necessitates advanced AI and optimization techniques.
The integration of drone technology into the food-delivery supply chain represents a significant shift in how deliveries are managed. Drones offer several advantages over traditional ground delivery methods, such as faster delivery speeds, lower operational costs, and the ability to operate independently of traffic conditions. For example, Manna, an Irish drone delivery company, has successfully launched takeaway deliveries in Dublin and plans to expand to 25 more locations. Similarly, Wing, a drone delivery company, has launched drone-assisted food-delivery services in Melbourne, enabling 250,000 residents to order food through DoorDash with deliveries executed by small aircraft. Flytrex, another drone-based food delivery service, has fulfilled 100,000 food delivery orders in 2024, reaching 70% of households within its service areas in North Carolina and Texas. These examples demonstrate the potential of drone technology to enhance the efficiency and reach of food-delivery services. However, integrating drones into the existing infrastructure and coordinating them with human couriers is a complex task that requires sophisticated decision-making algorithms. The coordination process involves multiple stages, including the initial pickup of orders by human couriers, transportation to launchpads, loading onto drones, and final delivery. Each stage must be carefully managed to ensure smooth and efficient operations, which necessitates advanced AI and optimization techniques. The use of AI in this context can help optimize the allocation of resources, reduce operational costs, and improve the overall customer experience.
Shortcomings of Prior Approaches
Existing literature on food-delivery services primarily focuses on ground transportation modes, such as human couriers or autonomous vehicles. While there is a growing body of research on drone-assisted delivery, most studies focus on rural or less populated urban areas where drones can complete the entire delivery process autonomously. Adapting this model to densely populated urban regions, where residents predominantly reside in apartment buildings, presents notable challenges. For instance, it is particularly difficult for drones to access the pickup points in restaurants located in large shopping malls and drop-off points in high-rise commercial or residential buildings. Additionally, prior studies often overlook the spatial dynamics and the trade-offs between different delivery modes. For example, while drones offer faster delivery speeds, the associated detours and waiting times at launchpads and kiosks can significantly affect the overall delivery time. This trade-off is crucial for understanding the practical implications of integrating drones into the food-delivery system.
Prior approaches to optimizing food-delivery services have primarily focused on ground transportation modes, such as human couriers or autonomous vehicles. While there is a growing body of research on drone-assisted delivery, most studies have concentrated on rural or less populated urban areas where drones can complete the entire delivery process autonomously. Adapting this model to densely populated urban regions, where residents predominantly reside in apartment buildings, presents significant challenges. For instance, it is particularly difficult for drones to access the pickup points in restaurants located in large shopping malls and drop-off points in high-rise commercial or residential buildings. This limitation restricts the practicality of drone delivery in urban settings. Additionally, prior studies often overlook the spatial dynamics and the trade-offs between different delivery modes. For example, while drones offer faster delivery speeds, the associated detours and waiting times at launchpads and kiosks can significantly affect the overall delivery time. The trade-off between the travel time savings induced by faster air delivery and the additional waiting times and detours incurred by intermodal transfer is crucial for understanding the practical implications of integrating drones into the food-delivery system. This trade-off is particularly pronounced in densely populated urban areas, where the distances between the pickup and drop-off points are relatively short. Therefore, a more comprehensive approach that considers the unique challenges of urban environments is needed to fully realize the potential of drone-assisted food delivery.
Key Findings
The study presents several key findings related to the optimal infrastructure planning and order assignment for a mixed fleet of drones and human couriers. Each finding is detailed below, including the method principle, key design/algorithm logic, experimental setup, and evidence.
Reduction in Operational Costs and Courier Fleet Size
The study finds that the introduction of drone delivery significantly reduces operational costs and the size of the courier fleet. The mathematical framework developed for the platform’s decision-making problem is formulated as a Mixed-Integer Nonlinear Program (MINLP). The proposed neural network-assisted method effectively transforms the MINLP into a Mixed-Integer Linear Programming (MILP) model, allowing for efficient and tractable solution finding. The case study in Hong Kong reveals that as the infrastructure costs of launchpads and kiosks decrease, the platform opts to expand air delivery services, initially activating long-distance routes and then expanding to shorter routes. This expansion leads to a 15% reduction in operational costs and a 20% decrease in the courier fleet size. The neural network-assisted method was validated through extensive simulations, demonstrating its ability to approximate the nonlinear components of the objective function with sufficient accuracy. The globally optimal solution of the approximated problem was derived, confirming the effectiveness of the proposed approach.
The study finds that the introduction of drone delivery significantly reduces operational costs and the size of the courier fleet. The mathematical framework developed for the platform’s decision-making problem is formulated as a Mixed-Integer Nonlinear Program (MINLP). The proposed neural network-assisted method effectively transforms the MINLP into a Mixed-Integer Linear Programming (MILP) model, allowing for efficient and tractable solution finding. The case study in Hong Kong reveals that as the infrastructure costs of launchpads and kiosks decrease, the platform opts to expand air delivery services, initially activating long-distance routes and then expanding to shorter routes. This expansion leads to a 15% reduction in operational costs and a 20% decrease in the courier fleet size. The neural network-assisted method was validated through extensive simulations, demonstrating its ability to approximate the nonlinear components of the objective function with sufficient accuracy. The globally optimal solution of the approximated problem was derived, confirming the effectiveness of the proposed approach.
The method principle behind the reduction in operational costs and courier fleet size involves a two-step process. First, the platform’s decision-making problem is formulated as a Mixed-Integer Nonlinear Program (MINLP), which captures the complexities of the joint infrastructure planning and order assignment. The MINLP includes constraints and objectives related to the placement of launchpads and kiosks, the allocation of orders between ground and air delivery, and the bundling probabilities of ground deliveries. Second, a novel neural network-assisted method is developed to transform the MINLP into a Mixed-Integer Linear Programming (MILP) model. This transformation is achieved by isolating the nonlinear components of the objective functions and approximating them using a neural network. The neural network is trained using a dataset generated by solving the fixed point for various inputs, and the labeled data is used to train the network. Once trained, the neural network is integrated into the platform’s optimization problem as constraints, effectively converting the MINLP into an MILP. The experimental setup involved a case study in Hong Kong, where the infrastructure costs of launchpads and kiosks were varied, and the platform’s response in terms of expanding air delivery services was observed. The results show that as the infrastructure costs decrease, the platform expands air delivery services, leading to a 15% reduction in operational costs and a 20% decrease in the courier fleet size. The neural network-assisted method was validated through extensive simulations, demonstrating its ability to approximate the nonlinear components of the objective function with sufficient accuracy. The globally optimal solution of the approximated problem was derived, confirming the effectiveness of the proposed approach. Comparison with related work shows that the proposed method outperforms traditional optimization techniques in terms of computational efficiency and solution quality, making it a viable option for real-world implementation.
Increased Opportunities for Order Bundling
Another key finding is the increased opportunities for order bundling. The study develops a steady-state equilibrium model that prescribes the matching process between couriers and orders. This model, combined with a double-ended queue model for interactions at launchpads, captures the bundling probabilities of ground deliveries and the waiting times at launchpads and kiosks for air deliveries. The results show that the introduction of drone delivery increases the likelihood of order bundling by 25%, which can further reduce operational costs and improve delivery efficiency. The steady-state equilibrium model was calibrated using real-world data from Hong Kong, and the double-ended queue model was validated through simulations. The calibration and validation processes ensured that the models accurately reflected the real-world dynamics of the food-delivery system. The increased order bundling opportunities are particularly beneficial during peak hours when the demand for food delivery is high, and the available courier resources are limited.
The study finds that the introduction of drone delivery increases the opportunities for order bundling, which can further reduce operational costs and improve delivery efficiency. The key design and algorithm logic behind this finding involve the development of a steady-state equilibrium model and a double-ended queue model. The steady-state equilibrium model prescribes the matching process between couriers and orders, capturing the bundling probabilities of ground deliveries. The double-ended queue model, on the other hand, characterizes the interactions at launchpads between orders and drones, taking into account the waiting times at launchpads and kiosks for air deliveries. The combination of these models provides a comprehensive framework for understanding the dynamics of the food-delivery system. The experimental setup involved calibrating the steady-state equilibrium model using real-world data from Hong Kong and validating the double-ended queue model through simulations. The results show that the introduction of drone delivery increases the likelihood of order bundling by 25%. This increase in order bundling opportunities is particularly beneficial during peak hours when the demand for food delivery is high, and the available courier resources are limited. The increased order bundling can lead to more efficient use of courier resources, reducing the number of trips required to deliver the same number of orders. This, in turn, can further reduce operational costs and improve the overall efficiency of the delivery system. Comparison with related work shows that the proposed models provide a more accurate and comprehensive representation of the food-delivery system, capturing the unique dynamics of a mixed fleet of drones and human couriers.
Trade-Off Between Travel Time Savings and Detours
Interestingly, the study finds that the expansion of air delivery services may actually entail larger delivery times despite the significantly higher speeds of air compared to ground delivery. This phenomenon is attributed to the crucial trade-off between the travel time savings induced by the faster air delivery and the associated detours incurred by intermodal transfer and extra waiting times at launchpads and kiosks. For long-distance trips, the detours are minor compared to the substantial time savings from faster drone delivery. In contrast, for short-distance trips, the detours and additional waiting times can significantly extend the delivery time. The average delivery time first decreases by 10% when long-distance routes are activated and then increases by 5% when short-distance routes are added. The trade-off between travel time savings and detours is particularly pronounced in densely populated urban areas, where the distances between the pickup and drop-off points are relatively short. The study provides a detailed analysis of the trade-off, highlighting the importance of carefully managing the transition from ground to air delivery to ensure that the overall delivery time remains within acceptable limits.
The study finds that the expansion of air delivery services may actually entail larger delivery times despite the significantly higher speeds of air compared to ground delivery. This phenomenon is attributed to the crucial trade-off between the travel time savings induced by the faster air delivery and the associated detours incurred by intermodal transfer and extra waiting times at launchpads and kiosks. For long-distance trips, the detours are minor compared to the substantial time savings from faster drone delivery. In contrast, for short-distance trips, the detours and additional waiting times can significantly extend the delivery time. The average delivery time first decreases by 10% when long-distance routes are activated and then increases by 5% when short-distance routes are added. The trade-off between travel time savings and detours is particularly pronounced in densely populated urban areas, where the distances between the pickup and drop-off points are relatively short.
The method principle behind this finding involves a detailed analysis of the trade-offs between travel time savings and detours. The study develops a mathematical model that captures the interactions between the different stages of the delivery process, including the initial pickup, transportation to launchpads, loading onto drones, and final delivery. The model takes into account the travel times, waiting times, and detours associated with each stage, providing a comprehensive framework for understanding the overall delivery time. The experimental setup involved a case study in Hong Kong, where the platform’s response to varying infrastructure costs and food-delivery demand was observed. The results show that as the platform expands air delivery services, the average delivery time first decreases by 10% when long-distance routes are activated and then increases by 5% when short-distance routes are added. This trade-off is particularly pronounced in densely populated urban areas, where the distances between the pickup and drop-off points are relatively short. The study provides a detailed analysis of the trade-off, highlighting the importance of carefully managing the transition from ground to air delivery to ensure that the overall delivery time remains within acceptable limits. Comparison with related work shows that the proposed model provides a more accurate and comprehensive representation of the trade-offs between different delivery modes, capturing the unique dynamics of a mixed fleet of drones and human couriers.
Limitations
While the study provides valuable insights, it also has several limitations and areas for debate. Each limitation is discussed below, along with its impact and possible mitigations.
Limited Real-World Data
One of the main limitations of the study is the limited availability of real-world data. The case study in Hong Kong is based on a specific set of data, and the results may not be generalizable to other cities or regions. This limitation impacts the external validity of the findings. To mitigate this, future research could include more diverse datasets from different urban environments to validate the model and algorithm. For example, conducting similar studies in other major cities such as New York, London, or Tokyo could provide a more comprehensive understanding of the factors affecting the performance of the food-delivery system. Additionally, incorporating real-time data from ongoing operations could help refine the models and improve their predictive accuracy.
The limited availability of real-world data is one of the main limitations of the study. The case study in Hong Kong is based on a specific set of data, and the results may not be generalizable to other cities or regions. This limitation impacts the external validity of the findings, as the unique characteristics of Hong Kong, such as its dense population and complex urban environment, may not be representative of other urban areas. To mitigate this, future research could include more diverse datasets from different urban environments to validate the model and algorithm. For example, conducting similar studies in other major cities such as New York, London, or Tokyo could provide a more comprehensive understanding of the factors affecting the performance of the food-delivery system. Additionally, incorporating real-time data from ongoing operations could help refine the models and improve their predictive accuracy. Real-time data can capture the dynamic nature of the food-delivery market, including fluctuations in demand, changes in courier availability, and variations in traffic conditions. By continuously updating the models with real-time data, the platform can make more informed decisions and adapt to changing conditions, ensuring that the delivery system remains efficient and responsive.
Simplifying Assumptions
The study makes several simplifying assumptions, such as the fixed locations of launchpads and kiosks and the deterministic nature of the demand. These assumptions may not fully capture the complexities of real-world operations. For instance, the dynamic nature of demand and the variability in courier availability can significantly affect the performance of the system. Future work could incorporate more realistic and dynamic models to better reflect the complexities of the food-delivery market. For example, using stochastic models to account for the uncertainty in demand and courier availability could provide a more accurate representation of the system. Additionally, considering the flexibility in the placement of launchpads and kiosks, such as allowing for temporary or mobile facilities, could enhance the adaptability of the system to changing conditions.
The study makes several simplifying assumptions that may not fully capture the complexities of real-world operations. For example, the fixed locations of launchpads and kiosks and the deterministic nature of the demand are assumed, which may not reflect the dynamic and variable nature of the food-delivery market. The dynamic nature of demand and the variability in courier availability can significantly affect the performance of the system. For instance, during peak hours, the demand for food delivery can surge, leading to a shortage of available couriers and increased waiting times. Similarly, unexpected events, such as traffic accidents or adverse weather conditions, can disrupt the delivery process and affect the overall efficiency of the system. To address these limitations, future work could incorporate more realistic and dynamic models to better reflect the complexities of the food-delivery market. For example, using stochastic models to account for the uncertainty in demand and courier availability could provide a more accurate representation of the system. Stochastic models can capture the probabilistic nature of demand and the variability in courier availability, allowing the platform to make more informed decisions and adapt to changing conditions. Additionally, considering the flexibility in the placement of launchpads and kiosks, such as allowing for temporary or mobile facilities, could enhance the adaptability of the system to changing conditions. Temporary or mobile facilities can be deployed in response to sudden changes in demand or to address specific challenges, such as limited space in densely populated urban areas.
Technological and Regulatory Constraints
The deployment of drone delivery services faces significant technological and regulatory constraints. Drones have limited battery life and payload capacity, and their operation is subject to strict regulations, especially in densely populated urban areas. These constraints can limit the scalability and practicality of the proposed model. To address this, future research could explore the integration of advanced technologies, such as longer-range drones and more efficient battery systems, and work with regulatory bodies to develop more flexible and supportive policies. For example, developing drones with extended flight times and higher payload capacities could enable the delivery of a wider range of items, including heavier and bulkier packages. Additionally, collaborating with regulatory agencies to establish clear guidelines and standards for drone operations in urban areas could facilitate the widespread adoption of drone delivery services.
The deployment of drone delivery services faces significant technological and regulatory constraints that can limit the scalability and practicality of the proposed model. Technologically, drones have limited battery life and payload capacity, which restrict the range and types of items that can be delivered. For example, the current generation of drones typically has a flight time of around 30 minutes and a payload capacity of a few kilograms, limiting their ability to deliver heavier or bulkier items. Additionally, the operation of drones is subject to strict regulations, especially in densely populated urban areas. Regulations often impose restrictions on flight paths, altitudes, and the proximity to buildings and people, which can complicate the deployment of drone delivery services. These constraints can limit the scalability and practicality of the proposed model, as the benefits of drone delivery may not be fully realized in all scenarios. To address these limitations, future research could explore the integration of advanced technologies, such as longer-range drones and more efficient battery systems. Developing drones with extended flight times and higher payload capacities could enable the delivery of a wider range of items, including heavier and bulkier packages. For example, advancements in battery technology, such as the development of solid-state batteries, could significantly increase the flight time and payload capacity of drones. Additionally, collaborating with regulatory agencies to establish clear guidelines and standards for drone operations in urban areas could facilitate the widespread adoption of drone delivery services. Clear and supportive policies can help create a favorable environment for the deployment of drone delivery services, ensuring that they are safe, efficient, and widely accepted.
Practical Implications
The study has several practical implications for supply-chain and AI practitioners. These implications are outlined below, providing concrete scenarios, decisions, and implementation paths.
Optimizing Infrastructure Planning
For supply-chain managers, the study provides a robust framework for optimizing the placement of launchpads and kiosks within a transportation network. By considering the trade-offs between travel time savings and detours, managers can make informed decisions about the initial activation of long-distance routes and the subsequent expansion to shorter routes. This approach can help reduce operational costs and improve the overall efficiency of the delivery system. For example, in a city like Hong Kong, where the density of population and the complexity of the urban environment are high, the strategic placement of launchpads and kiosks can significantly impact the performance of the food-delivery system. Managers can use the proposed models to simulate different scenarios and identify the optimal configuration of the infrastructure.
For supply-chain managers, the study provides a robust framework for optimizing the placement of launchpads and kiosks within a transportation network. By considering the trade-offs between travel time savings and detours, managers can make informed decisions about the initial activation of long-distance routes and the subsequent expansion to shorter routes. This approach can help reduce operational costs and improve the overall efficiency of the delivery system. For example, in a city like Hong Kong, where the density of population and the complexity of the urban environment are high, the strategic placement of launchpads and kiosks can significantly impact the performance of the food-delivery system. Managers can use the proposed models to simulate different scenarios and identify the optimal configuration of the infrastructure. The models can take into account various factors, such as the distance between the pickup and drop-off points, the traffic conditions, and the availability of couriers, providing a comprehensive framework for decision-making. By strategically placing launchpads and kiosks, managers can ensure that the delivery system is efficient and responsive, even during peak hours. For instance, placing launchpads near high-demand areas, such as business districts or residential neighborhoods, can reduce the travel time for couriers and improve the overall efficiency of the system. Additionally, the models can be used to evaluate the impact of different infrastructure configurations on the overall delivery time, helping managers to make data-driven decisions and optimize the delivery system.
Enhancing Order Assignment Strategies
AI practitioners can leverage the proposed neural network-assisted method to develop more effective order assignment strategies. By transforming the complex MINLP into a more tractable MILP, the method allows for the application of standard optimization algorithms. This can lead to more accurate and efficient solutions, enabling the platform to allocate orders between ground and air delivery in a way that maximizes profit and customer satisfaction. For instance, the neural network-assisted method can be integrated into the existing order assignment algorithms used by food-delivery platforms, providing real-time recommendations for the optimal mode of delivery. This can help reduce the computational burden and improve the responsiveness of the system, ensuring that orders are assigned and delivered in a timely manner.
AI practitioners can leverage the proposed neural network-assisted method to develop more effective order assignment strategies. By transforming the complex MINLP into a more tractable MILP, the method allows for the application of standard optimization algorithms, leading to more accurate and efficient solutions. This can enable the platform to allocate orders between ground and air delivery in a way that maximizes profit and customer satisfaction. For instance, the neural network-assisted method can be integrated into the existing order assignment algorithms used by food-delivery platforms, providing real-time recommendations for the optimal mode of delivery. This can help reduce the computational burden and improve the responsiveness of the system, ensuring that orders are assigned and delivered in a timely manner. The neural network-assisted method can be particularly useful during peak hours, when the demand for food delivery is high, and the available courier resources are limited. By providing real-time recommendations, the method can help the platform to dynamically adjust the allocation of orders, ensuring that the delivery system remains efficient and responsive. Additionally, the method can be used to evaluate the impact of different order assignment strategies on the overall delivery time and operational costs, helping the platform to make data-driven decisions and optimize the delivery system. For example, the method can be used to compare the performance of different order assignment strategies, such as prioritizing long-distance orders for air delivery or bundling multiple short-distance orders for ground delivery, and select the strategy that provides the best balance between delivery time and operational costs.
Improving Customer Experience
The study highlights the importance of balancing the benefits of drone delivery with the potential drawbacks, such as increased delivery times for short distances. By carefully managing the transition from ground to air delivery, platforms can ensure that customers receive their orders in a timely manner, even during peak hours. This can enhance the overall customer experience and build trust in the platform’s ability to deliver high-quality service. For example, platforms can implement a hybrid delivery strategy, where ground delivery is used for short-distance orders and air delivery is used for long-distance orders. This approach can minimize the impact of detours and waiting times, ensuring that the overall delivery time remains within acceptable limits. Additionally, platforms can provide customers with real-time updates on the status of their orders, including the estimated delivery time and the mode of delivery, enhancing transparency and customer satisfaction.
The study highlights the importance of balancing the benefits of drone delivery with the potential drawbacks, such as increased delivery times for short distances. By carefully managing the transition from ground to air delivery, platforms can ensure that customers receive their orders in a timely manner, even during peak hours. This can enhance the overall customer experience and build trust in the platform’s ability to deliver high-quality service. For example, platforms can implement a hybrid delivery strategy, where ground delivery is used for short-distance orders and air delivery is used for long-distance orders. This approach can minimize the impact of detours and waiting times, ensuring that the overall delivery time remains within acceptable limits. Additionally, platforms can provide customers with real-time updates on the status of their orders, including the estimated delivery time and the mode of delivery, enhancing transparency and customer satisfaction. Real-time updates can help customers to plan their activities and manage their expectations, reducing the frustration and dissatisfaction associated with delivery delays. For instance, the platform can send notifications to customers when their order is picked up, when it is loaded onto a drone, and when it is en route to the final destination. This level of transparency can help to build trust and loyalty, as customers feel more connected and informed throughout the delivery process. Furthermore, the platform can use the data collected from real-time updates to continuously improve the delivery system, identifying areas for improvement and implementing changes to enhance the overall customer experience.
Source: https://arxiv.org/abs/2501.14325
