Although it is impossible to project with certainty what uses of drones will arrive in the future, or the performance characteristics and costs of future successful drones, we have identified six key areas where we believe advances in drone research may occur in the coming years in the operations research/network optimization literature.
Later City News: A paper published in 'Special Issue on Celebrating 50 Years of Networks: Part 1' highlighted future directions in drone routing research. The published article explains that optimization of routing problems using drones (unmanned aerial vehicles or UAVs) has become an important area of academic research.
The purpose of this article is to look to the future and help stimulate drone routing research in directions we hope will prove interesting and fruitful. This article discusses opportunities for better modeling of (1) drone capabilities for both existing drones and those likely to be used in the future, (2) constraints on drone performance and operations, (3) different objectives for various drone services, and (4) alternative delivery modes, as well as some areas for methodological advances and some possible new applications. While much of the research to date has leveraged existing TSP (traveling salesman problem), VRP (vehicle routing problem), and arc routing models, Stefan Poikonen and James F. Campbell authors the paper look forward to new contributions from drone research that use better models of more realistic drone types and new drone applications.
The paper concluded that there exists great promise in the development and expansion of drone routing applications, and with this great promise comes great uncertainty. Although it is impossible to project with certainty what uses of drones will arrive in the future, or the performance characteristics and costs of future successful drones, we have identified six key areas where we believe advances in drone research may occur in the coming years in the operations research/network optimization literature. In addition, we ask many questions that we believe may be fruitful for the research community to answer in the coming years. Despite the separation of ideas into six areas for the sake of exposition, these areas are certainly not disjoint and are in fact highly interdependent.
This is an exciting research area and we offer some broad summary suggestions for future research here. First, models should consider more details to better represent relevant aspects of the drone service provided (e.g., delivery, or surveillance), including the business objectives (e.g., cost), any operations with multiple drones, and any scheduling and coordination issues with other vehicles (including drones). Second, models should better reflect the drone range constraints, especially the fundamental range limits (time and/or distance) based on realistic energy modeling, and realistic flight profiles that include the vertical component and likely hovering. Third, models and solution methods should be developed to handle more of the stochastic aspects of drone operations (e.g., travel times, delivery times, etc.), especially due to weather conditions. Fourth, there should be more field testing of drones similar to those likely to be used in actual operations (e.g., for large‐scale drone delivery) and with realistic flight profiles (which depend on the environment, regulation, UTM, etc.). Fifth, given the current uncertainty about future drone designs and operations, and the corresponding lack of accurate data for particular settings, researchers should perform a sensitivity analysis with a very wide range of values (e.g., drone energy consumption rates) to ensure the robustness of their findings.
With the rapid growth in academic interest in drone operations, there is potential to accelerate drone applications if academics, regulatory bodies, and the private sector share information and ideas. We especially hope to see drone routing research motivated by actual or proposed real‐world drone applications in the public and private sector, so that the research can have a greater impact. We also hope to see more drone research that might inform policy (especially drone regulations), improve practice, impact people positively (e.g., with a positive impact on environmental and social systems) and advance theory. We encourage, in particular, the formation of a subgroup (e.g., within an INFORMS or Euro subgroup) that could coordinate OR approaches to drone problems, ideally by hosting interdisciplinary international meetings that include researchers from engineering, policy, drone companies, etc. Additionally, the creation of realistic standardized datasets may be extremely useful to researchers and practitioners alike to test future drone routing models, solutions, and algorithms. This includes a database of measured, reliable drone parameters (e.g., speed/energy consumption profiles, battery capacity, package-carrying capabilities, costs, emissions, etc.) that accurately reflect the drones used in various settings. Standard sets of test instances, ideally based on real applications and customer locations, will facilitate comparison between future models and help advance research.
Comments