Urban unmanned aerial vehicle logistics ground station dynamic layout method considering user requirements and time-space characteristics

Abstract

The invention discloses an urban unmanned aerial vehicle logistics ground station dynamic layout method considering user requirements and time-space characteristics, and relates to the technical field of urban governance, and the method comprises the steps: determining the layout principle of ground stations based on the types, levels and number ranges of the ground stations, and constructing a ground station layout scheme sketch; obtaining geographic information data, population demand data, alternative site data, unmanned aerial vehicle parameters and ground station parameters according to the ground station layout scheme sketch, and obtaining a mathematical optimization model by constructing an objective function and constraint conditions; based on the mathematical optimization model, solving by using an adaptive large neighborhood search algorithm, and confirming an optimal solving result; and analyzing according to a solving result, and evaluating the performance of the mathematical optimization model and optimizing parameters in combination with spatial distribution, time change and cost effectiveness of the ground station. According to the invention, the efficiency and benefit of the unmanned aerial vehicle logistics system are improved, the operation cost is reduced, and user demands can be responded more quickly.

Description

Technical Field

[0001] The present invention relates to the field of urban governance technology, and in particular to a method for dynamically arranging urban drone logistics ground stations taking into account user needs and spatiotemporal characteristics. Background Art

[0002] In the current rapidly developing urban logistics and transportation, traditional ground logistics systems face many challenges. Problems such as ground traffic congestion, low distribution efficiency and high operating costs are becoming increasingly prominent, especially during urban peak hours and in remote areas. In order to improve logistics efficiency, reduce costs and improve user experience, drone logistics systems have emerged. Drone logistics has brought new solutions to logistics and distribution with its advantages such as high speed and freedom from ground traffic restrictions.

[0003] However, there are still some significant shortcomings in existing drone logistics technology: 1) Unreasonable planning of ground stations: In the current drone logistics system, the location, number and type of ground stations often lack scientific optimization planning. This causes drones to fly longer distances during transportation, increasing transportation time and operating costs. At the same time, unreasonable ground station layout may also affect the delivery efficiency of drones, and they cannot give full play to their speed advantage.

[0004] 2) Insufficient sustainability of urban logistics: Although drone logistics can theoretically avoid ground traffic congestion and reduce pressure on urban roads, the layout and optimization of ground stations in existing systems often fail to fully consider this advantage. In addition, although drones are usually powered by electricity and have zero or low emissions, the lack of systematic planning and optimization strategies has prevented this environmental advantage from being fully utilized.

[0005] 3) User experience needs to be improved: Although drone logistics has obvious advantages in emergency delivery and delivery in remote areas, the existing system still has problems such as unstable delivery time and inflexible delivery methods. This affects the further improvement of user experience, especially in meeting the diverse needs of users. At the same time, the existing drone logistics system has not been able to provide sufficiently convenient and economical solutions for logistics services in remote areas.

[0006] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0007] In response to the problems in the related technology, the present invention proposes a dynamic layout method of urban drone logistics ground stations that takes into account user needs and spatiotemporal characteristics to overcome the above-mentioned technical problems existing in the existing related technology.

[0008] To this end, the specific technical solution adopted by the present invention is as follows: A method for dynamic layout of urban UAV logistics ground stations considering user needs and spatiotemporal characteristics includes the following steps: S1. Determine the layout principles of ground stations based on the types, levels and quantity range of ground stations, and construct a draft of the ground station layout plan; S2. According to the sketch of the ground station layout plan, obtain geographic information data, population demand data, candidate site data, drone parameters and ground station parameters; S3, based on geographic information data, population demand data, candidate site data, UAV parameters and ground station parameters, obtain a mathematical optimization model by constructing objective functions and constraints; S4. Based on the mathematical optimization model, the adaptive large neighborhood search algorithm is used to solve the problem. The ground station site selection and allocation scheme is optimized through the iterative destruction and repair operators to confirm the best solution. S5. Analyze the solution results and evaluate the performance of the mathematical optimization model and optimize the parameters based on the spatial distribution, temporal variation, and cost-effectiveness of the ground stations.

[0009] Optionally, the step of determining the layout principle of the ground stations based on the types, levels and quantity ranges of the ground stations and constructing a draft of the ground station layout plan comprises the following steps: S11. Determine the type of ground station based on the functions and application scenarios of the drone; S12. Divide the hierarchical structure of the ground station according to the function and positioning of the ground station, determine the ground stations of different levels, and determine the type and level of the ground station in combination with the type of the ground station; S13. Determine the number of ground stations and formulate the layout principles of ground stations based on the city size, population density, expected demand for drone logistics, and budget constraints, combined with the types and levels of ground stations; S14. Based on the layout principles of the ground station and by connecting with other transportation systems, a draft layout plan for the ground station is constructed.

[0010] Optionally, the other transportation systems include ground transportation and other drone systems, and the ground transportation includes roads, railways, and waterways.

[0011] Optionally, the obtaining of geographic information data, population demand data, candidate site data, drone parameters and ground station parameters according to the ground station layout plan sketch comprises the following steps: S21. Collect urban road network data, perform path analysis, and obtain geographic information data; S22, collect mobile communication base station data, social media check-in data, and commercial sales data and make predictions to obtain population demand data; S23, collecting location information for building ground stations, and dividing the locations according to location characteristics to obtain candidate site data; S24. Obtaining the parameters of the drone according to the maximum cruising range of the drone; S25. Obtain ground station parameters according to construction costs and capacities of different types of ground stations.

[0012] Optionally, collecting mobile communication base station data, social media check-in data, and commercial sales data and making predictions to obtain population demand data includes the following steps: S221, collecting mobile communication base station data, social media check-in data, and commercial sales data; S222. Use time series analysis and regression analysis to predict the population demand of grid units in different time periods and obtain population demand data.

[0013] Optionally, obtaining a mathematical optimization model by constructing an objective function and constraints based on geographic information data, population demand data, candidate site data, drone parameters and ground station parameters comprises the following steps: S31. Define symbols and variables based on geographic information data, population demand data, candidate site data, UAV parameters and ground station parameters; S32. Based on the defined symbols and variables, the objective function and constraint conditions are defined in the form of weighted summation to obtain a mathematical optimization model.

[0014] Optionally, the weighted summation expression is: ; Where MaxZ represents the total service capacity, n represents the number of population grid units, and m represents the total number of construction sites. represents the weight coefficient, represents the weight coefficient, T represents the set of all time periods, and t represents any time period in the set of all time periods. represents the population demand of grid cell i in time period t, represents the binary decision variable with respect to which grid cell i’s demand is served by location j at time period t, represents the cost of building a large transfer station at location j, represents the binary decision variable for building a large transfer station at location j, represents the cost of building a small distribution point at location j, represents the binary decision variable for building a small distribution point at location j, J1 represents the set of all locations that can be used to build a large transfer station, and J2 represents the set of all locations that can be used to build a small distribution point.

[0015] Optionally, the method of solving the problem based on a mathematical optimization model using an adaptive large neighborhood search algorithm, iteratively optimizing the ground station site selection and allocation scheme by using a destruction and repair operator, and confirming the best solution result comprises the following steps: S41. Based on the mathematical optimization model, an adaptive large neighborhood search algorithm is used to generate an initial solution according to the capacity and distance of the ground station, and a destruction operator and a repair operator are set; S42, initializing the score and usage count of each destruction operator and repair operator, and updating the score of the operator in each iteration in combination with the adaptive large neighborhood search algorithm; S43. Update the weights periodically according to the score ratios of the operators, and use the roulette wheel selection method to select the destruction operator and the repair operator according to the weights of the operators in each iteration to confirm the best solution result.

[0016] Optionally, the initializing the score and usage count of each destruction operator and repair operator, and updating the score of the operator in each iteration in combination with the adaptive large neighborhood search algorithm comprises the following steps: S421. If the solution result is better than the currently known global optimal solution, the highest score is assigned to the used destruction operator and repair operator, and the updated global optimal solution is confirmed; S422, if the solution is better than the current solution but not better than the known global optimal solution, assigning the second highest score to the used destruction and repair operators; S423. If the solution is not as good as the current solution but meets the acceptance criteria, the lowest score is assigned to the destruction and repair operators used.

[0017] Optionally, analyzing the solution results and evaluating the performance of the mathematical optimization model and optimizing the parameters in combination with the spatial distribution, temporal variation and cost-effectiveness of the ground stations comprises the following steps: S51. Draw maps and charts to show the locations of ground stations, analyze spatial distribution characteristics, count the number and proportion of ground stations of different types and levels, and evaluate the rationality of the results of ground station site selection; S52. Count the demand satisfaction rates in different time periods and regions, calculate the demand satisfaction rate index, and evaluate the distribution efficiency in combination with the average delivery time of the goods; S53. Analyze the usage of ground stations in different time periods, including changes in busyness and service areas, and evaluate the adaptability of the solution to time-varying needs; S54. Based on the evaluation results, the parameters of the mathematical optimization model are optimized.

[0018] The beneficial effects of the present invention are: 1. The present invention improves the efficiency and benefits of the UAV logistics system. Through scientific optimization algorithms, the location, number and type of ground stations are rationally planned, the transportation distance and time of UAVs are minimized, the distribution efficiency is improved, the operating costs are reduced, and user needs can be responded to more quickly.

[0019] 2. The present invention promotes the sustainable development of urban logistics. UAV logistics can effectively avoid ground traffic congestion and reduce the pressure on urban roads. By optimizing the layout of ground stations, the advantages of UAV logistics can be maximized and the problem of urban traffic congestion can be alleviated. At the same time, compared with traditional fuel vehicles, UAVs are usually powered by electricity and have the characteristics of zero or low emissions. By developing UAV logistics, urban air pollution can be reduced and environmental quality can be improved.

[0020] 3. The drone logistics in the present invention has the advantage of high speed, and can deliver goods to users faster, especially in emergency delivery and delivery in remote areas, shortening delivery time and improving user experience. Drones are not restricted by ground traffic conditions and can provide more flexible delivery methods, such as fixed-point delivery and precise delivery, to meet the diverse needs of users; in addition, drone logistics can effectively solve these problems, provide more convenient and economical logistics services for remote areas, and promote coordinated development of urban and rural areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 is a flow chart of a method for dynamic layout of urban drone logistics ground stations taking into account user needs and spatiotemporal characteristics according to an embodiment of the present invention; Figure 2 It is a structural block diagram of a method for dynamic layout of urban drone logistics ground stations taking into account user needs and spatiotemporal characteristics according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in the field should be able to understand other possible implementations and advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0024] According to an embodiment of the present invention, a method for dynamic layout of urban drone logistics ground stations taking into account user needs and spatiotemporal characteristics is provided.

[0025] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1 and Figure 2 As shown, according to one embodiment of the present invention, a method for dynamic layout of urban drone logistics ground stations considering user needs and spatiotemporal characteristics is provided, and the dynamic layout method includes the following steps: S1. Based on the type, level and quantity range of ground stations, determine the layout principles of ground stations and construct a sketch of the ground station layout plan.

[0026] In one embodiment, determining the layout principle of the ground stations based on the type, level and quantity range of the ground stations and constructing a ground station layout plan sketch comprises the following steps: S11. Determine the type of ground station based on the functions and application scenarios of the drone; S12. Divide the hierarchical structure of the ground station according to the function and positioning of the ground station, determine the ground stations of different levels, and determine the type and level of the ground station in combination with the type of the ground station; S13. Determine the number of ground stations and formulate the layout principles of ground stations based on the city size, population density, expected demand for drone logistics, and budget constraints, combined with the types and levels of ground stations; S14. Based on the layout principles of the ground station and by connecting with other transportation systems, a draft layout plan for the ground station is constructed.

[0027] It is necessary to explain that the type and level of the ground station is determined according to the functions and application scenarios of the drone. For example, it can include take-off and landing platforms, charging / swapping stations, maintenance and repair stations, cargo sorting / storage centers, etc. This patent mainly focuses on ground stations that serve as cargo transfer and distribution functions, namely large transfer stations and small distribution points.

[0028] The levels are divided into the hierarchical structure of ground stations to clarify the functions and positioning of ground stations at different levels, which can be divided into: A first-level hub station (a large transit station in this plan) is responsible for long-distance, large-volume cargo transportation and connects logistics networks between cities or regions. It is usually large in scale, well-equipped and has a wide range of services.

[0029] The secondary distribution station (the small distribution point in this plan) is responsible for the "last mile" delivery, serves specific communities or commercial areas, is relatively small in scale, and mainly carries out the rapid distribution of goods; combining types and levels, for example, there can be "first-level hub-type take-off and landing platform", "second-level distribution-type charging station", etc., to more finely describe the functions and characteristics of the ground station.

[0030] Based on factors such as city size, population density, expected demand for drone logistics, and budget constraints, a preliminary estimate of the number of ground stations of various types and levels that need to be built can be made. This can be done through simple estimates or empirical judgments. For example, a preliminary estimate can be made based on how many people are served by setting up a small distribution point. The existing logistics network can be used as a reference, such as the number and distribution of express outlets and distribution sites.

[0031] In addition, the layout principles for determining ground stations include: 1) Coverage principle: Ensure that ground stations can cover the main areas of the city and meet user needs to the greatest extent possible.

[0032] 2) Accessibility principle: The ground station should be set up in a location with convenient transportation to facilitate the take-off and landing of drones and the transportation of goods, and consider the connection with the existing transportation system, such as close to highway entrances and exits, urban expressways, bus hubs, etc.

[0033] 3) Economic principle: Minimize the construction and operation costs of ground stations while meeting service needs.

[0034] 4) Safety principle: The site selection of the ground station should meet safety standards and avoid sensitive areas such as densely populated areas, high-rise buildings, and airport clearance areas.

[0035] 5) Environmental friendliness principle: Try to choose locations that have less impact on the environment, such as using existing vacant land, roofs, etc.

[0036] 7) Coordination with existing urban planning: The choice of location needs to take into account factors such as the city's overall planning and land use planning to ensure that it is consistent with the direction of urban development.

[0037] In one embodiment, the other transportation systems include ground transportation and other drone systems, and the ground transportation includes roads, railways, and waterways.

[0038] It is necessary to explain that when considering the connection with other transportation systems, when considering the connection with ground transportation, the connection between the ground station and other transportation modes such as roads, railways, and waterways should be considered to achieve the coordinated operation of multiple transportation modes. For example, the first-level hub station can be set up near the freight railway station or port, and the second-level distribution station can be set up in the commercial area or public parking lot near the community; when connecting with other drone systems, consider the connection with other drone systems, such as the drone traffic management system (UTM), to achieve effective management of airspace and safe operation of drones.

[0039] In addition, according to the above principles and considerations, a preliminary ground station layout plan sketch is drawn. The sketch does not need to be very accurate, but it should clearly indicate the location and approximate service range of different types and levels of ground stations.

[0040] S2. According to the sketch of the ground station layout plan, obtain geographic information data, population demand data, alternative site data, UAV parameters and ground station parameters.

[0041] In one embodiment, obtaining geographic information data, population demand data, candidate site data, drone parameters and ground station parameters according to the ground station layout plan sketch includes the following steps: S21. Collect urban road network data, perform path analysis, and obtain geographic information data.

[0042] S22. Collect mobile communication base station data, social media check-in data, and commercial sales data and make predictions to obtain population demand data.

[0043] In one embodiment, the collecting of mobile communication base station data, social media check-in data, and commercial sales data and forecasting to obtain population demand data includes the following steps: S221, collecting mobile communication base station data, social media check-in data, and commercial sales data; S222. Use time series analysis and regression analysis to predict the population demand of grid units in different time periods and obtain population demand data.

[0044] S23. Collect location information for building a ground station, divide the location according to its characteristics, and obtain candidate site data.

[0045] S24. Obtaining the parameters of the drone according to the maximum cruising range of the drone.

[0046] S25. Obtain ground station parameters according to construction costs and capacities of different types of ground stations.

[0047] It should be explained that geographic information data: collects urban road network data (such as OpenStreetMap) to calculate the road network distance between grid cells and candidate sites , Use GIS software (such as ArcGIS, QGIS) for path analysis, Population demand data: Collect or predict grid unit population demand data for different time periods t . Data sources include but are not limited to: mobile communication base station data, social media check-in data, commercial sales data, etc. Use time series analysis, regression analysis and other methods to make predictions based on historical data. Alternative site data: Collect information on locations that can be used to build ground stations, including location, cost, area, etc. Based on the characteristics of the locations, they are divided into a set J1 where large transfer stations can be built and a set J2 where small distribution points can be built. Drone parameters: Determine the maximum range R of the drone. Determine the construction cost of different types of ground stations and capacity .

[0048] S3. Based on geographic information data, population demand data, candidate site data, UAV parameters and ground station parameters, a mathematical optimization model is obtained by constructing objective functions and constraints.

[0049] In one embodiment, the obtaining of the mathematical optimization model by constructing an objective function and constraints based on geographic information data, population demand data, candidate site data, drone parameters, and ground station parameters comprises the following steps: S31. Define symbols and variables based on geographic information data, population demand data, candidate site data, UAV parameters and ground station parameters; S32. Based on the defined symbols and variables, the objective function and constraint conditions are defined in the form of weighted summation to obtain a mathematical optimization model.

[0050] In one embodiment, the weighted summation expression is: ; Where MaxZ represents the total service capacity, n represents the number of population grid units, and m represents the total number of construction sites. represents the weight coefficient, represents the weight coefficient, T represents the set of all time periods, and t represents any time period in the set of all time periods. represents the population demand of grid cell i in time period t, represents the binary decision variable with respect to which grid cell i’s demand is served by location j at time period t, represents the cost of building a large transfer station at location j, represents the binary decision variable for building a large transfer station at location j, represents the cost of building a small distribution point at location j, represents the binary decision variable for building a small distribution point at location j, J1 represents the set of all locations that can be used to build a large transfer station, and J2 represents the set of all locations that can be used to build a small distribution point.

[0051] It should be explained that the symbols and variable definitions include: I: The set of all grid cells, with i representing any one of them (i=1, 2,…, n).

[0052] J: The set of all available locations, J=J1∪J2, with j representing any one of them.

[0053] J1: A collection of all locations that can be used to build large transfer stations.

[0054] J2: A collection of all locations that can be used to build small distribution points.

[0055] T: The set of all time periods, with t representing any one of them.

[0056] : Population demand of grid cell i in time period t.

[0057] : The road network distance from grid unit i to location j.

[0058] : The cost of building a large transfer station at location j.

[0059] : The cost of building a small distribution point at location j.

[0060] K1: The maximum number of large transfer stations allowed to be built.

[0061] K2: The maximum number of small distribution points allowed to be built.

[0062] : The capacity of location j as a large transfer station.

[0063] : The capacity of location j as a small distribution point.

[0064] R: The maximum endurance distance of the drone.

[0065] : binary decision variables, Indicates the construction of a large transfer station at location j, otherwise .

[0066] : binary decision variables, Indicates building a small distribution point at location j, otherwise .

[0067] : binary decision variables, Indicates that the demand of grid cell i is served by location j in time period t, otherwise .

[0068] α and β: weight coefficients used to balance the service population and construction costs.

[0069] Additionally, constraint definitions include: 1) The demand of each grid cell in each time period can only be served by one location: ; This holds for all i (i=1, 2, …, n) and t (t=1, 2, …, |T|).

[0070] 2) Only when location j is built can it provide services: ; This holds for all i (i=1, 2, …, n), j (j=1, 2, …, m), and t (t=1, 2, …, |T|).

[0071] 3) The number of locations of various types to be built cannot exceed the limit: ; ; 4) The demand for services cannot exceed the capacity of the location: ; This holds true for all j (j=1, 2, …, m).

[0072] 5) Maximum cruising range constraints: ; This holds for all i (i=1, 2, …, n), j (j=1, 2, …, m), and t (t=1, 2, …, |T|).

[0073] 6) Only one type of ground equipment can be built at one location: ; This holds true for all j (j=1, 2, …, m).

[0074] 7) The range of decision variables: , for all Established; , for all Established; , for all i (i=1, 2,…, n), j (j=1, 2,…, m), and t (t=1, 2,…, |T|).

[0075] S4. Based on the mathematical optimization model, the adaptive large neighborhood search algorithm is used to solve the problem. The ground station site selection and allocation plan are iteratively optimized through the destruction and repair operators to confirm the best solution.

[0076] In one embodiment, the method of solving the problem based on a mathematical optimization model using an adaptive large neighborhood search algorithm, iteratively optimizing the ground station site selection and allocation scheme by using a destruction and repair operator, and confirming the best solution result includes the following steps: S41. Based on the mathematical optimization model, an adaptive large neighborhood search algorithm is used to generate the initial solution according to the capacity and distance of the ground station, and the destruction operator and the repair operator are set.

[0077] S42, initializing the score and usage count of each destruction operator and repair operator, and updating the score of the operator in each iteration in combination with the adaptive large neighborhood search algorithm.

[0078] In one embodiment, the steps of initializing the score and usage count of each destruction operator and repair operator and updating the score of the operator in each iteration in combination with the adaptive large neighborhood search algorithm include the following steps: S421. If the solution result is better than the currently known global optimal solution, the highest score is assigned to the used destruction operator and repair operator, and the updated global optimal solution is confirmed; S422, if the solution is better than the current solution but not better than the known global optimal solution, assigning the second highest score to the used destruction and repair operators; S423. If the solution is not as good as the current solution but meets the acceptance criteria, the lowest score is assigned to the destruction and repair operators used.

[0079] S43. Update the weights periodically according to the score ratios of the operators, and use the roulette wheel selection method to select the destruction operator and the repair operator according to the weights of the operators in each iteration to confirm the best solution result.

[0080] It should be explained that ALNS is a meta-heuristic algorithm that searches the solution space by constantly destroying and repairing the current solution. It maintains a set of destruction operators and repair operators, and adaptively adjusts the selection probability according to their performance during the search process. Among them, the initial solution generation: a heuristic allocation method based on capacity and distance, including: 1) Randomly select K1 locations as large transfer stations and K2 locations as small distribution points.

[0081] 2) Initialize the remaining capacity of each ground station j .

[0082] 3) For each grid cell i and each time period t, calculate the distance from i to all selected ground stations j .

[0083] 4) Select the closest and with a remaining capacity greater than Ground station: ; If there is no ground station that meets the conditions, the demand is assigned to the nearest ground station and the solution is marked as an infeasible solution. .

[0084] 5) Update the remaining capacity of j*: .

[0085] In addition, the destruction operator is set. According to the ALNS algorithm, three operators are mainly set, including: 1) Removal Operator of Station Serving Fewest Grids: This operator is highly targeted and directly focuses on ground stations with low utilization, which helps optimize resource allocation. In drone logistics scenarios, the demand in different regions varies greatly, and some ground stations may not be highly utilized. This operator can effectively identify and remove these low-utilization sites, thereby exploring more efficient layout solutions in the subsequent repair process.

[0086] Algorithm steps: Determine the number of ground stations that need to be removed ,generally The value range of is [1, k1+k2], which can be a fixed value or a random value. Randomly select from the selected ground station set ground station, remove it from the solution, that is, the corresponding x j 1 and x j 2 The variable is set to 0, and the grid cells served by these removed ground stations need to be reallocated. For each affected grid cell i and each time period t, the allocation relationship These grid cells need to be reallocated later through the repair operator.

[0087] Let S be the set of selected ground stations and R be the set of randomly removed ground stations, then: ; For all j∈R: ; ; 2) Shaw Removal Operator: The Shaw Removal Operator takes into account the correlation between geographic location and demand, and can remove a group of related ground stations at one time, thereby causing a large disturbance to the solution and helping to escape the local optimal solution. In drone logistics, demands in similar areas often have similar time patterns. Using the Shaw operator can effectively adjust the ground station configuration in these areas and improve overall efficiency.

[0088] Algorithm steps: Randomly select a grid cell i.

[0089] Calculate the correlation R of all other grid cells i' with i ii' , which takes into account both geographical distance and demand similarity.

[0090] ; In the formula, max D represents the maximum distance between all grid cell pairs; P i represents the demand of i in all time periods; represents the correlation between grid cells i and i'; represents the road network distance between grid cells i and i'; represents the demand vector of grid unit i in all time periods; max P It represents the maximum value of the demand vector difference between all grid unit pairs, α represents the weight parameter of the distance factor, and β represents the weight parameter of the demand similarity factor.

[0091] in, The higher the value, the stronger the correlation between the two grid cells, and the more likely they are to be removed at the same time. It can be expressed as a vector, each element of which represents the demand of the grid unit in different time periods, max P This value is used to normalize the demand vector difference to between 0 and 1 to avoid the absolute value of the demand from having too much influence on the correlation calculation. PThe method is to traverse all possible pairs of grid cells, calculate the absolute value of the difference between their demand vectors, and find the maximum value; the larger the value of α, the greater the proportion of the distance factor in the correlation calculation. In other words, the closer the two grid cells are, the higher their correlation is. α is usually a value between 0 and 1; β is the weight parameter of the demand similarity factor. The larger the value of β, the greater the proportion of the demand similarity factor in the correlation calculation. In other words, the more similar the demand patterns of two grid cells are, the higher their correlation is. β is also usually a value between 0 and 1. α+β does not necessarily need to be equal to 1 and can be adjusted according to the specific problem.

[0092] Select the one with the highest correlation with i The ground stations serving these grid cells are then removed. If multiple grid cells are served by the same ground station, the ground station is removed only once.

[0093] 3) Time Period Removal Operator: Considering that this solution emphasizes "time-varying demand", the demand and distribution of ground stations will change in different time periods. The time period removal operator can be used to directly adjust the configuration of ground stations in a specific time period, such as removing ground stations that only provide services during peak hours, or merging and adjusting ground stations during off-peak hours. This helps the algorithm find a dynamic layout solution that adapts to the needs of different time periods.

[0094] Algorithm steps: Determine the number of time periods that need to be removed ,generally The value range of is [1, |T|].

[0095] Random Selection time period t.

[0096] For each selected time period t, remove all assignments between grid cell i and ground station j, i.e. If a ground station only provides service during these removed time periods, it can also be removed (that is, the corresponding x j 1 or x j 2 set to 0).

[0097] In addition, setting up the repair operator includes: Algorithm steps: For each grid cell i whose allocation relationship is removed by the corrupted operator and each time period t, calculate the distance from i to all ground stations j that have not been removed and have sufficient remaining capacity , select the nearest ground station .

[0098] If no ground station that meets the capacity requirement is found, the nearest ground station is selected and the solution is recorded as an infeasible solution, or a penalty measure is taken and a penalty term is added to the objective function. .

[0099] Update the remaining capacity of j*: .

[0100] Some mathematical expressions include choosing the closest ground station as: ; Where j* represents a set of all ground stations that have not been removed, argmin j Indicates S j remaining It is an array that stores the current remaining capacity of each ground station. During the repair process, as grid cells are assigned to ground stations, their remaining capacity will continue to decrease; is a three-dimensional matrix, which represents the distribution relationship between grid cells and ground stations. Indicates that in time period t, grid cell i is assigned to ground station j; It represents unassigned, which is the output of the repair operator and part of the ALNS algorithm solution.

[0101] (2) Greedy Insertion Operator: When inserting each grid cell, this operator selects the location that has the greatest improvement in the objective function after insertion. The objective function can be minimizing the total transportation distance, maximizing the service population, etc.

[0102] Algorithm steps: For each grid cell i and each time period t where the assignment relationship of the destroyed operator is removed: Traverse all ground stations j that have not been removed and have sufficient remaining capacity, and calculate the increment Δf of the objective function after assigning i to j ij .

[0103] Select the ground station that maximizes the improvement of the objective function , if the goal is to maximize some value, choose argmax.

[0104] If no ground station that meets the capacity requirement is found, the ground station that minimizes the increment of the objective function is selected and the solution is recorded as an infeasible solution, or a penalty measure is taken to add a penalty term to the objective function. .

[0105] Update the remaining capacity of j*: .

[0106] Part of the mathematical expression involves selecting the ground station that maximizes the improvement in the objective function, specifically: ; 3) Time Window Insertion Operator: If there is a time window constraint in the problem (for example, delivery needs to be completed within a specific time period), the time window constraint needs to be considered during the insertion process. This operator is used in conjunction with the above operator.

[0107] Algorithm steps (taking nearest neighbor insertion as an example): For each grid cell i and each time period t where the assignment relationship of the destroyed operator is removed: Calculate the distance from i to all ground stations j that have not been removed, have sufficient remaining capacity and meet the time window constraints .

[0108] Select the nearest ground station j*.

[0109] If no ground station that meets all requirements is found, the corresponding processing strategy is adopted (for example, delaying delivery, re-planning the route, etc., or recording the solution as an infeasible solution), and the subsequent steps are the same as the nearest neighbor insertion operator.

[0110] 4) Combined with the repair of the original allocation, the original allocation of each grid cell can be recorded before the destruction operator is executed. When repairing, you can try to restore the grid cell to its original assigned ground station first. If the original ground station is still available and has sufficient capacity, the allocation is restored directly; otherwise, other repair operators are used for allocation. This strategy helps to preserve the structure of the better solution.

[0111] In addition, the operator score and usage count are initialized. In the ALNS algorithm, each destroy operator and repair operator maintains a score and a usage count. This information is used to dynamically adjust the selection probability of the operator, so that the algorithm can adaptively select operators with better performance.

[0112] 1) Score initialization: All damage operators d i and the repair operator r j The initial scores are all set to the same value, usually 0 or 1, mathematical expression: Among them, S init is the initial score, which can be set to 1, for example.

[0113] 2) Initialize with count: all destruction operators d i and the repair operator r j The initial usage count is set to 0, the mathematical expression: .

[0114] 3) Weight initialization: The initial weights of all destruction operators d_i and repair operators r_j are also set to the same value. Usually, the initial weights of all operators are equal, so that at the beginning of the algorithm, the probability of each operator being selected is also equal. The mathematical expression is: or , |D| is the number of destruction operators, The first method is to repair the number of operators. The second method normalizes the weights so that the sum of the weights of all operators is 1, which is convenient for subsequent roulette wheel selection.

[0115] In the formula, w represents the weight corresponding to the operator, and D represents the set of all destruction operators.

[0116] In addition, when updating operators, in each iteration, the ALNS algorithm selects a destruction operator and a repair operator to operate on the current solution to generate a new solution. According to the quality of the new solution, the score of the operator used is updated, and the update score is based on the following three criteria: 1) Find the updated global optimal solution: If the new solution is better than the currently known global optimal solution, assign the highest score σ1 to the destruction and repair operators used.

[0117] 2) Find a better solution than the current one: If the new solution is better than the current one, but not better than the known global optimal solution, assign the next highest score σ2 to the destruction and repair operators used.

[0118] 3) Find an acceptable solution: If the new solution is not as good as the current one, but is still acceptable according to some acceptance criterion (such as the acceptance probability in simulated annealing), assign the lowest score σ3 to the destruction and repair operators used.

[0119] If the new solution is worse than the current solution and is not accepted, the operator score is not updated. σ1>σ2>σ3 usually holds, for example, it can be set to σ1=10, σ2=5, σ3=1.

[0120] Mathematical expression (score update): If the updated global optimal solution is found: ; If a better solution than the current one is found: ; If an acceptable solution is found: ; Use count update: Each time a destruction or repair operator is used, its use count is incremented by 1.

[0121] The mathematical expression is: ; In the formula, dj represents the jth destruction operator, rj represents the jth repair operator, s represents the score of the corresponding destruction operator or repair operator, di represents the ith destruction operator, represents the count of the i-th destruction operator, Represents the count of the i-th repair operator.

[0122] In addition, the operator weights are updated regularly. In order to enable the algorithm to adapt to changes in the search process, the weights of the operators need to be updated regularly (for example, every τ iterations) according to their scores. The commonly used update method is to update the weights according to the score ratio, and update the weight of each operator to its average score in the last τ iterations.

[0123] ; Among them, ρ is a reaction factor, usually in the range of (0, 1), which is used to control the magnitude of weight update. The larger the value, the more sensitive it is to recent performance, and vice versa, it is more dependent on historical performance. or To avoid the error of dividing by zero, you can set a smaller number instead of 0, such as 1. Or keep the original weight unchanged when count is 0. After updating the weight, you need to clear the score and count of each operator to prepare for the next round of statistics.

[0124] The mathematical expression for clearing is: ; During operator selection, in each iteration, a roulette wheel selection method is used to select a destruction operator and a repair operator according to the weights of the operators.

[0125] step: Compute the sum of all destruction operator weights: .

[0126] Generate a 0 to W d A random number between d .

[0127] Traverse all destruction operators and accumulate their weights until the accumulated sum is greater than or equal to r d , then select the current destruction operator. The selection method of the repair operator is similar.

[0128] Through the above steps, the ALNS algorithm can dynamically adjust its selection probability according to the historical performance of the operator, thereby searching the solution space more effectively and finding a better solution.

[0129] S5. Analyze the solution results and evaluate the performance of the mathematical optimization model and optimize the parameters based on the spatial distribution, temporal variation, and cost-effectiveness of the ground stations.

[0130] In one embodiment, the analysis based on the solution results and the evaluation of the performance of the mathematical optimization model and the optimization of parameters in combination with the spatial distribution, temporal variation and cost-effectiveness of the ground stations include the following steps: S51. Draw maps and charts to show the locations of ground stations, analyze spatial distribution characteristics, count the number and proportion of ground stations of different types and levels, and evaluate the rationality of the results of ground station site selection; S52. Count the demand satisfaction rates in different time periods and regions, calculate the demand satisfaction rate index, and evaluate the distribution efficiency in combination with the average delivery time of the goods; S53. Analyze the usage of ground stations in different time periods, including changes in busyness and service areas, and evaluate the adaptability of the solution to time-varying needs; S54. Based on the evaluation results, the parameters of the mathematical optimization model are optimized.

[0131] It should be explained that when analyzing the site selection results of ground stations, in the spatial distribution analysis, maps or charts are drawn to clearly show the location of the selected ground stations. Analyze the spatial distribution characteristics of ground stations, such as whether they are evenly distributed, whether they are concentrated in certain areas, and whether they are consistent with population density or commercial activity centers; in the type and level proportion analysis, the number and proportion of ground stations of different types (such as take-off and landing platforms, charging stations, maintenance stations) and levels (such as first-level hub stations, second-level distribution stations) are counted. Analyze whether the configuration of ground stations of different types and levels is reasonable and whether it meets the logistics needs at different levels. In the correlation analysis with existing facilities: analyze the correlation between the selected ground stations and existing transportation facilities (such as highways, railways, bus hubs) and logistics facilities (such as warehouses, distribution centers). Evaluate whether the scheme effectively utilizes existing resources and realizes the coordinated operation of multiple modes of transportation.

[0132] The needs fulfillment analysis includes: 1) In the demand fulfillment rate analysis, the demand fulfillment rates in different time periods and different regions are counted to evaluate whether the solution effectively meets the user's logistics needs.

[0133] ; Where Demand_Satisfaction_Rate indicates the degree to which the ground station layout plan satisfies user needs, Total_Delivered_Demand indicates the number of logistics that have been successfully delivered, and Total_Demand indicates the total number of logistics delivery demands of users.

[0134] 2) In the average delivery time analysis, the average delivery time of the goods is calculated to evaluate the delivery efficiency of the plan and whether it has achieved the expected service level.

[0135] ; In the formula, Average_Delivery_Time represents the average delivery time, Total_Delivery_Time_of_All_orders represents the total delivery time of all goods, and Total_Number_of_Orders represents the total number of goods.

[0136] 3) Analyze the differences in the satisfaction of different types of demands (such as emergency delivery, ordinary delivery, etc.) and evaluate whether the solution can effectively handle different types of demands.

[0137] In addition, when analyzing the time-varying characteristics, the usage of ground stations in different time periods is analyzed, such as busyness, service area, etc., to evaluate whether the solution effectively adapts to the time-varying needs.

[0138] Based on the above three analysis results, the performance of the model is comprehensively evaluated. According to the evaluation results, the parameters of the model are fine-tuned to further optimize the model calculation results.

[0139] In summary, with the help of the above technical solutions of the present invention, the present invention improves the efficiency and benefits of the drone logistics system. Through scientific optimization algorithms, the location, number and type of ground stations are reasonably planned to minimize the transportation distance and time of drones, improve distribution efficiency, reduce operating costs, and respond to user needs faster. The present invention promotes the sustainable development of urban logistics. UAV logistics can effectively avoid ground traffic congestion and reduce the pressure on urban roads. By optimizing the layout of ground stations, the advantages of drone logistics can be maximized to alleviate urban traffic congestion. At the same time, compared with traditional fuel vehicles, drones are usually powered by electricity and have the characteristics of zero or low emissions. By developing drone logistics, urban air pollution can be reduced and environmental quality can be improved. The drone logistics in the present invention has the advantage of fast speed, and can deliver goods to users faster, especially in emergency delivery and remote area delivery. It has obvious advantages, shortens delivery time, and improves user experience. UAVs are not restricted by ground traffic conditions and can provide more flexible delivery methods, such as fixed-point delivery and precise delivery, to meet the diverse needs of users. In addition, drone logistics can effectively solve these problems, provide more convenient and economical logistics services for remote areas, and promote coordinated development of urban and rural areas.

[0140] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for dynamic layout of urban UAV logistics ground stations considering user needs and spatiotemporal characteristics, characterized in that: The dynamic layout method includes the following steps: S1. Determine the layout principles of ground stations based on the types, levels and quantity range of ground stations, and construct a draft of the ground station layout plan; S2. According to the sketch of the ground station layout plan, obtain geographic information data, population demand data, candidate site data, drone parameters and ground station parameters; S3, based on geographic information data, population demand data, candidate site data, UAV parameters and ground station parameters, obtain a mathematical optimization model by constructing objective functions and constraints; S4. Based on the mathematical optimization model, the adaptive large neighborhood search algorithm is used to solve the problem. The ground station site selection and allocation scheme is optimized through the iterative destruction and repair operators to confirm the best solution. S5. Analyze the solution results and evaluate the performance of the mathematical optimization model and optimize the parameters based on the spatial distribution, temporal variation, and cost-effectiveness of the ground stations.

2. According to claim 1, a method for dynamic layout of urban drone logistics ground stations considering user needs and spatiotemporal characteristics is characterized in that: Determining the layout principle of the ground station based on the type, level and quantity range of the ground station and constructing a ground station layout plan sketch includes the following steps: S11. Determine the type of ground station based on the functions and application scenarios of the drone; S12. Divide the hierarchical structure of the ground station according to the function and positioning of the ground station, determine the ground stations of different levels, and determine the type and level of the ground station in combination with the type of the ground station; S13. Determine the number of ground stations and formulate the layout principles of ground stations based on the city size, population density, expected demand for drone logistics, and budget constraints, combined with the types and levels of ground stations; S14. Based on the layout principles of the ground station and by connecting with other transportation systems, a draft layout plan for the ground station is constructed.

3. According to claim 2, a method for dynamic layout of urban drone logistics ground stations considering user needs and spatiotemporal characteristics is characterized in that: The other transportation systems include ground transportation and other drone systems, and the ground transportation includes roads, railways, and waterways.

4. According to claim 1, a method for dynamic layout of urban drone logistics ground stations considering user needs and spatiotemporal characteristics is characterized in that: The step of obtaining geographic information data, population demand data, candidate site data, drone parameters and ground station parameters according to the ground station layout plan sketch includes the following steps: S21. Collect urban road network data, perform path analysis, and obtain geographic information data; S22, collect mobile communication base station data, social media check-in data, and commercial sales data and make predictions to obtain population demand data; S23, collecting location information for building ground stations, and dividing the locations according to location characteristics to obtain candidate site data; S24. Obtaining the parameters of the drone according to the maximum cruising range of the drone; S25. Obtain ground station parameters according to construction costs and capacities of different types of ground stations.

5. According to claim 4, a method for dynamic layout of urban drone logistics ground stations considering user needs and spatiotemporal characteristics is characterized in that: The collecting of mobile communication base station data, social media check-in data, and commercial sales data and forecasting to obtain population demand data includes the following steps: S221, collecting mobile communication base station data, social media check-in data, and commercial sales data; S222. Use time series analysis and regression analysis to predict the population demand of grid units in different time periods and obtain population demand data.

6. According to claim 1, a method for dynamic layout of urban drone logistics ground stations considering user needs and spatiotemporal characteristics is characterized in that: The method of obtaining a mathematical optimization model based on geographic information data, population demand data, candidate site data, drone parameters and ground station parameters by constructing an objective function and constraint conditions includes the following steps: S31. Define symbols and variables based on geographic information data, population demand data, candidate site data, UAV parameters and ground station parameters; S32. Based on the defined symbols and variables, the objective function and constraint conditions are defined in the form of weighted summation to obtain a mathematical optimization model.

7. The method for dynamic layout of urban UAV logistics ground stations considering user needs and spatiotemporal characteristics according to claim 6 is characterized in that: The expression of the weighted summation is: ； Where MaxZ represents the total service capacity, n represents the number of population grid units, and m represents the total number of construction sites. represents the weight coefficient, represents the weight coefficient, T represents the set of all time periods, and t represents any time period in the set of all time periods. represents the population demand of grid cell i in time period t, represents the binary decision variable with respect to which grid cell i’s demand is served by location j at time period t, represents the cost of building a large transfer station at location j, represents the binary decision variable for building a large transfer station at location j, represents the cost of building a small distribution point at location j, represents the binary decision variable for building a small distribution point at location j, J1 represents the set of all locations that can be used to build a large transfer station, and J2 represents the set of all locations that can be used to build a small distribution point.

8. According to claim 1, a method for dynamic layout of urban drone logistics ground stations considering user needs and spatiotemporal characteristics is characterized in that: The method is based on a mathematical optimization model, uses an adaptive large neighborhood search algorithm to solve, and iteratively optimizes the ground station site selection and allocation plan through destruction and repair operators to confirm the best solution, including the following steps: S41. Based on the mathematical optimization model, an adaptive large neighborhood search algorithm is used to generate an initial solution according to the capacity and distance of the ground station, and a destruction operator and a repair operator are set; S42, initializing the score and usage count of each destruction operator and repair operator, and updating the score of the operator in each iteration in combination with the adaptive large neighborhood search algorithm; S43. Update the weights periodically according to the score ratios of the operators, and use the roulette wheel selection method to select the destruction operator and the repair operator according to the weights of the operators in each iteration to confirm the best solution result.

9. The method for dynamic layout of urban UAV logistics ground stations considering user needs and spatiotemporal characteristics according to claim 8 is characterized in that: Initializing the score and usage count of each destruction operator and repair operator, and updating the score of the operator in each iteration in combination with the adaptive large neighborhood search algorithm includes the following steps: S421. If the solution result is better than the currently known global optimal solution, the highest score is assigned to the used destruction operator and repair operator, and the updated global optimal solution is confirmed; S422. If the solution is better than the current solution but not better than the currently known global optimal solution, assign the second highest score to the used destruction and repair operators; S423. If the solution is not as good as the current solution but meets the acceptance criteria, the lowest score is assigned to the destruction and repair operators used.

10. The method for dynamic layout of urban UAV logistics ground stations considering user needs and spatiotemporal characteristics according to claim 1, characterized in that: The analysis based on the solution results, combined with the spatial distribution, time variation and cost-effectiveness of the ground stations, evaluating the performance of the mathematical optimization model and optimizing the parameters includes the following steps: S51. Draw maps and charts to show the locations of ground stations, analyze spatial distribution characteristics, count the number and proportion of ground stations of different types and levels, and evaluate the rationality of the results of ground station site selection; S52. Count the demand satisfaction rates in different time periods and regions, calculate the demand satisfaction rate index, and evaluate the distribution efficiency in combination with the average delivery time of the goods; S53. Analyze the usage of ground stations in different time periods, including changes in busyness and service areas, and evaluate the adaptability of the solution to time-varying needs; S54. Based on the evaluation results, the parameters of the mathematical optimization model are optimized.

Images