An intelligent scheduling method and system for air-ground collaborative distribution of emergency supplies

Abstract

The application discloses an emergency material ground-air cooperative distribution intelligent scheduling method and system, which firstly acquires the state information and external environment information of a UAV and an unmanned transport vehicle in real time; a scheduling scheme is established by using the acquired information, and a genetic algorithm is adopted to optimize the scheduling scheme for multiple generations; the optimal task allocation and path length planning are calculated by comprehensively considering multiple performance indexes such as distribution time, transportation cost and path length; the unmanned transport vehicle transports the materials from a material reserve center to a gathering point or a temporary resettlement point in the periphery of a disaster area according to the optimized scheduling scheme; and the UAV accurately delivers the materials from the gathering point or the temporary resettlement point to a target area. The two-stage mode of 'unmanned transport vehicle distribution + UAV end delivery' is adopted to execute the distribution task, the scheduling scheme suitable for the environment of the disaster area can be accurately formulated and efficiently executed, and the timeliness and accuracy of the emergency material distribution in the earthquake disaster area are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of collaborative delivery and intelligent scheduling technology of drones and unmanned transport vehicles, specifically to an intelligent scheduling method and system for the ground-air collaborative delivery of emergency supplies using drones and unmanned transport vehicles in earthquake-stricken areas. Background Technology

[0002] In earthquake-stricken areas, the timely and accurate delivery of emergency supplies is directly related to the safety of affected people, the efficiency of rescue work, and the initial guarantee for post-disaster reconstruction. Therefore, intelligent scheduling of emergency supply delivery in earthquake-stricken areas is crucial. Traditional earthquake-stricken area supply delivery relies heavily on manual scheduling combined with a single mode of transport. This method has many shortcomings. On the one hand, manual scheduling is greatly affected by subjective and objective factors such as the experience of dispatchers and the timeliness of information access to the disaster area, which can easily lead to problems such as unreasonable route planning and improper allocation of material priorities, resulting in delivery delays or waste of resources. On the other hand, a single mode of transport has obvious limitations. Relying solely on vehicle transport is easily restricted by road conditions such as road collapses and aftershocks, while relying solely on drone transport is constrained by factors such as limited load capacity and insufficient range, making it difficult to meet the multi-point, dispersed, and diverse emergency supply needs of the disaster area, and failing to achieve overall coordination and dynamic optimization of supply delivery.

[0003] With the development of drone and unmanned vehicle technologies, ground-air collaborative delivery models are gradually being applied to the field of emergency material transportation. However, existing collaborative scheduling methods still have certain limitations. For example, general ground-air collaborative scheduling models are not optimized for the special scenarios of earthquake-stricken areas and do not fully adapt to the core characteristics of disaster areas, such as sudden changes in road conditions, unstable communication signals, and large differences in the priority of material needs. This results in insufficient matching between scheduling plans and actual rescue needs. At the same time, some models have weak capabilities in integrating and processing real-time status of drones and unmanned vehicles and dynamic environmental information in disaster areas. Under complex conditions such as extreme weather and local communication interruptions, scheduling accuracy and response speed will decrease significantly, making it difficult to reliably and stably complete the collaborative delivery of various types of materials such as medicines, food, and rescue equipment.

[0004] Therefore, there is a need for an intelligent scheduling method and system for ground-air collaborative delivery that can adapt to the complex environment of earthquake-stricken areas and take into account the characteristics of transportation units and the priority of material demand. This is to solve the problems of low efficiency, poor adaptability and insufficient reliability in the existing scheduling methods, and to meet the core needs of earthquake-stricken areas for rapid, accurate and efficient material delivery in emergency rescue. Summary of the Invention

[0005] To achieve the above objectives, the present invention provides an intelligent scheduling method and system for the coordinated air-ground delivery of emergency supplies using drones and unmanned transport vehicles in earthquake-stricken areas.

[0006] The present invention adopts the following technical solution:

[0007] A method for intelligent scheduling of emergency supplies' ground-air coordinated delivery includes the following steps:

[0008] S1. Real-time acquisition of status information of drones and unmanned transport vehicles, as well as external environmental information such as road conditions, weather, traffic conditions and disaster areas;

[0009] S2. Use the information obtained in step S1 to establish a scheduling scheme, and use a genetic algorithm to optimize the scheduling scheme through selection, crossover and mutation operations for multiple generations.

[0010] S3. Taking into account multiple performance indicators such as delivery time, transportation cost, and route length, calculate the optimal scheduling scheme that includes task allocation and route length planning;

[0011] S4. According to the optimized scheduling plan, unmanned transport vehicles use restored roads or cleared areas to transport supplies from the material storage center to the assembly point or temporary resettlement point on the outskirts of the disaster area.

[0012] S5. Based on optimized path length planning and scheduling schemes, drones can accurately deliver supplies from assembly points or temporary settlement points to target areas.

[0013] Furthermore, the genetic algorithm employs selection, crossover, and mutation operations, and by evaluating the fitness function, it comprehensively considers delivery time, transportation costs, and path length, enabling delivery tasks to be completed collaboratively in the complex environment of disaster areas.

[0014] Furthermore, the selection operation is used to select the most suitable individual from the current population to pass on to the next generation, and then select parent individuals proportionally based on the fitness values ​​of different individuals;

[0015] Crossover operations are used to generate new individuals. Offspring are produced by exchanging the genetic information of parent individuals. Demand points are allocated to different unmanned transport vehicles and different drones by exchanging the task allocation gene segments of parent individuals. The path sequence gene segment is the order and path segment of unmanned transport vehicles visiting the assembly point, and the order and flight path segment of drones visiting the terminal delivery point.

[0016] Mutation operations are used to introduce randomness and increase diversity. These include exchanging, inserting, and reversing the sequence of unmanned transport vehicles visiting assembly points or drone delivery sequences to bypass congested or collapsed road sections or shorten flight routes; resource mutations that change a demand point from one unmanned transport vehicle to another or a delivery point from one drone to another; and reselecting handover points from candidate assembly points to adapt to handover point mutations caused by changes in accessibility due to road conditions or weather identified in step S1.

[0017] Furthermore, when performing multi-generation optimization on the scheduling scheme, the scheduling scheme in the initial planning or environmental stabilization phase is based on a comprehensive objective function. The fitness of individuals in a genetic algorithm is used for multi-generation optimization; when the external environment changes or during the real-time optimization phase in a dynamic environment, the fitness of the scheduling scheme after changes in the external environment is comprehensively considered. As a genetic algorithm, the individual fitness of the genetic algorithm is used to re-evaluate and update the population, and obtain a scheduling scheme that is adapted to the current environment.

[0018] Furthermore, the comprehensive objective function is specifically as follows:

[0019]

[0020] in, These are the weighting coefficients for delivery time, transportation cost, and route length, respectively. The total delivery time of the scheduling plan Delivery time metrics after normalization The total transportation cost of the scheduling plan The transportation cost indicator after normalization. It refers to the path length of the scheduling scheme. The path length index after normalization.

[0021] Furthermore,

[0022]

[0023]

[0024]

[0025] in, It is the total delivery time of the scheduling plan. This refers to the delivery time of the unmanned transport vehicle. It refers to the delivery time of the drone. It is the total transportation cost of the scheduling plan. It's the transportation cost of the unmanned transport vehicle. The cost is the transportation cost of the drone. It is the path length of the scheduling scheme. It is the path length of the unmanned transport vehicle. It is the path length of the drone.

[0026] Furthermore,

[0027]

[0028]

[0029] in, The unmanned transport vehicle completed the task. Delivery time, The unmanned transport vehicle completed the task. Delivery time Indicates the number of unmanned transport vehicle missions; The drone completed the mission. Delivery time, The drone completed the mission. Delivery time Indicates the number of drone missions;

[0030]

[0031]

[0032] in, The unmanned transport vehicle completed the task. Fuel consumption and transportation costs It is a task and tasks The distance between them The unmanned transport vehicle completed the task. Other transportation costs during the execution process, Indicates the number of unmanned transport vehicle missions. The drone completed the mission. Fuel consumption and transportation costs It is a task and tasks The distance between them The drone completed the mission. Other transportation costs during the execution process, Indicates the number of drone missions;

[0033]

[0034]

[0035] in, It is the [number]th [unit] in the unmanned transport vehicle scheduling scheme, arranged in the order of execution. Each delivery node This task involves calculating the road network map of the disaster area based on the real-time road conditions and traffic status information obtained in step S1. and tasks The shortest passable path length or the minimum travel cost It is the [number]th drone in the drone scheduling scheme arranged in execution order. Each delivery node This refers to the task calculated on the road network map of the disaster area based on the real-time weather information and airspace constraint information obtained in step S1. and tasks The shortest possible flight path length or the minimum travel cost.

[0036] Furthermore, the adaptability of the scheduling scheme after changes in the external environment are comprehensively considered. The calculation formula is:

[0037]

[0038]

[0039] in, The total delivery time of the scheduling plan Delivery time metrics after normalization The total transportation cost of the scheduling plan The transportation cost indicator after normalization. It refers to the path length of the scheduling scheme. The path length index after normalization These are the weighting coefficients for delivery time, transportation cost, and route length, respectively. It is an environmental factor. It is the adjustment coefficient for the impact of traffic on the objective function. It is an adjustment factor for the influence of weather on the objective function. It is a factor influencing traffic conditions. It is an influencing factor of weather conditions. It is the overall impact adjustment coefficient. It is an influencing factor of other external factors.

[0040] This application also provides an intelligent dispatching system for the coordinated ground-air delivery of emergency supplies, including:

[0041] The data acquisition module acquires real-time status information of drones and unmanned transport vehicles, as well as external environmental information such as road conditions, weather, traffic conditions, and the disaster area.

[0042] The genetic algorithm module uses information acquired by the data acquisition module to establish a scheduling scheme. It then uses a genetic algorithm to optimize the scheduling scheme through multiple generations of selection, crossover, and mutation operations.

[0043] The real-time optimization module comprehensively considers multiple performance indicators such as delivery time, transportation cost, and route length to calculate the optimal task allocation and route length planning.

[0044] The execution control module, through real-time feedback and monitoring system, directs unmanned transport vehicles to transport supplies from the material storage center to assembly points or temporary resettlement points on the periphery of the disaster area according to the optimized scheduling plan, using restored roads or cleared areas.

[0045] In the delivery module, drones, based on optimized path length planning and scheduling schemes, accurately deliver supplies from assembly points or temporary settlement points to target areas.

[0046] Furthermore, the genetic algorithm module includes a fitness function design module, which continuously adjusts the scheduling scheme through iterative optimization to achieve the optimization goal of minimizing delivery time, transportation costs, and path length.

[0047] Beneficial effects:

[0048] This invention presents an intelligent scheduling method and system for the coordinated air-ground delivery of emergency supplies using drones and unmanned transport vehicles in earthquake-stricken areas. It utilizes sensors and IoT devices to collect multi-dimensional data in real time, and combines this with a genetic algorithm to dynamically optimize the scheduling scheme. Employing a two-stage delivery model of "unmanned transport vehicle delivery + drone last-mile delivery," the system can accurately formulate and efficiently execute scheduling schemes adapted to the disaster area environment, effectively improving the timeliness and accuracy of emergency supply delivery in earthquake-stricken areas. Furthermore, through multi-generation optimization via selection, crossover, and mutation of the genetic algorithm, and dynamic fitness adjustment, it can adapt to changes in road conditions, weather, and other environmental factors in the disaster area in real time. Data archiving also preserves the entire process information, providing data support for delivery process review, scheduling algorithm iteration, and the development of emergency rescue plans for similar disasters, thus contributing to improved management and control of emergency relief supplies in earthquake-stricken areas. Attached Figure Description

[0049] Figure 1 This is a schematic flowchart of the method of the present invention;

[0050] Figure 2 This is a flowchart of the genetic algorithm of the present invention;

[0051] Figure 3 This is a schematic diagram of the system modules of the present invention;

[0052] Figure 4 This is a schematic diagram of the genetic algorithm of the present invention. Detailed Implementation

[0053] To make the technical means, creative features, objectives, and effects of this invention easier to understand, the following description is provided in conjunction with... Figures 1 to 4 Specific embodiments are described and illustrated.

[0054] Example 1

[0055] This implementation discloses an intelligent scheduling method for the ground-air collaborative delivery of emergency supplies using drones and unmanned transport vehicles in earthquake-stricken areas, such as... Figure 1 As shown, it includes the following steps:

[0056] Phase 1: Delivery by unmanned transport vehicles

[0057] S1. Data Acquisition and Aggregation: Unmanned transport vehicles and drones collect operational status and disaster area environmental information through their own positioning, power, fault diagnosis, and other status sensors, as well as environmental perception modules such as vehicle-mounted cameras, weather sensors, and airborne cameras. At the same time, they combine with the IoT network to obtain real-time external environmental information such as road conditions, weather, and traffic in the disaster area, including real-time road accessibility, road damage, traffic congestion, and traffic speed. This information is updated at fixed intervals or instantly when an event is triggered. The data is then input into the S2 genetic algorithm scheduling optimization steps to construct candidate scheduling scheme constraints and cost parameters and drive replanning.

[0058] S2. Using the operational status information and disaster area environmental information obtained in step S1, construct the feasibility constraints and cost parameters of the candidate scheduling schemes, and employ a genetic algorithm, such as... Figure 2 As shown, the genetic algorithm population consists of several individuals, each representing a candidate ground-to-air collaborative delivery scheduling scheme, obtained through selection, crossover, and mutation operations, which correspond to the comprehensive objective value and fitness value of the scheduling scheme. This refers to the complete decision set for ground-air coordinated delivery, including allocating various demand points to different unmanned transport vehicles and drones, as well as the driving paths and access order of unmanned transport vehicles, the flight paths and delivery order of drones, and the path length. This refers to the total path length index corresponding to the path planning in the scheduling scheme;

[0059] The selection operation is used to choose a better individual from the candidate scheduling schemes of the current population and pass it on to the next generation. For any individual... Based on the real-time status of the unmanned transport vehicle and drone, and the external environmental information of the disaster area collected in step S1, the delivery time, transportation cost, route length, and other indicators corresponding to the plan are calculated, and fitness is formed accordingly, with fitness values ​​for different individuals. The probability of an individual being selected is calculated using the following formula:

[0060]

[0061] in, Individual The probability of being selected, and It is the i-th or j-th individual. , Individuals and fitness value, It is the sum of the fitness values ​​of all individuals, where i and j both range from [1, n]. It is the number of individuals in the population.

[0062] Crossover operations generate new individuals. By exchanging the task allocation gene segments of the parent individuals, the demand points are assigned to different unmanned transport vehicles and different drones. The path sequence gene segment represents the order and path fragments of unmanned transport vehicles visiting the assembly point, and the order and flight path fragments of drones visiting the terminal delivery point. After crossover, feasibility verification and repair are performed based on the real-time state and environmental constraints of step S1, thereby ensuring that the generated offspring scheduling scheme is executable. The formula is:

[0063]

[0064] in, It is the crossover rate at the current delivery time t. It is the maximum crossover rate. It is a constant. It is population diversity.

[0065] The mutation operation is used to randomly fine-tune candidate scheduling schemes to increase diversity and avoid local optima. In the context of ground-air coordination in earthquake-stricken areas, mutations include exchanging, inserting, and reversing the sequence of unmanned transport vehicles accessing assembly points or drone delivery sequences to bypass congestion, landslides, or shorten flight routes; resource mutations such as shifting a demand point from one unmanned transport vehicle to another, or a delivery point from one drone to another; and reselecting handover points from candidate assembly points to accommodate handover point mutations caused by road blockages or weather deterioration identified in step S1. After mutation, constraints are checked and corrected based on the data from step S1 to ensure the new scheme is executable. The formula is:

[0066]

[0067] in, It is the rate of variation of the current delivery time t. It is the maximum mutation rate. It is the adjustment constant. It is the fitness value of the current population.

[0068] S3. In the complex environment of the disaster area, the optimal task allocation and path length planning are calculated by comprehensively considering multiple performance indicators such as delivery time, transportation cost, and path length.

[0069] Optimizing the scheduling plan requires considering how to reduce delivery time to improve efficiency. The formula for total delivery time is:

[0070]

[0071] in, This is the total delivery time of the unmanned transport vehicle scheduling plan. The unmanned transport vehicle completed the task. Delivery time, The unmanned transport vehicle completed the task. Delivery time This indicates the number of unmanned transport vehicle missions.

[0072]

[0073] in, This is the total delivery time of the drone dispatch plan. The drone completed the mission. Delivery time, The drone completed the mission. Delivery time This indicates the number of drone missions.

[0074] The optimization objective is to reduce transportation costs while ensuring delivery efficiency. The formula for total transportation cost is:

[0075]

[0076] in, It is the total transportation cost of the unmanned transport vehicle scheduling scheme. The unmanned transport vehicle completed the task. Fuel consumption and transportation costs It is a task and tasks The distance between them The unmanned transport vehicle completed the task. Other transportation costs during the execution process, This indicates the number of unmanned transport vehicle missions.

[0077]

[0078] in, It is the total transportation cost of the drone dispatching scheme. The drone completed the mission. Fuel consumption and transportation costs It is a task and tasks The distance between them The drone completed the mission. Other transportation costs during the execution process, This indicates the number of drone missions.

[0079] Path length The shortest passable path length or minimum travel cost is calculated in the disaster area road network by accumulating the information on road accessibility, damage level, and congestion collected in step S1 from adjacent delivery nodes. Based on this, the genetic algorithm avoids damaged roads and selects the best temporary passable roads. The formula for the path length is:

[0080]

[0081] in, It is the path length of the unmanned transport vehicle scheduling scheme. It is the [number]th [unit] in the unmanned transport vehicle scheduling scheme, arranged in the order of execution. Each delivery node This task involves calculating the road network map of the disaster area based on real-time road condition information such as road accessibility, road damage, traffic congestion, and traffic speed obtained in step S1. and tasks The shortest passable path length or the minimum passable cost.

[0082]

[0083] in, It is the path length of the drone scheduling scheme. It is the [number]th drone in the drone scheduling scheme arranged in execution order. Each delivery node This refers to the task calculated on the road network map of the disaster area based on the real-time weather information and airspace constraint information obtained in step S1, including temporary control information for no-fly zones and terrain obstacle risks. and tasks The shortest possible flight path length or minimum passage cost; when the weather deteriorates or a no-fly zone is in effect. The corresponding increase or determination of infeasibility will suppress the route segment in the genetic algorithm.

[0084]

[0085]

[0086]

[0087] in, It is the total delivery time of the scheduling plan. It is the total transportation cost of the scheduling plan. It is the path length of the scheduling scheme.

[0088] Normalize the time, cost, and path length respectively:

[0089]

[0090]

[0091]

[0092]

[0093] in, The total delivery time of the scheduling plan Delivery time metrics after normalization It is the total delivery time of the scheduling plan. It is the shortest delivery time. That is the longest delivery time. The total transportation cost of the scheduling plan The transportation cost indicator after normalization. It is the total transportation cost of the scheduling plan. It is the shortest transportation cost. It is the highest transportation cost. It refers to the path length of the scheduling scheme. The path length index after normalization It is the path length of the scheduling scheme. It is the shortest path length. It is the longest path length.

[0094] It is the comprehensive objective value function used by genetic algorithms to evaluate the quality of a complete scheduling scheme. These are dynamic weighting coefficients, specifically weighting coefficients for delivery time, transportation cost, and route length.

[0095] Real-time optimization steps include scheduling adjustments based on environmental and individual fitness assessments to ensure that the scheduling plan is updated in real time in dynamic environments. , , All data are calculated and mapped from real-time sensor or IoT monitoring data in step S1, and are obtained during the initial planning or relatively stable environmental phases, using a comprehensive objective function. The fitness of individuals in a genetic algorithm is used for multi-generation optimization. When the external environment changes or during the real-time optimization phase in a dynamic environment, the fitness of the scheduling scheme is comprehensively considered after changes in the external environment. The fitness of individuals in the genetic algorithm is used to re-evaluate and update the population, and to update and re-plan the scheduling scheme, thereby obtaining a scheduling scheme adapted to the current environment. The formula is as follows:

[0096]

[0097]

[0098] in, It comprehensively considers the adaptability of the scheduling scheme after changes in the external environment. When the environment is relatively stable, it uses a comprehensive objective value function. The fitness evaluation criterion for genetic algorithms is used for multi-generational optimization. When the external environment changes, the algorithm is optimized accordingly. As a fitness evaluation criterion during the real-time optimization phase, the scheduling scheme is updated and replanned. It is an environmental factor. It represents the intensity of traffic impact, an adjustment factor for the effect of traffic on the objective function. It is the intensity of weather impact, an adjustment factor for the influence of weather on the objective function. These are factors influencing traffic conditions, such as the road congestion index, road capacity, and road blockage probability. These are factors influencing weather conditions, such as wind speed level, rainfall or snowfall intensity, and visibility level. It is an adjustment factor, a scaling factor for the overall impact of "other external factors," used to balance their relative weighting to traffic or weather. Other external factors include communication disruption levels, aftershock risk levels, no-fly zones or temporary air traffic control information, and terrain obstacle risks.

[0099] S4. Based on the optimized scheduling plan, the unmanned transport vehicles utilize partially restored roads and cleared areas to transport emergency supplies from the material storage center to assembly points or temporary resettlement sites on the outskirts of the disaster area. Using the external environmental information data provided in step S1, the unmanned transport vehicles avoid damaged roads within the disaster area and select temporary roads with better accessibility to transport supplies to assembly points closer to the center of the disaster area.

[0100] Phase Two: Ground-Air Collaborative Delivery Using Unmanned Vehicles and Drones

[0101] S5. After the unmanned transport vehicles transport supplies to the assembly points or temporary resettlement points on the outskirts of the disaster area, they need to further distribute the supplies to areas that cannot be reached by ground transportation.

[0102] S6. The unmanned transport vehicle hands over the supplies to the drone, which then performs the "last mile" delivery according to an optimized scheduling plan. This scheduling plan includes information such as the drone's final path length planning and delivery sequence. This ensures that emergency supplies are accurately delivered to disaster-stricken areas, including ruins, mountains, high-rise buildings, or other severely damaged areas.

[0103] The genetic algorithm employs a fitness function design module, iteratively optimizing and continuously adjusting the scheduling scheme to minimize objectives such as delivery time, transportation cost, and path length. The formula is:

[0104] .

[0105] S7. Based on the mission assignment and the actual situation in the disaster area, the drones avoid obstacles, fly over mountains, ruins, etc., and quickly deliver urgently needed supplies from the assembly point to the places where they are most needed; the delivery target areas include high-rise ruins, disaster-stricken mountainous areas, or places where transportation is completely cut off.

[0106] In step S2, when optimizing the genetic algorithm, a dynamic adaptation strategy is adopted for the three core operations of selection, crossover, and mutation during the multi-generation optimization of the scheduling scheme. Reasonable randomness is introduced to avoid the algorithm getting trapped in local optima.

[0107] In step S7, after the drone delivers the goods accurately, there is also a data archiving and traceability step. This step saves relevant information such as the status information of the transportation unit, environmental monitoring data of the disaster area, genetic algorithm scheduling parameters, and two-stage delivery execution records to the emergency rescue database. The purpose is to provide data support for subsequent delivery process review, scheduling algorithm iteration optimization, and emergency dispatch in similar disaster scenarios.

[0108] This method consists of two main phases. The first phase utilizes unmanned transport vehicles to achieve "trunk line transportation" from the material storage center to the periphery of the disaster area. Firstly, a widely deployed sensor network and IoT network comprehensively perceive the system status and the disaster area environment. The unmanned transport vehicles and drones themselves integrate BeiDou positioning modules, power sensors, load sensors, and fault diagnosis modules, and have built-in LoRa, Wi-Fi-Mesh, and BeiDou short message modules. The unmanned transport vehicles are equipped with 360-degree cameras and weather sensors to collect real-time road conditions and weather information while driving. Before and after the delivery drones perform their missions, they can execute rapid reconnaissance routes, capturing videos and images with their onboard cameras to analyze road conditions in real time. Subsequently, a genetic algorithm is used to deeply fuse and optimize the collected multi-source data, generating an optimal ground delivery plan that balances time, cost, and route. Finally, the unmanned transport vehicles follow the plan, utilizing the partially restored road network to complete the initial forward delivery of materials.

[0109] The second phase achieves precise delivery of supplies for the "last mile" from the assembly point to the final disaster site. Unmanned transport vehicles and drones complete the handover of supplies at the assembly point; the drones, leveraging their aerial maneuverability, fly over complex areas where ground transportation is completely disrupted, following optimized aerial paths, and avoid dynamic obstacles through advanced airborne perception systems, ultimately delivering emergency supplies precisely to previously inaccessible disaster sites.

[0110] Throughout the process, the central scheduling system based on genetic algorithms served as the core "intelligent brain." It not only generated high-quality initial plans through multiple generations of evolution but also, by introducing environmental fitness factors, endowed the system with the ability to dynamically respond to the rapidly changing environment of the disaster area and optimize in real time. After the mission was completed, the archiving and storage of all process data provided valuable data support for the accumulation of rescue experience and the continuous iteration of algorithm models.

[0111] Example 2

[0112] like Figure 3 As shown, this embodiment discloses an intelligent dispatching system for the ground-air coordinated delivery of emergency supplies, including:

[0113] The data acquisition module acquires real-time status information of drones and unmanned transport vehicles, as well as road conditions, weather, traffic conditions and external environmental information in disaster areas through sensors and IoT devices. It adopts an edge computing architecture to clean, merge and compress the collected data locally, and then uploads it to the central dispatch platform through fixed period or event triggering to ensure the real-time performance and reliability of the data.

[0114] The genetic algorithm module is the core scheduling engine of the system, such as... Figure 4 As shown, a scheduling scheme is established using information obtained from the data acquisition module, and a genetic algorithm is used to optimize the scheduling scheme through selection, crossover, and mutation operations over multiple generations.

[0115] Taking into account three performance indicators—delivery time, transportation cost, and route length—a fitness function is constructed, and each indicator is first normalized.

[0116]

[0117]

[0118]

[0119]

[0120] in, The total delivery time of the scheduling plan Delivery time metrics after normalization It is the total delivery time of the scheduling plan. It is the shortest delivery time. That is the longest delivery time. The total transportation cost of the scheduling plan The transportation cost indicator after normalization. It is the total transportation cost of the scheduling plan. It is the shortest transportation cost. It is the highest transportation cost. It refers to the path length of the scheduling scheme. The path length index after normalization It is the path length of the scheduling scheme. It is the shortest path length. It is the longest path length.

[0121] It is the comprehensive objective value function used by genetic algorithms to evaluate the quality of a complete scheduling scheme. These are dynamic weighting coefficients, specifically weighting coefficients for delivery time, transportation cost, and route length.

[0122] The real-time optimization module comprehensively considers multiple performance indicators such as delivery time, transportation cost, and route length to calculate the optimal task allocation and route length planning.

[0123] , , All data are calculated and mapped from real-time monitoring data of S1's sensors or IoT, using the following formula:

[0124]

[0125]

[0126] in, It is the adaptability of the scheduling scheme after comprehensively considering changes in the external environment. It is an environmental factor. It represents the intensity of traffic impact, an adjustment factor for the effect of traffic on the objective function. It is the intensity of weather impact, an adjustment factor for the influence of weather on the objective function. These are factors influencing traffic conditions, such as the road congestion index, road capacity, and road blockage probability. These are factors influencing weather conditions, such as wind speed level, rainfall or snowfall intensity, and visibility level. It is an adjustment factor, a scaling factor for the overall impact of "other external factors," used to balance their relative weighting to traffic or weather. Other external factors include communication disruption levels, aftershock risk levels, no-fly zones or temporary air traffic control information, and terrain obstacle risks.

[0127] The execution control module, through real-time feedback and monitoring system, directs unmanned transport vehicles to transport supplies from the material storage center to the assembly point or temporary resettlement point on the periphery of the disaster area according to the optimized scheduling plan, using restored roads or cleared areas. In communication blind spots or areas with weak signals, drones are used as aerial relay nodes to build an integrated air-ground communication network to ensure reliable transmission of commands and data.

[0128] The delivery module is responsible for executing a two-stage delivery task: "unmanned transport vehicle trunk line transportation + drone last-mile delivery". Based on the optimized path length planning and scheduling scheme, the drones accurately deliver materials from the assembly point or temporary settlement point to the target area.

[0129] The genetic algorithm module generates new task allocation schemes by selecting scheduling schemes with high fitness and performing crossover and mutation operations, and optimizes material delivery time, transportation cost, and path length in the collaborative work of unmanned transport vehicles and drones.

[0130] The genetic algorithm module includes a fitness function design module. Through iterative optimization, it continuously adjusts the scheduling scheme to minimize delivery time, transportation costs, and path length. The formula is:

[0131]

[0132] The real-time optimization module includes an environmental adaptability assessment module. When traffic congestion occurs or the weather deteriorates, the environmental adaptability module adjusts the weighting coefficients. This allows for appropriate adjustments to objectives such as delivery time, transportation costs, and route length, thereby optimizing the execution of delivery tasks.

[0133] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and modifications according to the present invention, but these corresponding changes and modifications should all fall within the protection scope of the appended claims.

Claims

1. A method for intelligent scheduling of ground-air coordinated delivery of emergency supplies, characterized in that, Includes the following steps: S1. Real-time acquisition of status information of drones and unmanned transport vehicles, as well as external environmental information such as road conditions, weather, traffic conditions and disaster areas; S2. Use the information obtained in step S1 to establish a scheduling scheme, and use a genetic algorithm to optimize the scheduling scheme through selection, crossover and mutation operations for multiple generations. S3. Taking into account multiple performance indicators such as delivery time, transportation cost, and route length, calculate the optimal scheduling scheme that includes task allocation and route length planning; S4. According to the optimized scheduling plan, unmanned transport vehicles use restored roads or cleared areas to transport supplies from the material storage center to the assembly point or temporary resettlement point on the outskirts of the disaster area. S5. Based on optimized path length planning and scheduling schemes, drones can accurately deliver supplies from assembly points or temporary settlement points to target areas.

2. The scheduling method according to claim 1, characterized in that, The genetic algorithm described employs selection, crossover, and mutation operations. By evaluating the fitness function and considering factors such as delivery time, transportation cost, and path length, it enables delivery tasks to be completed collaboratively in the complex environment of disaster areas.

3. The scheduling method according to claim 2, characterized in that, The selection operation is used to select the most suitable individual from the current population to pass on to the next generation, and then select parent individuals proportionally based on the fitness values ​​of different individuals. Crossover operations are used to generate new individuals. Offspring are produced by exchanging the genetic information of parent individuals. Demand points are allocated to different unmanned transport vehicles and different drones by exchanging the task allocation gene segments of parent individuals. The path sequence gene segment is the order and path segment of unmanned transport vehicles visiting the assembly point, and the order and flight path segment of drones visiting the terminal delivery point. Mutation operations are used to introduce randomness and increase diversity. These include exchanging, inserting, and reversing the sequence of unmanned transport vehicles visiting assembly points or drone delivery sequences to bypass congested or collapsed road sections or shorten flight routes; resource mutations that change a demand point from one unmanned transport vehicle to another or a delivery point from one drone to another; and reselecting handover points from candidate assembly points to adapt to handover point mutations that adapt to changes in accessibility caused by road conditions or weather changes identified in step S1.

4. The scheduling method according to claim 1, characterized in that, When performing multi-generation optimization on a scheduling scheme, the scheduling scheme, in the initial planning or environmental stabilization phase, uses a comprehensive objective function. The fitness of individuals in a genetic algorithm is used for multi-generation optimization; when the external environment changes or during the real-time optimization phase in a dynamic environment, the fitness of the scheduling scheme after changes in the external environment is comprehensively considered. As a genetic algorithm, the individual fitness of the genetic algorithm is used to re-evaluate and update the population, and obtain a scheduling scheme that is adapted to the current environment.

5. The scheduling method according to claim 4, characterized in that, The comprehensive objective function is as follows: in, These are the weighting coefficients for delivery time, transportation cost, and route length, respectively. The total delivery time of the scheduling plan Delivery time metrics after normalization The total transportation cost of the scheduling plan The transportation cost indicator after normalization. It refers to the path length of the scheduling scheme. The path length index after normalization.

6. The scheduling method according to claim 5, characterized in that, in, It is the total delivery time of the scheduling plan. This refers to the delivery time of the unmanned transport vehicle. It refers to the delivery time of the drone. It is the total transportation cost of the scheduling plan. It's the transportation cost of the unmanned transport vehicle. The cost is the transportation cost of the drone. It is the path length of the scheduling scheme. It is the path length of the unmanned transport vehicle. It is the path length of the drone.

7. The scheduling method according to claim 6, characterized in that, in, The unmanned transport vehicle completed the task. Delivery time The unmanned transport vehicle completed the task. Delivery time Indicates the number of unmanned transport vehicle missions; The drone completed the mission. Delivery time The drone completed the mission. Delivery time Indicates the number of drone missions; in, The unmanned transport vehicle completed the task. Fuel consumption and transportation costs It is a task and tasks The distance between them The unmanned transport vehicle completed the task. Other transportation costs during the execution process, The drone completed the mission. Fuel consumption and transportation costs It is a task and tasks The distance between them The drone completed the mission. Other transportation costs incurred during the execution process; in, It is the [number]th [unit] in the unmanned transport vehicle scheduling scheme, arranged in the order of execution. Each delivery node This task involves calculating the road network map of the disaster area based on the real-time road and traffic information obtained in step S1. and tasks The shortest passable path length or the minimum travel cost It is the [number]th drone in the drone scheduling scheme arranged in execution order. Each delivery node This refers to the task calculated on the road network map of the disaster area based on the real-time weather information and airspace constraint information obtained in step S1. and tasks The shortest possible flight path length or the minimum travel cost.

8. The scheduling method according to claim 4, characterized in that, The adaptability of the scheduling scheme after comprehensive consideration of changes in the external environment The calculation formula is: in, The total delivery time of the scheduling plan Delivery time metrics after normalization The total transportation cost of the scheduling plan The transportation cost indicator after normalization. It refers to the path length of the scheduling scheme. The path length index after normalization These are the weighting coefficients for delivery time, transportation cost, and route length, respectively. It is an environmental factor. It is the adjustment coefficient for the impact of traffic on the objective function. It is an adjustment factor for the impact of weather on the objective function. It is a factor influencing traffic conditions. It is an influencing factor of weather conditions. It is the overall impact adjustment coefficient. It is an influencing factor of other external factors.

9. An intelligent dispatching system for the coordinated ground-air delivery of emergency supplies, characterized in that, include: The data acquisition module acquires real-time status information of drones and unmanned transport vehicles, as well as external environmental information such as road conditions, weather, traffic conditions, and the disaster area. The genetic algorithm module uses information acquired by the data acquisition module to establish a scheduling scheme. It then uses a genetic algorithm to optimize the scheduling scheme through multiple generations of selection, crossover, and mutation operations. The real-time optimization module comprehensively considers multiple performance indicators such as delivery time, transportation cost, and route length to calculate the optimal task allocation and route length planning. The execution control module, through real-time feedback and monitoring system, directs unmanned transport vehicles to transport supplies from the material storage center to assembly points or temporary resettlement points on the periphery of the disaster area according to the optimized scheduling plan, using restored roads or cleared areas. In the delivery module, drones, based on optimized path length planning and scheduling schemes, accurately deliver supplies from assembly points or temporary settlement points to target areas.

10. The system according to claim 9, characterized in that, The genetic algorithm module includes a fitness function design module, which continuously adjusts the scheduling scheme through iterative optimization to achieve the optimization goal of minimizing delivery time, transportation cost and path length.

Images