A method and system for optimizing the mission execution process of a long-flight unmanned aerial vehicle

Abstract

This application provides a method and system for optimizing the mission execution process of a long-flight UAV, including: obtaining mission information of the long-flight UAV to be optimized, the mission information including monitoring mission information, mission control mission information, and flight-related information; setting a Multi-U&G IOP model for the integrated optimization problem of multi-flight UAV mission planning and multi-GCS control resource allocation; and optimizing the mission of the long-flight UAV to be optimized based on the mission information and the Multi-U&G IOP using a preset three-stage iterative optimization algorithm TSIOA. The three-stage optimization algorithm allows for rapid and accurate optimization solutions to be obtained while considering multiple long-flight UAVs and multiple variables.

Description

Technical Field

[0001] The present application relates to the field of drones, and in particular to an optimization method and system for a long-flight drone to perform a mission. Background Art

[0002] In recent years, drones, as advanced aerial vehicles, have demonstrated significant potential in a wide range of fields, including military strike, reconnaissance and surveillance, communications relay, disaster management, and remote sensing (RS) mapping. Based on their endurance, drones can generally be categorized as short-endurance and long-endurance. Short-endurance drones, limited by their payload and energy configuration, have slower flight speeds and shorter ranges. Long-endurance drones, however, offer longer flight times and wider coverage, enabling them to carry out more sustained and extensive missions. Consequently, various countries have developed a variety of long-endurance drones, including China's Rainbow-5 and the United States' RQ-4 Global Hawk and MQ-9 Reaper.

[0003] Currently, long-endurance drones (UAVs) possess certain intelligence and autonomy features. They can automatically cruise along a manually set route, avoiding obstacles, and so on. However, during extended flights, they still require human oversight, and their onboard computers are not yet capable of automatically executing military missions such as attack, reconnaissance, and surveillance. Operators are still required to send commands through ground stations to control the drone's execution, a practice known as human-in-the-loop command and control. For example, the Northrop Grumman RQ-4 Global Hawk UAV, after takeoff, is controlled by a pilot and an image sensor operator within the GCS, operating the drone and conducting reconnaissance operations using onboard sensors.

[0004] The shortage of long-flight UAV operators has limited GCS control resources, which in turn limits the number of UAVs that can perform missions and hinders the advantages of multi-UAV operations. Therefore, when planning UAV missions, it is necessary to consider the constraints and allocation of control resources.

[0005] The problem of multi-UAV mission planning is well-studied, with scholars exploring numerous mission planning algorithms, including mathematical programming and heuristic algorithms. However, to our knowledge, most current research focuses on short-endurance UAVs and is based on the assumption that they possess fully autonomous intelligence and require no control resources, or that UAVs have sufficient control resources. However, in actual military applications, long-endurance UAVs lack full autonomy and require GCS intervention and control at certain stages. Regarding the relationship between GCSs and UAVs, current research has only considered a fixed, one-to-one pairing between ground stations and UAVs, failing to consider the flexible matching mechanisms between UAVs and ground stations. Summary of the Invention

[0006] The main purpose of the embodiments of the present invention is to provide an optimization method and optimization system for the long-flight UAV mission execution process. By setting a three-stage optimization algorithm, an optimization solution can be obtained quickly and accurately while considering multiple long-flight UAVs and multiple variables.

[0007] In the first aspect, a method for optimizing the mission execution process of a long-flight UAV is provided, the optimization method comprising:

[0008] Obtain the task information of the long-flight UAV to be optimized, the task information includes: monitoring task information, control task information and flight related information; the monitoring task information is: every interval Time, spent through a ground control station GCS The flight status and working status of the onboard sensor of the long-flight UAV to be optimized are checked at the time; the control task information is: when the long-flight UAV to be optimized is at the target point During the mission, the long-flight UAV to be optimized is controlled by a GCS to perform the target mission; the flight-related information includes: the departure time of the long-flight UAV to be optimized , the longest flight time , latest return time , the flight time required from GCS node a to GCS node b , at the mission target point The number of resources required is , the maximum number of resources carried by long-flight drones ;

[0009] Set up a Multi-U&GIOP model for the integrated optimization problem of multi-flight UAV mission planning and multi-GCS control resource allocation;

[0010] The mission of the long-flight UAV to be optimized is optimized according to the mission information and the Multi-U&G IOP through the preset three-stage iterative optimization algorithm TSIOA.

[0011] In one possible implementation, the performing of mission optimization on the long-flight UAV to be optimized according to the mission information and the Multi-U&G IOP by using a preset TSIOA includes:

[0012] Obtaining a first mission plan of the long-flight UAV to be optimized by a preset hybrid genetic algorithm HGA, where the first mission plan is a mission control mission plan;

[0013] Obtaining a second mission plan for the long-flight UAV to be optimized through a preset lightweight mixed-integer linear programming model and a heuristic adjustment strategy, where the second mission plan is a monitoring mission plan;

[0014] The control task and monitoring task are issued according to the first and second task plans using a preset heuristic method.

[0015] In one possible implementation, the first mission plan of the long-flight UAV to be optimized is obtained by using a preset hybrid genetic algorithm HGA, where the first mission plan is a mission control mission plan, including:

[0016] Obtaining an initial solution of the Multi-U&G IOP model;

[0017] Changing the domain structure set of the initial solution by LS-VND to obtain an optimized initial solution;

[0018] The optimized initial solution is tested using a preset fitness function.

[0019] In one possible implementation, the second mission plan of the long-flight UAV to be optimized is obtained by using a preset lightweight mixed integer linear programming model MILP and a heuristic adjustment strategy, where the second mission plan is a monitoring mission plan:

[0020] Solve the lightweight MILP model and make decisions on monitoring tasks;

[0021] According to the statistical quartiles of the start time distribution of the mission sequence of the long-flight UAV to be optimized, a heuristic strategy is designed to solve the situation where the MILP model has no feasible solution, and the feasible solution is the second mission plan.

[0022] In one possible implementation, the issuing of the control task and the monitoring task according to the first and second task plans by using a preset heuristic method includes:

[0023] Arrange the targets and monitoring tasks in ascending order according to the start time to obtain the task set H;

[0024] Obtain the minimum task h of the task set H and distribute the tasks according to the preset distribution principle, which includes: obtaining the total working time of GCS , for all GCS according to the Sort in ascending order; get The smallest GCS is obtained, and the working status of the smallest GCS is obtained. If the working status is idle, h is directly assigned to the smallest GCS. If the working status is busy, the completion time of the task being executed by the smallest GCS is obtained. If the completion time is later than the start time of h, h is assigned to The second smallest GCS.

[0025] Secondly, a system for optimizing the mission execution process of a long-flight UAV is provided. The optimization system includes:

[0026] The mission information acquisition module is used to obtain the mission information of the long-flight UAV to be optimized. The mission information includes: monitoring mission information, mission control mission information and flight related information; the monitoring mission information is: Time, spent through a ground control station GCS The flight status and working status of the onboard sensors of the long-flight UAV to be optimized are checked at the time; the mission control mission information is: when the long-flight UAV to be optimized is at the mission target point During the mission, the long-flight UAV to be optimized is controlled by a GCS to perform the mission control task; the flight related information includes: the departure time of the long-flight UAV to be optimized , the longest flight time , latest return time , the flight time required from GCS node a to GCS node b , at the mission target point The number of resources required is , the maximum number of resources carried by long-flight drones ;

[0027] Multi-U&G IOP model setting module, used to set the Multi-U&G IOP model for the integrated optimization problem of multi-flight-duration UAV mission planning and multi-GCS control resource allocation;

[0028] The mission optimization module is used to optimize the mission of the long-flight UAV to be optimized according to the mission information and the Multi-U&G IOP through a preset three-stage iterative optimization algorithm TSIOA.

[0029] In one possible implementation, the task optimization module includes:

[0030] A first mission plan acquisition unit is configured to acquire a first mission plan of the long-flight UAV to be optimized by using a preset hybrid genetic algorithm (HGA), wherein the first mission plan is a mission control mission plan;

[0031] a second mission plan acquiring unit, configured to acquire a second mission plan for the long-flight UAV to be optimized by using a preset lightweight mixed integer linear programming model and a heuristic adjustment strategy, wherein the second mission plan is a monitoring mission plan;

[0032] A task issuing unit is used to issue the control task and the monitoring task according to the first and second task plans using a preset heuristic method.

[0033] In one possible implementation, the first task plan acquisition unit includes:

[0034] An initial solution obtaining unit, configured to obtain an initial solution of the Multi-U&G IOP model;

[0035] An optimized initial solution obtaining unit, configured to change the domain structure set of the initial solution through LS-VND to obtain an optimized initial solution;

[0036] A detection unit is used to detect the optimized initial solution through a preset fitness function.

[0037] In a possible implementation, the second task plan acquisition unit includes:

[0038] A monitoring task decision unit, used to solve the lightweight MILP model and decide on monitoring tasks;

[0039] The second task planning unit is used to design a heuristic strategy to solve the situation where the MILP model has no feasible solution based on the statistical quartiles of the start time distribution of the task sequence of the long-flight UAV to be optimized, where the feasible solution is the second task plan.

[0040] In one possible implementation, the task issuing unit includes:

[0041] The task set acquisition unit is used to arrange the targets and monitoring tasks in ascending order according to the start time to obtain the task set H;

[0042] The task allocation unit is used to obtain the minimum task h of the task set H and allocate tasks according to a preset allocation principle. The allocation principle includes: obtaining the total working time of the GCS , for all GCS according to the Sort in ascending order; get The smallest GCS is obtained, and the working status of the smallest GCS is obtained. If the working status is idle, h is directly assigned to the smallest GCS. If the working status is busy, the completion time of the task being executed by the smallest GCS is obtained. If the completion time is later than the start time of h, h is assigned to The second smallest GCS. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for describing the embodiments of the present application.

[0044] Figure 1 A flowchart of a method for optimizing a long-flight UAV mission execution process provided by one embodiment of the present invention;

[0045] Figure 2 A structural diagram of a system for optimizing the mission execution process of a long-flight UAV provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0046] The following describes embodiments of the present application in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar modules or modules having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application and are not to be construed as limiting the present invention.

[0047] It will be understood by those skilled in the art that, unless expressly stated otherwise, the singular forms "a", "an", "said" and "the" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the specification of this application refers to the presence of the features, integers, steps, operations, modules and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, modules, components and / or groups thereof. It should be understood that when we refer to a module as being "connected" or "coupled" to another module, it may be directly connected or coupled to the other module, or there may be an intermediate module. In addition, "connected" or "coupled" as used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any modules and all combinations of one or more associated listed items.

[0048] In order to make the objectives, technical solutions and advantages of this application clearer, the implementation of this application will be further described in detail below with reference to the accompanying drawings.

[0049] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0050] For ease of understanding, the English abbreviations involved in the present application are explained in advance:

[0051] LE-UAV: Long Endurance Unmanned Aerial Vehicle, long-endurance UAV;

[0052] GCS: Ground control station, ground control station:

[0053] Multi-U&G IOP: Integrated optimization problem of multi-LE-UAV missionplanning and multi-GCS control resource allocation, integrated optimization problem of multi-LE-UAV mission planning and multi-GCS control resource allocation;

[0054] HGA: hybrid genetic algorithm, hybrid genetic algorithm;

[0055] LS-VND: local search variable neighborhood decline, local search and variable neighborhood decline algorithm;

[0056] TSIOA: Three-Stage Iterative Optimization Algorithm, three-stage iterative optimization algorithm;

[0057] MILP model: Mixed Integer Linear Programming, mixed integer linear programming model.

[0058] like Figure 1 FIG2 is a flow chart of a method for optimizing a long-flight UAV mission execution process according to an embodiment of the present invention. The method includes:

[0059] Step S101, obtaining the task information of the long-flight UAV to be optimized, the task information includes: monitoring task information, control task information and flight related information, the monitoring task information is: every interval Time, spent by a bottom control station GCS The flight status and working status of the onboard sensor of the long-flight UAV to be optimized are checked at the time. The control task information is: when the long-flight UAV to be optimized is at the target point During the mission, the long-flight UAV to be optimized is controlled by a GCS to perform the target mission. The flight-related information includes: the departure time of the long-flight UAV to be optimized; , the longest flight time , latest return time , the flight time required from GCS node a to GCS node b , at the mission target point The number of resources required is , the maximum number of resources carried by long-flight drones ;

[0060] Step S102, setting a Multi-U&G IOP model for the integrated optimization problem of multi-flight-time UAV mission planning and multi-GCS control resource allocation;

[0061] Step S103: performing mission optimization on the long-flight UAV to be optimized according to the mission information and the Multi-U&G IOP using a preset three-stage iterative optimization algorithm TSIOA.

[0062] In an embodiment of the present invention, for the mission optimization of long-flight UAVs, this application presupposes the following conditions: 1. There is only one airport, and all UAVs depart from the airport and return to the original airport after completing the mission; 2. The total duration of the UAV mission is fixed, and the service time and resource requirements of the target point are fixed; 3. The ground station, UAV, and satellite are visible in real time, and the signals are not interfered with; 4. Within a given time, each control mission cannot be interrupted once it starts; 5. Any GCS can only control one UAV at a time, and a UAV can only be controlled by one GCS at a time; 6. The working capabilities of each GCS are the same.

[0063] Assume a fully connected graph , where GCS node , arc set Node 0 represents the airport where the drone takes off and lands. It is the place where the drone departs and returns. Represents the GCS nodes that the long-endurance UAV passes through during its mission.

[0064] Set time , we consider a group of identical drones exist Depart from the airport at any time and head to the target point to perform the mission. The drone's flight time shall not exceed ,exist Once the drone starts to go to the target point to perform the mission, the mission process cannot be interrupted. Service time is the time it takes to execute a task. , the flight time required from GCS node a to GCS node b is Target point The number of resources required is , the maximum number of resources carried by the drone is .if Describe the target point For reconnaissance missions, only the drone's fuel is consumed.

[0065] set up Represents the GCS set of the drone. During the autonomous flight of the drone traversing the GCS nodes, every time , you need a GCS cost Timely check of the drone's flight status and sensor working status, referred to as monitoring tasks, each drone Monitoring task set The calculation formula is In addition, the drone is at the target point During the mission, a GCS is required to control the UAV to perform the mission, referred to as the mission control mission.

[0066] The optimization goal is to complete all tasks in the shortest possible time. To minimize the time it takes for the drone to traverse the target points, we need to determine two sets of decisions: assigning target points to drones and sorting these target points; and assigning monitoring and control tasks to the GCS and sorting these tasks.

[0067] The MILP model of Multi-U&G IOP is:

[0068] (1)

[0069] Among them, the Multi-U&G IOP model (1) minimizes the time it takes for each UAV to traverse the target and minimizes the time it takes for all control tasks to be completed.

[0070] Accordingly, the following constraints are set for the Multi-U&G IOP model:

[0071] (2)

[0072] (3)

[0073] (4)

[0074] (5)

[0075] (6)

[0076] (7)

[0077] (8)

[0078] (9)

[0079] (10)

[0080] (11)

[0081] (12)

[0082] (13)

[0083] (14)

[0084] (15)

[0085] (16)

[0086] (17)

[0087] (18)

[0088] (19)

[0089] (20)

[0090] (twenty one) (twenty two)

[0091] Among them, constraints (2)-(3) ensure that each aircraft departs from the airport and eventually returns to the airport.

[0092] Constraint (4) is the conservation of flow for each target.

[0093] Constraint (5) ensures that each target is traversed by only one drone.

[0094] Constraints (6)-(7) couple the routing decision with the time schedule, where M is a large positive constant. (6) means that if =1, then the drone The first target traversed from airport 0 The arrival time is equal to the departure time Add from airport 0 to target The flight time of . Constraint (7) means that if , namely drones From the target Visit the target directly after departure , then the drone arrive The time to arrive is equal to Time plus goal Service time and goals To the target The flight time between.

[0095] Constraints (8) ensure that drones Virtual nodes that have not been visited of is 0.

[0096] Constraints (9)-(10) ensure and coupling.

[0097] Constraint (11) is the flight duration limit, which ensures that each drone can return home.

[0098] Constraint (12) ensures that each UAV has enough resources to meet the goals it traverses.

[0099] Constraint (13) ensures that the interval between monitoring tasks of the same UAV is at least .

[0100] Constraint (14) ensures that the maximum interval between monitoring tasks of the same UAV is .

[0101] Constraint (15) ensures that there is only one GCS controlling each time during the monitoring phase.

[0102] Constraint (16) indicates the coupling between the decision-making and time scheduling of the mission control task. GCS is Execute mission control tasks The start time is equal to the time when the drone reaches the target time.

[0103] Constraint (17) indicates the guarantee target The mission control is performed by one and only one GCS.

[0104] Constraint (18) couples the monitoring decision and the control task decision, ensuring that the GCS can only control one UAV at a time.

[0105] Constraints (19)–(22) are used to define the domain of the decision variables.

[0106] It should be pointed out that the Multi-U&G IOP model (1) is presented in a nonlinear form of minimizing the maximum function. We define an auxiliary variable , through the auxiliary variable, the constraint (1) is transformed into the following linear expression:

[0107] (twenty three)

[0108] (twenty four)

[0109] Due to the complexity of the Multi-U&G IOP model and the difficulty in solving it when dealing with large-scale instances, we seek to obtain a lower bound for the problem to provide a benchmark reference for the heuristic algorithm. If we do not consider the impact of GCS on multiple drones, assume that the number of GCSs at any time meets the requirements of multiple drones, and only consider the MILP model of drone task allocation, then the result can be regarded as a lower bound for the original problem model. This is because when considering control factors, the total time required to complete the task will inevitably exceed the total time when ignoring control factors. Therefore, we reduce the scale of the original problem by relaxing the control resource constraint (constraints (13)-(18)).

[0110] Since constraint (16) is relaxed, the variable To couple task allocation with GCS allocation. The variable dimension is high, and the take-off time of each drone can be 0, that is, Therefore, we relax constraints (6)-(11) and express the flight time limit through the new constraint (30). However, relaxing constraints (6)-(7) will make the arrival time relationship between nodes (the order in which drones arrive) disappear, causing a sub-loop problem in the solution. Therefore, we add auxiliary variables based on the above MILP model parameters. , in order to eliminate the sub-loop.

[0111] (25)

[0112] (26)

[0113] (27)

[0114] (28)

[0115] (29)

[0116] (30)

[0117] (31)

[0118] (32)

[0119] Among them, constraints (25)-(29) are consistent with the constraints of the aforementioned MILP model. Constraint (30) is the flight time limit; constraint (31) is the sub-loop elimination; constraint (32) is the capacity limit constraint. Constraints (31) and (32) can make the model tighter.

[0120] The step of optimizing the mission of the long-flight UAV to be optimized according to the mission information and the Multi-U&G IOP by using the preset TSIOA includes:

[0121] Obtaining a first mission plan of the long-flight UAV to be optimized by a preset hybrid genetic algorithm HGA, where the first mission plan is a mission control mission plan;

[0122] Obtaining a second mission plan for the long-flight UAV to be optimized through a preset lightweight mixed-integer linear programming model and a heuristic adjustment strategy, where the second mission plan is a monitoring mission plan;

[0123] The control task and monitoring task are issued according to the first and second task plans using a preset heuristic method.

[0124] The first mission plan of the long-flight UAV to be optimized is obtained by using a preset hybrid genetic algorithm HGA, and the first mission plan is a mission control mission plan, including:

[0125] Obtaining an initial solution of the Multi-U&G IOP model;

[0126] Changing the domain structure set of the initial solution by LS-VND to obtain an optimized initial solution;

[0127] The optimized initial solution is tested using a preset fitness function.

[0128] In an embodiment of the present invention, the construction of the initial solution of the Multi-U&G IOP model includes: ,in, Indicates that the target value in the initial solution is the maximum value of all current target completion times; It is the GCS control resource allocation conflict label, obtained by the control conflict resolution program; It is the time list of GCS to start the mission control of each UAV. A chromosome is a collection of multiple sub-chromosomes. ,Each sub-chromosome represents a UAV mission planning scheme. The target visit order list in the sub-chromosome is encoded using integer encoding, where positive integers represent target points and '0' represents the UAV airport.

[0129] In the sub-chromosome corresponding to each UAV mission planning scheme, It is a valid label of the sub-chromosome, which is used to verify whether the drone of the sub-chromosome takes off from the airport, traverses some target points, and finally returns to the airport. It is the requirement of all targets in each drone’s planned route, is the flight time of each drone from takeoff to landing, Represents the time series of each drone’s take-off time and arrival time at each node.

[0130] The quality of the initial solution is crucial to the HGA solution results and convergence speed. To improve the quality and richness of the initial population solution, we use the following five strategies to generate the initial solution.

[0131] (1) Nearest Neighbor

[0132] The nearest neighbor method is a common approach for virtual reality (VRP) problems. We first select point A, the closest point to the origin. Then, using a greedy strategy, we select the nearest point with a 90% probability. We then randomly select a point with a 10% probability, and so on. This continues until the range and resources are insufficient for the drone to reach the next node, at which point it returns to its destination.

[0133] (2)Random average generation

[0134] Unlike traditional drone mission planning problems, the goal in military and disaster relief settings is often to complete all missions in the shortest possible time. Therefore, we evenly assign the goal to dispatchable drones, aiming to complete all missions as quickly as possible. This is accomplished through the following steps:

[0135] Step 1: Disrupt the task sequence;

[0136] Step 2: Assign the targets to the drones in set U one by one in order , until the last u is also assigned a target, and the selected target is deleted from the sequence;

[0137] Step 3: Move the unassigned target from U to the first Start allocation and loop step 2 until allocation is completed;

[0138] Step 4: Output the mission sequence of each drone.

[0139] (3) Mean nearest neighbor

[0140] Each drone selects the point closest to the current node in turn, thereby averaging the task load of each drone to speed up the overall task completion time.

[0141] Step 1: According to the order in set U, the drones The target is selected in the task sequence using the Nearest Neighbor method, and the selected target is deleted from the sequence;

[0142] Step 2: Loop step 1 until the task sequence is empty;

[0143] Step 3: Output the mission sequence of each drone.

[0144] (4) Improved K-means

[0145] The K-means algorithm is the most commonly used algorithm for clustering task targets. It achieves the most satisfactory clustering results by iteratively modifying cluster centers and clusters. It has advantages such as simple computation and the ability to quickly process large data sets. The traditional K-means algorithm randomly selects initial cluster centers from the data objects, but the K-means algorithm is sensitive to these initial centers, and clustering results corresponding to different initial centers can vary significantly.

[0146] In order to ensure the clustering effect and avoid the situation where there is no task in some task clusters, the initial cluster centers are dispersed as much as possible. , define the task cluster center set , where the elements are cluster centers , the cluster centers to be generated The number is equal to the number of drones Mission Objectives and The Euclidean distance between , cluster center and mission objectives The Euclidean distance between ; for The number of cluster centers in the dataset is initially 0; for Middle Tasks to The sum of the distances to all cluster centers in , then:

[0147]

[0148]

[0149]

[0150] Randomly select a point from the input control task set As the first cluster center , the task from Pull out and put in In, then Extraction The biggest task As the newly selected cluster center Put in Repeat the extraction process until The number of cluster centers in equal , thus completing the calculation of the initial cluster center.

[0151] (5)Average Sweep Algorithm

[0152] The Sweep Algorithm was first proposed by Gillett and Miller in 1974 as a heuristic algorithm for solving Vehicle-Related Plans (VRPs) with an unlimited number of vehicles. Based on the characteristics of the problem, we have made the following improvements:

[0153] Step 1: Establish a polar coordinate system

[0154] Take the airport coordinates as the pole, and define the line connecting any target in the diagram and the pole as angle 0, establish a polar coordinate system, and then convert all target points into points in the polar coordinate system;

[0155] Step 2: Grouping

[0156] According to the number of drones U, the entire polar coordinate system is divided into U groups in a counterclockwise direction. For each group, the targets are added to the drone mission sequence one by one according to the Nearest Neighbor method. After adding the last target, return to the airport.

[0157] Step 3: Check whether the mission sequence of each drone exceeds the range or capacity limit. If so, try to repair it;

[0158] Step 4: Output all task sequences.

[0159] We evenly use the above five methods to generate the initial population. Next, we perform LS-VND on each chromosome and find a sequence with shorter time by adjusting the order of the targets in each drone mission sequence.

[0160] In this embodiment of the present invention, the purpose of LS-VND is to systematically change the neighborhood structure set of the current solution during the search process to expand the search range and find better solutions. For each UAV's mission sequence, we execute the LS-VND algorithm to search the neighborhood.

[0161] In LS-VND, we have made some innovative improvements to Using iterative greedy search, Find the current optimal After that, update , continue searching until the optimal .

[0162] For each new task sequence generated by LS-VND, the time cost is used To evaluate the quality of these sequences, we use a local search algorithm to update the task sequence based on the objective cost. From an algorithmic perspective, the adaptive LS-VND algorithm aims to enhance the intensification and diversification of the HGA. The core idea of ​​the LS-VND algorithm is to systematically change the neighborhood structure of the current solution, expand the search range, and then find the current optimal solution through a local search algorithm.

[0163] In HGA, fitness mainly consists of two parts. One part is the time it takes for the drone to complete all control tasks. , that is, the benefit, is equal to the solution ; Part of it is that the drone violates the control constraints ( =0), penalty value caused by range limitation and capacity limitation .

[0164] The fitness function can be expressed as:

[0165]

[0166] The calculation of is as follows:

[0167]

[0168] in, is the number of drones in this chromosome, Represents the drone in this chromosome The resource requirement of the task sequence, Q is the rated capacity of the UAV, For drones Flight duration, Rated flight time for the drone. is the current iteration number, is the total number of iterations, is the penalty value due to failure to satisfy the control constraints.

[0169] coefficient The calculation formula is as follows:

[0170]

[0171] in, is the task completion time of the optimal solution in the current population, is the sum of the service time of all control tasks, is a small positive number to ensure that the denominator is not zero.

[0172] It increases with the number of iterations, which means that the penalty term gradually becomes heavier as the algorithm progresses, pushing the algorithm to find a solution that satisfies all constraints as quickly as possible.

[0173] In this embodiment of the present invention, the fitness function is a performance evaluation metric for a solution or chromosome, influencing the genetic algorithm's search process and the quality of the final solution. The purpose of the fitness function is to provide a quantitative score indicating the quality of the chromosome, thereby guiding the algorithm's search towards a more optimal solution.

[0174] In HGA, fitness mainly consists of two parts. One part is the time it takes for the drone to complete all control tasks. , that is, the benefit, is equal to the solution ; Part of it is that the drone violates the control constraints ( =0), penalty value caused by range limitation and capacity limitation .

[0175] The fitness function can be expressed as:

[0176]

[0177] The calculation of is as follows:

[0178]

[0179] in, is the number of drones in this chromosome, Represents the drone in this chromosome The resource requirement of the task sequence, Q is the rated capacity of the UAV, For drones Flight duration, Rated flight time for the drone. is the current iteration number, is the total number of iterations, is the penalty value due to failure to satisfy the control constraints.

[0180] coefficient The calculation formula is as follows:

[0181]

[0182] in, is the task completion time of the optimal solution in the current population, is the sum of the service time of all control tasks, is a small positive number to ensure that the denominator is not zero.

[0183] It increases with the number of iterations, which means that the penalty term gradually becomes heavier as the algorithm progresses, pushing the algorithm to find a solution that satisfies all constraints as quickly as possible.

[0184] In the embodiment of the present invention, HGA can be evolved through the following strategies: elite retention strategy, crossover strategy, mutation operator and simulated annealing mechanism.

[0185] Elite Retention Strategy: In the HGA framework, offspring generation strategies are diverse and can be categorized into three main types: elite retention, elite reproduction, and general reproduction. These strategies work together to balance the algorithm's exploration and exploitation capabilities.

[0186] Elite retention: This refers to retaining a subset of chromosomes with the highest fitness in each generation, ensuring that the algorithm doesn't lose the best solutions it has already found. This strategy is a common technique in genetic algorithms and helps them converge quickly.

[0187] Elite reproduction: is a special strategy that aims to accelerate convergence by preserving and spreading excellent genes. Specifically, a portion of chromosomes with the highest fitness in the population, namely the elite chromosomes, are selected. Then, another chromosome is selected from the remaining population by the binary tournament method. A maximum overlap crossover is performed with it. The chromosome after the crossover undergoes recombination optimization, mutation, repair, and LS-VND acceleration to generate a high-quality daughter chromosome. This process helps combine the current optimal solution with other potentially beneficial features, thereby exploring even better solutions.

[0188] General reproduction: The remaining chromosomes in the population undergo general reproduction. Similarly, the binary tournament method is used to select two parent chromosomes. These are then followed by crossover, recombination optimization, mutation, repair, and LS-VND acceleration to generate two daughter chromosomes. Only the general reproduction process is subject to the crossover rate. This process aims to introduce new gene combinations into the population, which helps the algorithm avoid premature convergence to a local optimum.

[0189] Crossover strategies, a series of crossover strategies are designed based on the characteristics of the problem to improve the quality of the solution and the efficiency of the algorithm. These strategies include:

[0190] Biggest overlap crossover (BOC): This crossover operation targets the routes with the largest geographical overlap, with the goal of merging routes with favorable spatial structures. The process is shown in the figure. We start from the parent chromosome. Randomly select a sub-route and randomly select a fragment in the sub-route Then eliminate the elite chromosomes Elements in the fragment. Calculated later Find the minimum bounding rectangle of each sub-route and the fragment The sub-route with the largest overlapping area of ​​the minimum bounding rectangle Finally, insert the fragment into Generate a chromosome at the optimal position (the position with the shortest path distance after insertion among all possible insertion points) .

[0191] Subroute Exchange crossover (SEC): From the parent chromosome Randomly select a sub-route and randomly select a fragment in the sub-route . Then eliminate the parent chromosome In the fragment Elements. Put the fragment Insert At the optimal position of The same method is used to generate daughter chromosomes .

[0192] Subroute single point crossover (SSPC): From the parent chromosome Randomly select a sub-route and randomly select a fragment in the sub-route , then eliminate the parent chromosome In the fragment . The elements in are inserted one by one into At the optimal position of The same method is used to generate daughter chromosomes .

[0193] Simple Random crossover (SRC): From the parent chromosome Randomly select a sub-route and randomly select a fragment in the sub-route , then eliminate the parent chromosome In the fragment . The elements in are inserted one by one into The specific method is: random selection Insert the point into the optimal position of a sub-route in the sub-route. Finally, the offspring chromosome is generated. The same method is used to generate daughter chromosomes .

[0194] After crossover, the newly formed offspring chromosomes are subjected to further crossover operations, called recombination optimization operators (ROOs). These operations are repeated to find better offspring.

[0195] Mutation Operator,In order to improve the exploration ability of genetic algorithm, we designed two different,mutation operators:

[0196] Simple optimal mutation: Randomly select a sub-route from the chromosome. In the process, randomly select a node and then randomly select another sub-route , insert this point into the sub-route The optimal position of .

[0197] Simple random mutation: randomly select a subroute from the chromosome, randomly select a point in the subroute, and then randomly select another subroute , insert this point into the sub-route At a random position in .

[0198] After our tests, using simple optimal mutation in elite reproduction and simple random mutation in general reproduction are the best.

[0199] For the simulated annealing mechanism, we draw on the idea of ​​the "temperature" parameter in the simulated annealing algorithm. The higher the temperature, the greater the probability of accepting a worse solution. As the temperature gradually decreases, the probability of accepting a worse solution decreases, and the algorithm gradually stabilizes and focuses more on searching for the optimal solution or a near-optimal solution.

[0200] In the early stage of the HGA algorithm, a certain probability is allowed Accepting children that are worse than their parents will help the algorithm remain exploratory.

[0201]

[0202] In the later stages, we tend to accept offspring with higher fitness values ​​than their parents.

[0203] In this strategy, the algorithm allows some individuals with lower fitness to be retained in the population to maintain the diversity of the population and help avoid premature convergence.

[0204] By judiciously balancing convergence and diversity, our designed hybrid genetic algorithm is able to effectively solve the mission goal planning problem.

[0205] The repair operator is a key component in HGAs, crucial for maintaining the diversity and quality of the population. It is responsible for adjusting infeasible solutions (violating chromosomes) to satisfy the actual constraints of the problem. In HGAs, we design two repair operators: a control conflict resolution procedure and a route repair operator.

[0206] Control conflict resolution procedures

[0207] In this section, we propose and elaborate a strategy for resolving control conflicts in UAV scheduling tasks. Control conflicts typically occur when multiple UAVs are collaborating on a mission, but the limited number of ground-based GCSs (Ground Control Systems) makes it impossible to simultaneously manage all the UAVs. In such situations, UAVs may encounter a shortage of GCS resources upon reaching their mission point, necessitating measures to prevent mission interruption.

[0208] Our strategy is to design a rational UAV takeoff and mission execution scheduling plan based on maintaining mission efficiency and rationally utilizing control resources. For UAVs at the initial mission node, if GCS control resources are insufficient, takeoff can be delayed to reduce the possibility of waiting in the air. For UAVs mid-flight, if control resources are insufficient, they can choose to hover and wait until the GCS is free.

[0209] Based on these principles, we designed a control conflict resolution strategy based on the time-priority principle. The core purpose of this strategy is to ensure that the first drone to arrive at the task node receives control resources. This strategy is implemented using the Control Schedule Matrix (CSM). This matrix is ​​generated based on the time series of nodes in the sub-chromosomes and records the number of drones to be controlled at each time point. This ensures that at any time t, the sum of each column in the CSM does not exceed the upper limit S on the number of GCSs.

[0210] The implementation process of the strategy is as follows:

[0211] 1) Initialize the CSM and construct a two-dimensional matrix, where rows represent drones and columns represent time points. The values ​​of the matrix indicate whether the drone needs to be controlled at the corresponding moment;

[0212] 2) Starting from time t=0, traverse the CSM moment by moment and detect the sum of control demands at each moment;

[0213] 3) If the control demand exceeds the threshold S at a certain moment, all the drones that require control at that moment are identified and conflict analysis is performed;

[0214] 4) Sort the conflicting drones based on the start time of the conflicting mission, the number of postponements, and the completion time of the drone's last mission. The purpose of sorting is to prioritize the drones that complete their missions.

[0215] 5) The first S drones will be retained to continue the mission. The remaining drones will adopt a postponement strategy, calculate the postponement time, and update the drone's mission start time and CSM;

[0216] 6) When delaying drones, use incremental delay, that is, only delay one drone at a time and recalculate the CSM to verify whether the control requirements are met. If it is still not met, continue to delay other drones until the requirements are met or the preset maximum number of attempts is reached;

[0217] 7) If a suitable solution cannot be found within a limited number of attempts, the label of the drone violating the control constraint is recorded and the next stage of penalty value calculation is entered.

[0218] Through the above steps, we ensure that under limited control resources, the UAV mission can be executed according to the optimized scheduling plan, maximize the mission efficiency, and reasonably allocate control resources. , UAV 2's first mission Conflict with other drones, so directly postpone the takeoff time of UAV 2. arrive ', all subsequent tasks are postponed Units of time; UAV 3's second mission Conflict, so by hovering and waiting, the control time is postponed by to .

[0219] Route Repair Operator

[0220] In the Multi-U&G IOP problem, the primary purpose of the route repair operator is to ensure that each UAV's route satisfies constraints in terms of demand and flight time. If a constraint is violated, the repair operator restores the chromosome's feasibility by selecting routes with the highest demand and flight time and appropriately redistributing nodes between them. This redistribution is accomplished through a series of insertion and deletion operations, as shown in the figure. Reallocation of the target node involves removing the last target point from the extreme route. For flight time violations, this target point is assigned to the UAV with the lowest flight time. For capacity violations, this target point is assigned to the UAV with the lowest route demand.

[0221] The second mission plan of the long-flight UAV to be optimized is obtained through the preset lightweight MILP and heuristic adjustment strategy. The second mission plan is a monitoring mission plan:

[0222] Solve the lightweight MILP model to make decisions and monitor tasks;

[0223] Based on the statistical quartiles of the start time distribution of the mission sequence of the long-flight UAV to be optimized, a heuristic strategy is designed to solve the situation where the MILP model has no feasible solution. The feasible solution is the second mission plan.

[0224] In this embodiment of the present invention, the following parameters are set for the new Milp model:

[0225] :Indicates drone At the moment Is there a control task? This is a two-dimensional input data matrix.

[0226] :Indicates drone The take-off time is obtained from the HGA results.

[0227] :Indicates drone The landing time is obtained from the HGA results.

[0228] Indicates drone flight time period.

[0229] For drones The monitoring task set, .

[0230] Decision variables:

[0231] =1 if drone Monitoring tasks At the moment Start execution, otherwise 0.

[0232] =1 if drone Monitoring tasks At the moment is being executed, otherwise 0.

[0233] Building a lightweight MILP model:

[0234] (33)

[0235] (34)

[0236] (35)

[0237] (36)

[0238] (37)

[0239] (38)

[0240] (39)

[0241] The objective function means that there is no clear optimization goal, because we only need to find a feasible solution that satisfies the constraints.

[0242] Constraint (34) ensures that any time The number of drones that need GCS control tasks (including mission control tasks and monitoring tasks) cannot exceed .

[0243] Constraint (35) ensures that each UAV has additional The duration is monitoring tasks.

[0244] Constraints (36) Guarantees and The coupling from To begin, you need The drone was monitored at all times.

[0245] Constraint (37) ensures that the monitoring tasks of the same UAV meet the time interval .

[0246] Constraint (39) ensures that the scheduled monitoring tasks cannot conflict with the time of existing control tasks.

[0247] The heuristic adjustment strategy aims to find a feasible solution that satisfies the control constraints by re-adjusting the start time of the UAV's control mission while minimizing the increase in mission completion time.

[0248] The heuristic strategy will be adjusted based on the statistical quartiles of the time distribution of all drones to complete the task sequence. Quartiles are a key concept in statistics, which is used to divide the data points in the data set into four equal parts. This is particularly important when determining the threshold for adjusting the drone task time. By analyzing the completion time data, we can obtain the median of the task execution time (the second quartile, ), and first adjust the drone mission schedule accordingly to achieve a more balanced workload distribution. The key steps in strategy execution are as follows:

[0249] (1) Select the best solution. Select the four solutions with the smallest fitness values ​​from the first stage HGA solution results, and perform the following steps for each solution:

[0250] Quartile determination: Sort the completion time of all control tasks and determine the first quartile ( ), the second quartile( , also known as the median), the third quartile ( ).

[0251] Task time adjustment: First, As a benchmark, try to adjust the time of those controllable tasks that are ranked after the median and postpone them. If the MILP model still cannot find a solution after adjustment, then Adjust task time.

[0252] If the above adjustments do not provide a solution, try Adjust for the baseline.

[0253] If all adjustment attempts fail or the adjusted drone violates the flight time constraint, a penalty value is imposed on the solution and the mission completion time is set to M.

[0254] (2) Evaluation and selection of solutions, sorting the adjusted solutions according to task completion time,

[0255] 1) If the task completion time of all four solutions is equal to M, it indicates that there is no feasible solution. In this case, you need to select four solutions from the HGA results and repeat step (1).

[0256] 2) If there is a solution whose task completion time is less than M, the solution is selected as the final monitoring task solution.

[0257] The goal of this strategy is to find a feasible approximate solution when the MIP model solution conditions are not met, so as to ensure that all monitoring tasks can be completed within the specified time.

[0258] The issuing of the control task and the monitoring task according to the second task plan by using a preset heuristic method includes:

[0259] Arrange the control tasks in ascending order according to the start time to obtain the task set H;

[0260] Obtain the minimum task h of the task set H and distribute the tasks according to the preset distribution principle, which includes: obtaining the total working time of GCS , for all GCS according to the Sort in ascending order; get The smallest GCS is obtained, and the working status of the smallest GCS is obtained. If the working status is idle, h is directly assigned to the smallest GCS. If the working status is busy, the completion time of the control task being executed by the smallest GCS is obtained. If the completion time is later than the start time of h, h is assigned to The second smallest GCS.

[0261] like Figure 2 FIG. 1 is a diagram illustrating an optimization system for a long-flight UAV mission execution process provided by one embodiment of the present invention, the optimization system comprising:

[0262] The mission information acquisition module is used to obtain the mission information of the long-flight UAV to be optimized. The mission information includes: monitoring mission information, mission control mission information and flight related information. The monitoring mission information is: Time, spent through a ground control station GCS The flight status and working status of the onboard sensors of the long-flight UAV to be optimized are checked at the time. The control mission information is: when the long-flight UAV to be optimized is at the mission target point During the mission, the long-flight UAV to be optimized is controlled by a GCS to perform the mission control task, and the flight-related information includes: the departure time of the long-flight UAV to be optimized , the longest flight time , latest return time , the flight time required from GCS node a to GCS node b , at the mission target point The number of resources required is , the maximum number of resources carried by long-flight drones ;

[0263] Multi-U&G IOP model setting module, used to set the Multi-U&G IOP model for the integrated optimization problem of multi-flight-duration UAV mission planning and multi-GCS control resource allocation;

[0264] The mission optimization module is used to optimize the mission of the long-flight UAV to be optimized according to the mission information and the Multi-U&G IOP through a preset three-stage iterative optimization algorithm TSIOA.

[0265] Wherein, the task optimization module includes:

[0266] A first mission plan acquisition unit is configured to acquire a first mission plan of the long-flight UAV to be optimized by using a preset hybrid genetic algorithm (HGA), wherein the first mission plan is a mission control mission plan;

[0267] a second mission plan acquiring unit, configured to acquire a second mission plan for the long-flight UAV to be optimized by using a preset lightweight mixed integer linear programming model and a heuristic adjustment strategy, wherein the second mission plan is a monitoring mission plan;

[0268] A task issuing unit is used to issue the control task and the monitoring task according to the first and second task plans using a preset heuristic method.

[0269] The first task planning acquisition unit includes:

[0270] An initial solution obtaining unit, configured to obtain an initial solution of the Multi-U&G IOP model;

[0271] An optimized initial solution obtaining unit, configured to change the domain structure set of the initial solution through LS-VND to obtain an optimized initial solution;

[0272] A detection unit is used to detect the optimized initial solution through a preset fitness function.

[0273] The second task planning acquisition unit includes:

[0274] A monitoring task decision unit, used to solve the lightweight MILP model and decide on monitoring tasks;

[0275] The second task planning unit is used to design a heuristic strategy to solve the situation where the MILP model has no feasible solution based on the statistical quartiles of the start time distribution of the task sequence of the long-flight UAV to be optimized, where the feasible solution is the second task plan.

[0276] The task issuing unit includes:

[0277] The task set acquisition unit is used to arrange the targets and monitoring tasks in ascending order according to the start time to obtain the task set H;

[0278] The task allocation unit is used to obtain the minimum task h of the task set H and allocate tasks according to a preset allocation principle. The allocation principle includes: obtaining the total working time of the GCS , for all GCS according to the Sort in ascending order; get The smallest GCS is obtained, and the working status of the smallest GCS is obtained. If the working status is idle, h is directly assigned to the smallest GCS. If the working status is busy, the completion time of the task being executed by the smallest GCS is obtained. If the completion time is later than the start time of h, h is assigned to The second smallest GCS.

[0279] It should be understood that although the steps in the flowcharts of the accompanying drawings are shown in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some of the steps in the flowcharts of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0280] The above is only a partial implementation of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for optimizing the mission execution process of a long-flight UAV, characterized in that: The optimization method comprises: Obtain the task information of the long-flight UAV to be optimized, the task information includes: monitoring task information, control task information and flight related information; the monitoring task information is: every interval Time, spent through a ground control station GCS The flight status and working status of the onboard sensor of the long-flight UAV to be optimized are checked at the time; the mission information is: when the long-flight UAV to be optimized is at the target point During the mission, the long-flight UAV to be optimized is controlled by a GCS to perform the target mission; the flight-related information includes: the departure time of the long-flight UAV to be optimized , the longest flight time , latest return time , the flight time required from GCS node a to GCS node b , at the mission target point The number of resources required is , the maximum number of resources carried by long-flight drones ; Set up a Multi-U&G IOP model for the integrated optimization problem of multi-flight UAV mission planning and multi-GCS control resource allocation; Performing mission optimization on the long-flight UAV to be optimized according to the mission information and the Multi-U&G IOP using a preset three-stage iterative optimization algorithm TSIOA; The step of optimizing the mission of the long-flight UAV to be optimized according to the mission information and the Multi-U&G IOP by using the preset TSIOA includes: Obtaining a first mission plan of the long-flight UAV to be optimized by a preset hybrid genetic algorithm HGA, where the first mission plan is a mission control mission plan; Obtaining a second mission plan for the long-flight UAV to be optimized through a preset lightweight mixed-integer linear programming model and a heuristic adjustment strategy, where the second mission plan is a monitoring mission plan; The control task and monitoring task are issued according to the first and second task plans using a preset heuristic method.

2. The optimization method according to claim 1, wherein: The first mission plan of the long-flight UAV to be optimized is obtained by using a preset hybrid genetic algorithm HGA, where the first mission plan is a mission control mission plan, including: Obtaining an initial solution of the Multi-U&G IOP model; Changing the domain structure set of the initial solution by LS-VND to obtain an optimized initial solution; The optimized initial solution is tested using a preset fitness function.

3. The optimization method according to claim 1, wherein: The second mission plan of the long-flight UAV to be optimized is obtained by using a preset lightweight mixed integer linear programming model MILP and a heuristic adjustment strategy, wherein the second mission plan is a monitoring mission plan: Solve the lightweight MILP model and make decisions on monitoring tasks; According to the statistical quartiles of the start time distribution of the mission sequence of the long-flight UAV to be optimized, a heuristic strategy is designed to solve the situation where the MILP model has no feasible solution, and the feasible solution is the second mission plan.

4. The optimization method according to claim 1, wherein: The issuing of the control task and the monitoring task according to the first and second task plans by using a preset heuristic method includes: Arrange the targets and monitoring tasks in ascending order according to the start time to obtain the task set H; Obtain the minimum task h of the task set H and distribute the tasks according to the preset distribution principle, which includes: obtaining the total working time of GCS , for all GCS according to the Sort in ascending order; get The smallest GCS is obtained, and the working status of the smallest GCS is obtained. If the working status is idle, h is directly assigned to the smallest GCS. If the working status is busy, the completion time of the task being executed by the smallest GCS is obtained. If the completion time is later than the start time of h, h is assigned to The second smallest GCS.

5. An optimization system for a long-flight UAV mission execution process, characterized in that: The optimization system comprises: The mission information acquisition module is used to obtain the mission information of the long-flight UAV to be optimized. The mission information includes: monitoring mission information, mission control mission information and flight related information; the monitoring mission information is: Time, spent through a ground control station GCS The flight status and working status of the onboard sensors of the long-flight UAV to be optimized are checked at the time; the mission control mission information is: when the long-flight UAV to be optimized is at the mission target point During the mission, the long-flight UAV to be optimized is controlled by a GCS to perform the mission control task; the flight related information includes: the departure time of the long-flight UAV to be optimized , the longest flight time , latest return time , the flight time required from GCS node a to GCS node b , at the mission target point The number of resources required is , the maximum number of resources carried by long-flight drones ; Multi-U&G IOP model setting module, used to set the Multi-U&G IOP model for the integrated optimization problem of multi-flight-duration UAV mission planning and multi-GCS control resource allocation; A task optimization module is configured to optimize the task of the long-flight UAV to be optimized based on the task information and the Multi-U&G IOP using a preset three-stage iterative optimization algorithm TSIOA. The task optimization module includes: a first task plan acquisition unit, configured to acquire a first task plan of the long-flight UAV to be optimized using a preset hybrid genetic algorithm HGA, where the first task plan is a mission control task plan; a second mission plan acquiring unit, configured to acquire a second mission plan for the long-flight UAV to be optimized by using a preset lightweight mixed integer linear programming model and a heuristic adjustment strategy, wherein the second mission plan is a monitoring mission plan; A task issuing unit is used to issue the control task and the monitoring task according to the first and second task plans using a preset heuristic method.

6. The optimization system according to claim 5, wherein: The first task planning acquisition unit includes: An initial solution obtaining unit, configured to obtain an initial solution of the Multi-U&G IOP model; An optimized initial solution obtaining unit, configured to change the domain structure set of the initial solution through LS-VND to obtain an optimized initial solution; A detection unit is used to detect the optimized initial solution through a preset fitness function.

7. The optimization system according to claim 5, wherein: The second task plan acquisition unit includes: A monitoring task decision unit, used to solve the lightweight MILP model and decide on monitoring tasks; The second task planning unit is used to design a heuristic strategy to solve the situation where the MILP model has no feasible solution based on the statistical quartiles of the start time distribution of the task sequence of the long-flight UAV to be optimized, where the feasible solution is the second task plan.

8. The optimization system according to claim 5, wherein: The task issuing unit includes: The task set acquisition unit is used to arrange the targets and monitoring tasks in ascending order according to the start time to obtain the task set H; The task allocation unit is used to obtain the minimum task h of the task set H and allocate tasks according to a preset allocation principle. The allocation principle includes: obtaining the total working time of the GCS , for all GCS according to the Sort in ascending order; get The smallest GCS is obtained, and the working status of the smallest GCS is obtained. If the working status is idle, h is directly assigned to the smallest GCS. If the working status is busy, the completion time of the task being executed by the smallest GCS is obtained. If the completion time is later than the start time of h, h is assigned to The second smallest GCS.

Images