Emergency scene unmanned aerial vehicle task allocation method based on multi-objective optimization

Abstract

The application relates to the technical field of unmanned aerial vehicle scheduling, and discloses an emergency scene unmanned aerial vehicle task allocation method based on multi-target optimization, which realizes real-time acquisition of a task parameter set of an unmanned aerial vehicle through a standardized task interface, and simultaneously realizes real-time acquisition of an unmanned aerial vehicle state data set through an unmanned aerial vehicle management platform. Secondly, a multi-target particle swarm optimization algorithm and an improved PSO algorithm are adopted to calculate an optimal task allocation scheme, and a task preemption mechanism is combined to dynamically adjust a sudden emergency task and priority change; by providing three task allocation plans of a time optimal scheme, a resource optimal scheme and a parameter optimal scheme, it is ensured that a task can be completed in the shortest time or executed at the lowest cost, and the final scheme is provided to a command center for decision-making. The application improves the intelligence and flexibility of unmanned aerial vehicle scheduling, and realizes rapid response and global optimal allocation of resources of unmanned aerial vehicle scheduling in an emergency scene.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) scheduling technology, specifically to a method for allocating UAV tasks in emergency scenarios based on multi-objective optimization. Background Technology

[0002] In recent years, with the acceleration of standardization and intelligentization, drone scheduling has been deeply integrated into all aspects of social production, becoming a key carrier of new-quality productivity. Drone scheduling has always been a technological challenge, especially in complex environments. Traditional drone scheduling methods often rely on fixed algorithms, which cannot adapt to real-time changes in the environment or task requirements. However, AI-based drone scheduling systems can use deep learning and data analysis to comprehensively consider multiple factors such as weather, terrain, and task priority, optimizing flight paths and scheduling strategies in real time. This not only improves the efficiency of drone operations but also helps reduce energy consumption and extend operation time.

[0003] The existing technologies and their problems are as follows:

[0004] Heuristic scheduling methods: Heuristic algorithms seek suboptimal solutions that balance computational efficiency and allocation objectives when a direct optimal solution is difficult to obtain. Typical examples include tabu search (TS), simulated annealing (SA), and genetic algorithms (GA). These algorithms offer advantages such as ease of implementation, high computational speed, and broad applicability; however, they also have inherent limitations, including high computational complexity, sensitivity to initial parameters, and the inability to guarantee a globally optimal solution. In UAV task allocation, if pre-defined heuristic rules are used, this experience-based scheduling approach struggles to achieve globally optimal resource allocation when faced with complex and variable task requirements and limited UAV resources.

[0005] Scheduling methods based on centralized optimization algorithms: These methods use optimization algorithms such as integer programming or mixed integer programming for UAV mission scheduling. However, as the scale of the mission increases, the computational complexity of these methods becomes high, making it difficult to meet the real-time scheduling requirements.

[0006] Scheduling methods based on traditional Particle Swarm Optimization (PSO): The PSO algorithm originated from Kennedy and Eberhart's 1995 biomimetic research on bird flock foraging behavior. This algorithm dynamically adjusts inertia weights, social learning factors, and cognitive learning factors, combined with a fitness function to evaluate individual and global extrema, achieving optimal solution search under group cooperation. As an intelligent optimization method, PSO has advantages such as low computational complexity, strong robustness, outstanding global search capability, and high convergence efficiency. While using PSO for UAV mission optimization, its limitations in single-objective optimization make it difficult to simultaneously satisfy multiple optimization objectives.

[0007] The UAV scheduling method based on static task allocation first collects task requirements and UAV status information, then calculates a task allocation scheme through an optimization algorithm, and finally executes the scheduling. However, this method ignores the urgency of tasks, task dependencies, and dynamic adjustment of priorities, making it difficult to respond promptly to task changes in the event of emergencies or resource conflicts.

[0008] Drone scheduling methods with weak task preemption mechanisms: Some studies have explored task preemption mechanisms, but they usually only consider the insertion of emergency tasks without comprehensively modeling factors such as task dependencies, dynamic priority adjustment, and multi-task conflicts, resulting in insufficient system flexibility.

[0009] In summary, the shortcomings of existing technologies are as follows:

[0010] 1) Lack of dynamic task preemption capability:

[0011] Traditional scheduling systems use fixed task queues and lack a preemption mechanism based on real-time environmental changes (such as sudden tasks or resource shortages). Urgent tasks must wait for the queue to be cleared, resulting in delays in the response of critical tasks.

[0012] 2) Single optimization objective:

[0013] Existing technologies often employ a single algorithm to optimize a single metric (time or resource consumption), lacking the ability to optimize multiple objectives collaboratively, resulting in low resource utilization or increased risk of task timeout.

[0014] 3) Static matching model → Low device compatibility with task requirements:

[0015] Existing technologies use fixed weight allocation and do not dynamically adjust the matching degree model, resulting in a decrease in the matching degree between drones and tasks and an increase in the execution failure rate. Summary of the Invention

[0016] To address the aforementioned problems, the present invention aims to provide a UAV task allocation method for emergency scenarios based on multi-objective optimization. By combining multi-objective dynamic optimization with intelligent preemptive decision-making, it solves the multi-dimensional trade-offs in UAV scheduling under complex scenarios and provides hierarchical decision support for emergency response. Through a dynamic preemptive mechanism, it achieves second-level response for emergency tasks, reducing delays in critical missions. The technical solution is as follows:

[0017] The emergency scenario drone task allocation method based on multi-objective optimization includes the following steps:

[0018] Step 1: Collect various types of data through the external emergency command system, including geographical information of the disaster area, weather conditions, traffic conditions, and distribution of rescue resources. Obtain the set of task parameters from the external emergency command system in real time through the standardized task interface, and obtain the set of drone status data in real time through the drone management platform.

[0019] Step 2: Task allocation optimization is performed using a dual optimization channel, including a multi-objective particle swarm optimization algorithm and an improved particle swarm optimization algorithm. The objective particle swarm optimization algorithm defines the objective function as minimizing the latest task completion time and minimizing total resource consumption. It maintains the Pareto front solution set through non-dominated sorting and generates time-priority and resource-saving schemes by combining endurance, load, and time window constraints. The improved particle swarm optimization algorithm generates a dynamic parameter balancing scheme by combining a penalty function.

[0020] Step 3: Based on the dynamic task preemption mechanism, the urgency of the task queue is evaluated in real time. When a high-priority task is detected to be inserted or environmental parameters exceed the limit, the preemption mechanism is triggered to release the drone resources occupied by low-priority tasks and prioritize the allocation of drones that meet the set conditions to the urgent task.

[0021] Step 4: Push the generated three-level optimization plan to the external emergency command system through the visualization terminal for manual decision-making, and then execute task allocation based on the decision results, provide real-time feedback on the drone status and task progress, dynamically trigger secondary optimization or fault tolerance, and form a closed-loop scheduling process.

[0022] The beneficial effects of this invention are:

[0023] 1) This invention effectively balances efficiency and compliance in task allocation by combining the multi-objective particle swarm optimization (MOPSO) algorithm and an improved PSO algorithm with a multi-dimensional penalty function, rapidly generating feasible solutions under complex constraints. It optimizes UAV emergency task scheduling, achieving efficient task allocation and rational resource scheduling. Compared to traditional static task allocation methods, this invention introduces a dual-interface interaction mechanism between task and UAV status, ensuring real-time synchronization of task information and UAV status, improving scheduling accuracy and flexibility. Simultaneously, the improved MOPSO algorithm balances minimizing task completion time and minimizing total UAV energy consumption, ensuring the scheduling scheme considers both task response speed and resource utilization efficiency.

[0024] 2) This invention designs a three-level optimal task scheduling scheme generation mechanism, providing time-priority, energy-optimal, and parameter-based schemes for different emergency scenarios. These schemes are then pushed to the command center via a decision interface. A dynamic constraint model for wind resistance and endurance reduces task execution risk by quantifying flight redundancy time and matching it with capabilities. This achieves intelligent assisted decision-making and closed-loop feedback control. Compared to traditional heuristic scheduling methods, this invention can dynamically adjust and optimize strategies in complex task environments, improving the task execution capabilities of UAV swarms and providing reliable technical support for the application of UAVs in emergency rescue scenarios. Attached Figure Description

[0025] Figure 1 This is a flowchart of the emergency scenario drone task allocation method based on multi-objective optimization according to the present invention. Detailed Implementation

[0026] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0027] This invention proposes a method and system for UAV task allocation based on the multi-objective particle swarm optimization (MOPSO) algorithm, and introduces a task preemption mechanism, which can optimize task scheduling in dynamic environments, improve task completion efficiency and UAV utilization.

[0028] This invention processes real-time status data of UAVs and missions, including parameters such as UAV speed, payload capacity, wind resistance, endurance, and cost per flight, as well as mission information such as time constraints, resource requirements, urgency, and time windows. Secondly, it employs a multi-objective particle swarm optimization (MOPSO) algorithm and an improved PSO algorithm to calculate the optimal task allocation scheme. Simultaneously, it incorporates a task preemption mechanism to dynamically adjust for sudden emergency tasks and priority changes. By providing three task allocation contingency plans—time-optimal, resource-optimal, and parameter-optimal—it ensures that missions can be completed in the shortest possible time or at the lowest cost. The final plan is then provided to the command center for decision-making, improving the intelligence and flexibility of UAV scheduling and achieving rapid response and globally optimal resource allocation in emergency scenarios.

[0029] The solution process can be roughly divided into the following steps:

[0030] First, emergency task parameters, including timeliness requirements, resource needs, and urgency level, are injected in real time through the standardized task interface provided by the system. Simultaneously, real-time status data of the drone fleet uploaded by the drone management platform, including location, battery life, and load capacity, are synchronized. The system dynamically evaluates the task queue based on a multi-dimensional preemption rule base. When a high-priority task is detected to be inserted or environmental parameters exceed limits, a preemption mechanism is triggered to release low-priority task resources and prioritize the allocation of high-scoring drones to emergency tasks.

[0031] Subsequently, the multi-objective particle swarm optimization (MOPSO) algorithm was used for parallel computation: the UAV cluster was initialized as a particle swarm, and a dual objective function was defined: time optimization—minimizing the latest task completion time and resource optimization—minimizing total energy consumption. The Pareto front solution set was maintained by non-dominated sorting, and three-level contingency plans were generated by combining endurance, payload, time window constraints and penalty functions: time priority scheme, resource saving scheme and dynamic parameter balancing scheme.

[0032] The optimization results are pushed through the command center's visual terminal, supporting manual decision-making and selection. During execution, the drone's status and mission progress are fed back in real time, and secondary optimization or fault tolerance are dynamically triggered.

[0033] Finally, the archived task allocation data and algorithm parameters form a dynamic scheduling system covering the entire process from real-time response and multi-objective optimization to closed-loop feedback, enabling efficient allocation of drone resources and second-level preemptive response in emergency scenarios.

[0034] The specific flowchart of the present invention is as follows: Figure 1 As shown.

[0035] 1. Task access phase:

[0036] A. The external emergency command system dynamically injects a set of task parameters through a standardized task interface.

[0037] Let the task parameter set be: ;

[0038] in, For a set of tasks, For the task The set of attributes for each task It consists of multiple attributes:

[0039] ;

[0040] in, As a unique identifier for the task, Due to the time constraints of the task, Resources required for the task. The coordinates of the mission location. For task type, For the wind resistance required for the mission, Depending on the urgency of the task, The number of drones required for the mission. This is a list of drones that have been assigned to this task.

[0041] Meanwhile, the drone management platform uploads drone status data in real time via an interface. Let the drone status data set be:

[0042] ;

[0043] in, For a collection of drones, the status parameters of each drone are... for:

[0044] ;

[0045] in, For drone ID, Location of the drone. This refers to the actual flight time of the drone. Whether the drone is busy To score points for the drone, This refers to the drone's flight time (without weight). For the drone's flight speed, For the payload capacity of drones, For tasks that drones can perform, To enhance the wind resistance of drones, Cost per flight for the mission.

[0046] The main mission type parameters are shown in Table 1, and the main UAV type parameters are shown in Table 2.

[0047] Table 1 Main Task Type Parameters

[0048] .

[0049] Table 2. Parameters of Main UAV Types

[0050] .

[0051] 2. Combined application performance modules:

[0052] This emergency scene drone combined application efficiency algorithm primarily provides drone task priority analysis. Based on the emergency scenario and task urgency, the algorithm analyzes the priority of different tasks to ensure that high-priority tasks are executed first. Considering multi-drone collaborative task execution, it achieves collaborative planning of task allocation to improve emergency rescue efficiency. The algorithm intelligently analyzes drone technical capabilities, payload information, and other data to rationally allocate tasks, ensuring that each drone achieves maximum effectiveness. Specifically, this algorithm aims to dispatch high-efficiency, high-resource-consuming drones in urgent situations, and dispatch drones with low resource utilization that can complete the task within a specified time in less urgent situations.

[0053] A. The basic idea of ​​the efficiency algorithm for combined application of UAVs in emergency situations mainly includes the following aspects:

[0054] a) Modeling of UAV mission allocation;

[0055] Task allocation is modeled as a multi-objective optimization problem, with optimization objectives including minimizing task completion time and minimizing resource consumption.

[0056] Pareto optimal solutions are used to provide multiple candidate solutions to meet different scheduling requirements.

[0057] b) Optimization methods based on MOPSO;

[0058] The Multi-Objective PSO (MOPSO) problem is used to solve the UAV scheduling problem. The task allocation scheme is continuously optimized through individual optimal and global optimal update strategies. Non-dominated sorting and congestion calculation are used to maintain the Pareto front, ensuring the diversity and coverage of the solution set.

[0059] c) Dynamically adjust the drone's status to adapt to the mission;

[0060] During mission execution, the drone status is dynamically updated to ensure that the allocation plan matches the real-time situation. The drone management platform provides real-time feedback on mission execution and adjusts mission allocation accordingly.

[0061] d) Optimize the delivery and execution of scheduling plans;

[0062] The human-computer interaction terminal in the command center displays time priority and resource priority to help commanders make decisions.

[0063] The above measures can significantly improve the overall efficiency of large unmanned helicopters and medium-sized compound-wing UAVs and reduce energy consumption, thereby enabling them to play a greater role in various application scenarios.

[0064] B. Technical Approach:

[0065] a) Input of mission parameters and UAV status;

[0066] The external emergency command system injects a set of task parameters, including task type, location, time constraints, and resource requirements, through a standardized interface.

[0067] The drone management platform uploads the current status of the drone fleet, including location, battery level, payload capacity, wind resistance, and mission execution capability, through a capability interface.

[0068] b) Particle swarm initialization;

[0069] Within the mission area, drones can be used as decision variables for particles, with each particle's dimension being the number of drones. The particle positions (task allocation scheme) and velocities are initialized, and tasks are randomly assigned to drones.

[0070] In the multi-objective PSO (MOPSO) optimization process, the task allocation scheme is modeled as a particle swarm, where each particle represents a task allocation scheme.

[0071] The position of each particle is initialized to the task assignment of the drone (a drone can be assigned multiple tasks, and tasks can also be performed by multiple drones).

[0072] Set the initial velocity of the particles for subsequent update iterations.

[0073] Calculate the objective function for the particle:

[0074] Task completion time: The first objective is to minimize the latest task completion time, calculated using the following formula:

[0075] ;

[0076] in, For task completion time, For drones The completion time for its last task is calculated as follows:

[0077] ;

[0078] in, For drones The task start time, For drones Distance from current location to the task point For drones Flight speed, For drones Time required to perform the task;

[0079] ;

[0080] in, The distance to the task is precisely calculated using the Harvard function.

[0081] The goal is to minimize the completion time of the entire task set: .

[0082] Resource consumption: The objective is to minimize the total cost of all drones performing the mission.

[0083] ;

[0084] in, For resource consumption, Cost per flight for the mission.

[0085] c) Multi-objective PSO iterative optimization;

[0086] Calculation of individual optimality and global optimality:

[0087] Calculate the current fitness of the particles (task completion time + resource consumption) and update the individual optimal solution for each particle.

[0088] Maintain an external file to store the Pareto front solutions (the set of non-dominated solutions) of the current population.

[0089] Particle position and velocity updates:

[0090] Using the speed update formula of MOPSO:

[0091] ;

[0092] in, Inertial weights are used to control the smoothness of particle movement; and This is a learning factor used to determine the degree to which a particle is influenced by the individual optimal solution and the global optimal solution; and These are random numbers, used to increase the randomness of the search; This is the particle's current position. The current velocity of the particle. The velocity of the particle after the update; This is the optimal solution for the individual. This is the globally optimal solution.

[0093] Update particle positions:

[0094] ;

[0095] Perform non-dominated sorting, maintain the Pareto front, and select the optimal solution set for the next round of optimization.

[0096] Iteration termination condition:

[0097] The iteration terminates when the preset limit is reached or the optimal solution does not change significantly over multiple generations. The fitness of each particle (task completion time + resource consumption) is calculated, and its historical best is updated. A Pareto front solution set is maintained, and non-dominated solutions are selected as external archives for global guidance. Particle velocity and position are updated according to the particle velocity update formula, combining individual and global best values. Iteration continues until the maximum number of generations is reached, outputting the final optimized solution set.

[0098] d) Constraints;

[0099] During the optimization process, we need to ensure that the solution meets a series of physical and task constraints.

[0100] Task constraints:

[0101] Each task must be performed by at least one drone:

[0102] ;

[0103] in, It is the first decision variable, representing the drone. Should the task be executed? , A collection of drones.

[0104] Battery life constraints:

[0105] The total energy consumption of a drone for mission execution cannot exceed its current remaining battery life.

[0106] ;

[0107] in, For drones Maximum available battery life Due to time constraints;

[0108] Load constraints:

[0109] If any goods need to be transported, the drone's payload capacity must be met:

[0110] ;

[0111] in, For drones Maximum load capacity;

[0112] Task time window constraints:

[0113] Some tasks may have strict time windows, requiring them to be executed within the available time:

[0114]

[0115] in, and These are tasks The earliest and latest execution times, It is a drone The actual time taken to perform the task.

[0116] Uniqueness constraint for drone scheduling:

[0117] A drone can only perform one task at a time.

[0118] ;

[0119] Decision variables:

[0120] In the MOPSO optimization process, we mainly need to define decision variables applicable to the task scheduling problem:

[0121] Task allocation variables ;

[0122] This variable determines the drone Whether to be assigned to perform tasks .

[0123] Drone flight speed:

[0124] ;

[0125] This indicates the drone's flight speed, ranging between the minimum and maximum permissible values. Due to the complexity of emergency scenarios, the actual speed is controlled by the pilot.

[0126] Solution generation and decision delivery:

[0127] Generate a Level 2 contingency plan:

[0128] Optimal time solution: Prioritize ensuring the shortest possible task completion time.

[0129] Optimal resource solution: Minimize drone energy consumption within a limited time, or generate results that exceed the time limit.

[0130] 3. Parameter optimization module:

[0131] A. Solve the task-drone matching problem by mapping drone indices using particle encoding and combining them with a penalty function;

[0132] Particle coding:

[0133] ;

[0134] The objective function minimizes the total penalty value assigned to the task, and includes the following penalty terms:

[0135] Invalid index penalty The drone index is out of range.

[0136] Punishment for incompetence The drone cannot meet the mission's capabilities, wind resistance, or time requirements.

[0137] Reuse penalty The same drone was assigned to multiple missions.

[0138] Penalty for Duplicate Assignment of the Same Task The same drone was assigned to the same task multiple times.

[0139] The ultimate goal is to minimize the sum of all penalty terms to ensure task allocation:

[0140] ;

[0141] Define the second decision variable :

[0142] ;

[0143] Among them, the second decision variable It is a binary variable representing the drone. Is it assigned to a task? If it is 1, it indicates allocation; if it is 0, it indicates no allocation.

[0144] Invalid index penalty:

[0145] ;

[0146] Penalty for insufficient ability:

[0147] ;

[0148] Penalty for repeated use:

[0149] ;

[0150] Penalty for assigning the same task repeatedly:

[0151] ;

[0152] B. Constraints;

[0153] Task requirements constraints:

[0154] ;

[0155] Unmanned aerial vehicle (UAV) capability constraints:

[0156] The drone must support the mission type;

[0157] The wind resistance of the drone must meet the mission requirements;

[0158] The drone must be idle;

[0159] The mission time limit must be less than the drone's remaining flight time (round-trip flight time plus redundancy). ):

[0160] ;

[0161] ;

[0162] in, , , and These indicate whether the task type is supported, whether wind resistance is achieved, and whether it is a drone. If the drone is idle and its battery life exceeds the mission time, set the value to 1; otherwise, set the value to 0. This is for redundancy.

[0163] Resource uniqueness constraint:

[0164] A drone cannot be repeatedly assigned to multiple tasks, and the same drone cannot be repeatedly assigned to the same task.

[0165] ;

[0166] Urgent task priority constraints:

[0167] Define the emergency task:

[0168] ;

[0169] ;

[0170] in, This is a subset of emergency tasks.

[0171] Emergency mission resource constraints:

[0172] ;

[0173] in, For drones Assign to an emergency task? , For emergency mission Minimum number of drones required This represents the total number of available drones.

[0174] Urgent task priority constraints:

[0175] ;

[0176] The penalty coefficient for urgent tasks is much greater than that for non-urgent tasks, which makes the algorithm prioritize satisfying the constraints of urgent tasks.

[0177] C. Dynamic task preemption mechanism;

[0178] By using priority judgment and resource release rules, high-priority tasks can dynamically preempt drone resources from low-priority tasks, and idle drones can be allocated preferentially.

[0179] Conditions for seizing: and ;

[0180] Resource release: (Release the seized drones);

[0181] D. Solution generation and decision delivery;

[0182] Generate contingency plan: Optimal parameter contingency plan.

[0183] 4. The generated three-level optimization plan is pushed to the external emergency command system through the visualization terminal for manual decision-making. Based on the decision results, the task allocation is executed, the drone status and task progress are fed back in real time, and secondary optimization or fault tolerance is dynamically triggered to form a closed-loop scheduling process.

[0184] In summary, this invention effectively balances efficiency and compliance in task allocation by combining the Multi-Objective Particle Swarm Optimization (MOPSO) algorithm and an improved PSO algorithm with a multi-dimensional penalty function, rapidly generating feasible solutions under complex constraints. It optimizes UAV emergency task scheduling, achieving efficient task allocation and rational resource scheduling. Compared to traditional static task allocation methods, this invention introduces a dual-interface interaction mechanism between task and UAV status, ensuring real-time synchronization of task information and UAV status, improving scheduling accuracy and flexibility. Simultaneously, the improved MOPSO algorithm balances minimizing task completion time and minimizing total UAV energy consumption, ensuring the scheduling scheme considers both task response speed and resource utilization efficiency.

Claims

1. A method for allocating unmanned aerial vehicle (UAV) tasks in emergency scenarios based on multi-objective optimization, characterized in that: Includes the following steps: Step 1: Collect various types of data through the external emergency command system, including geographical information of the disaster area, weather conditions, traffic conditions, and distribution of rescue resources. Obtain the set of task parameters from the external emergency command system in real time through the standardized task interface, and obtain the set of drone status data in real time through the drone management platform. Step 2: Task allocation optimization is performed using a dual optimization channel, including a multi-objective particle swarm optimization algorithm and an improved particle swarm optimization algorithm. The objective particle swarm optimization algorithm defines the objective function as minimizing the latest task completion time and minimizing total resource consumption. It maintains the Pareto front solution set through non-dominated sorting and generates time-priority and resource-saving schemes by combining endurance, load, and time window constraints. The improved particle swarm optimization algorithm generates a dynamic parameter balancing scheme by combining a penalty function. Step 3: Based on the dynamic task preemption mechanism, the urgency of the task queue is evaluated in real time. When a high-priority task is detected to be inserted or environmental parameters exceed the limit, the preemption mechanism is triggered to release the drone resources occupied by low-priority tasks and prioritize the allocation of drones that meet the set conditions to the urgent task. Step 4: Push the generated plan to the external emergency command system through the visualization terminal for manual decision-making, and then execute task allocation based on the decision results, provide real-time feedback on the drone status and task progress, dynamically trigger secondary optimization or fault tolerance, and form a closed-loop scheduling process. Step 2, the task allocation optimization using the multi-objective particle swarm optimization algorithm, specifically includes: Step 311: Use the available drones within the task area as decision variables for particles, with the dimension of each particle being the number of drones; model the task allocation scheme as a particle swarm, where each particle represents a task allocation scheme; initialize the position of each particle to the task allocation status of the drones, and set the initial velocity of the particles for subsequent updates and iterations. Step 312: Calculate the objective function of the particle; The formula for minimizing the latest task completion time is: ； in, For task completion time, For drones The completion time of its last task is calculated using the following formula: ； in, For drones The task start time, For drones Distance from current location to the task point For drones Flight speed, For drones Time required to perform the task; ； in, For Harvard functions; Minimize the total cost of all drones performing the mission: ； in, For resource consumption, Cost per flight for the mission; Step 313: Iterative optimization of the multi-objective particle swarm optimization algorithm; Calculation of individual optimality and global optimality: Calculate the current fitness of the particles, i.e. the task completion time and resource consumption, and update the individual optimal solution for each particle; Maintain an external file to store the Pareto front solutions of the current population; Particle position and velocity updates: The speed update formula using the multi-objective particle swarm optimization algorithm: ； in, Inertial weights are used to control the smoothness of particle movement; and This is a learning factor used to determine the degree to which a particle is influenced by the individual optimal solution and the global optimal solution; and These are random numbers, used to increase the randomness of the search; This is the particle's current position. The current velocity of the particle. The velocity of the particle after the update; This is the optimal solution for the individual. This is the globally optimal solution; Update particle positions: ； in, The updated position of the particle; Perform non-dominated sorting, maintain the Pareto front, and select the optimal solution set for the next round of optimization; Iteration termination condition: The number of iterations reaches a preset limit, or the optimal solution is found in consecutive n... d If the degree of change in each generation does not exceed the expected level, the iteration is terminated; the fitness of each particle is calculated and the individual historical best is updated; the Pareto front solution set is maintained, and non-dominated solutions are selected as external archives for global guidance; the particle velocity and position are updated according to the particle velocity update formula, combined with the individual best and the global best; the iteration continues until the maximum number of generations is reached, and the final optimized solution set is output; Step 314: Set constraints; 1) Define decision variables; Define the first decision variable applicable to the task scheduling problem. : ； 2) Task constraints; Each task must be performed by at least one drone, as shown below: ； 3) Battery life constraints; The total energy consumption of a drone for mission execution cannot exceed its current remaining battery life, as shown below: ； in, For drones Maximum available battery life Due to time constraints; 4) Load constraints; Any transport of goods must meet the payload capacity of the drone, as shown below: ； in, For drones Maximum load capacity; 5) Task time window constraints; The task is executed within the available time, as shown below: ； in, and These are tasks The earliest and latest execution times, It is a drone The actual time it took to execute the task; 6) Uniqueness constraint for drone scheduling; A drone can only perform one task at a time, as shown below: ； 7) Unmanned aerial vehicle (UAV) flight speed constraints; ； Indicates the flight speed of the drone The range is within the minimum allowable speed. and maximum speed between; Step 2, which uses an improved particle swarm optimization algorithm to optimize task allocation, specifically includes: Step 321: Solve the task-drone matching problem by mapping drone indices through particle encoding and combining it with a penalty function; the particle encoding is: ； in, For the first Various task allocation schemes Assign locations to all drones globally; Step 322: Define the following penalty items: Invalid index penalty The drone index is out of range. Punishment for incompetence The drone cannot meet the mission's capabilities, wind resistance, or time requirements; Reuse penalty The same drone was assigned to multiple tasks; Penalty for Duplicate Assignment of the Same Task The same drone was assigned to the same task multiple times; The ultimate goal is to minimize the sum of all penalty terms: ； in, α , , and The weights of each item; Define the second decision variable : ； Among them, the second decision variable It is a binary variable representing the drone. Is it assigned to a task? If the value is 1, it indicates allocation; if the value is 0, it indicates no allocation; then: ； ； ； ； in, For effective drone indexing, For the first The drone was assigned to the first One task; , , and These indicate whether the drone is not a valid index, and the drone itself. Is the task unable to be performed? drones Is it assigned to the first One mission and the same drone Whether the user has been assigned to the same task multiple times; Step 323: Set constraints: 1) Task requirement constraints: ； in, To carry out the mission The required number of drones; 2) Unmanned aerial vehicle (UAV) capability constraints: The drone must support the mission type, its wind resistance must meet the mission requirements, the drone must be idle, and the mission time limit must be less than the drone's remaining flight time, as shown in the following formula: ； ； in, , , and These indicate whether the task type is supported, whether wind resistance is achieved, and whether it is a drone. Whether the drone is idle and whether its battery life exceeds the mission timeframe. This is for redundancy; 3) Resource uniqueness constraint: A drone cannot be reassigned to multiple tasks, and the same drone cannot be reassigned to the same task, as expressed in the following way: ； 4) Emergency Task Constraints: Define the emergency task: ； ； in, A subset of emergency tasks; Assess the urgency of the task; Emergency mission resource constraints: ； in, For drones Assign to an emergency task? , For emergency mission Minimum number of drones required This represents the total number of available drones. Urgent task priority constraints: ； in, , , and These are penalties for invalid indexes, insufficient capacity, reuse, and duplicate assignment of the same task for urgent tasks. , , and These are the corresponding weighting coefficients.

2. The emergency scenario UAV task allocation method based on multi-objective optimization according to claim 1, characterized in that, In step 1, the task parameter set is... Defined as: ； in, For a set of tasks, each task It consists of multiple attributes: ； in, For the task The set of attributes, As a unique identifier for the task, Due to the time constraints of the task, Resources required for the task. The coordinates of the mission location. For task type, For the wind resistance required for the mission, Depending on the urgency of the task, The number of drones required for the mission. List of drones already assigned to this task; The drone status data set Defined as: ； in, For a collection of drones, the status parameters of each drone are... for: ； in, For drone ID, Location of the drone. This refers to the actual flight time of the drone. Whether the drone is busy To score points for the drone, For drone battery life, For the drone's flight speed, For the payload capacity of drones, For tasks that drones can perform, To enhance the wind resistance of drones, Cost per flight for the mission.

3. The emergency scenario UAV task allocation method based on multi-objective optimization according to claim 2, characterized in that, Step 3 involves establishing a dynamic task preemption mechanism, specifically as follows: By using priority judgment and resource release rules, high-priority tasks can dynamically preempt drone resources from low-priority tasks, and idle drones can be allocated preferentially. Conditions for seizing: and ; Resource release: This means releasing the seized drones; in, , , and These are the newly arrived high-priority tasks, the current low-priority tasks, and the [number]th [task]. The endurance of drones and the collection of drones whose resources are being taken over by high-priority tasks.

Images