A particle swarm-based clean energy station multi-unmanned aerial vehicle task allocation method and device

Abstract

The present application relates to a kind of particle swarm-based clean energy station multi-unmanned aerial vehicle task allocation method, comprising: obtaining clean energy station area data information, and the division of patrolling area is carried out;Clean energy station multi-unmanned aerial vehicle task allocation model is constructed;According to individual optimal position and global optimal position, the speed and position of particle are adjusted, and objective function is optimized;When the iteration number reaches upper limit, the task allocation solution corresponding to global optimal particle is returned, otherwise, continue to update particle state, optimization solution.Nonlinear dynamic collaborative improvement is carried out to inertia weight and learning factor, inertia weight adopts nonlinear self-adaptive decreasing strategy, iteration initial period is kept larger value, enhances the global traversal ability of unmanned aerial vehicle formation, widely searches various sub-regions and task combination;Rapidly attenuate in iteration later period, strengthen local precision search, ensure fast convergence to optimal cost combination.Golden sinusoidal algorithm is fused to reconstruct position update mechanism, realize the dynamic balance of global and local search.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) task allocation technology, and in particular to a method and device for multi-UAV task allocation at clean energy power stations based on particle swarm optimization. Background Technology

[0002] With the continuous development of drone technology, drones have been widely used in various industries, especially in the monitoring and maintenance of clean energy plants, where they can effectively perform tasks such as equipment inspection and fault detection. To improve the operational efficiency of drone swarms, efficient task allocation within multi-drone systems has become a key issue.

[0003] Traditional task allocation methods, such as greedy algorithms and auction algorithms, can solve some simple task allocation problems, but they often fail to achieve global optimization when tasks are complex, there are many drones, or tasks have time constraints or resource limitations. This limitation becomes even more apparent in environments like clean energy power plants, where there is a large amount of equipment and heavy inspection workload.

[0004] Particle Swarm Optimization (PSO) is a swarm intelligence optimization algorithm with strong global search capabilities, but it may face problems such as local optima or slow convergence speed in task allocation. Summary of the Invention

[0005] To address the problems of existing technologies, such as difficulty in achieving global optimization, poor adaptability to clean energy power station scenarios, and inability to meet the efficient collaborative requirements of large-scale inspection tasks, the primary objective of this invention is to provide a particle swarm optimization-based multi-UAV task allocation method for clean energy power stations that achieves global optimization, enhances adaptability to complex terrains at clean energy power stations, and effectively meets the actual needs of large-scale inspection tasks for dynamic collaboration and efficient scheduling.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a multi-UAV task allocation method for clean energy power stations based on particle swarm optimization, the method comprising the following sequential steps:

[0007] (1) Obtain regional data information of clean energy power stations and divide the inspection areas;

[0008] (2) Construct a multi-UAV task allocation model for clean energy stations based on the divided inspection areas; integrate the golden sine algorithm into the particle swarm algorithm, and improve the inertial weight and learning factor of the particle swarm algorithm;

[0009] (3) Calculate the fitness value based on the objective function, update the individual optimal position of the particle, determine the global optimal position, and adjust the particle's velocity and position based on the individual optimal position and the global optimal position to optimize the objective function;

[0010] (4) Set the total number of iterations. Each iteration will update the particle's fitness value, individual optimal position, global optimal position, and adjust the particle's speed and position. When the number of iterations reaches the upper limit, return the task allocation solution corresponding to the global optimal particle. Otherwise, continue to update the particle state and optimize the solution.

[0011] Step (1) specifically refers to: obtaining the geographical location information of the overall inspection area of ​​the clean energy station, performing data modeling on the overall inspection area, and using the modeling data to divide the overall inspection area into sub-regions to obtain the boundaries and main features of each inspection area. The main features include task attributes, spatial geometric attributes, and environmental constraint attributes. The task attributes include equipment type, load requirements, and defect density. The spatial geometric attributes include area, aspect ratio, and terrain undulation. The environmental constraint attributes include electromagnetic interference, wind speed, and take-off and landing conditions.

[0012] In step (2), the clean energy station multi-UAV task allocation model includes constraints and objective functions; the constraints include one task per UAV, single UAV energy consumption upper limit constraint, and task priority timing constraint.

[0013] The one-task-one-machine constraint is as follows:

[0014] Each task can only be assigned to one drone, each particle's dimension can only take one value, and each task can only be executed by one drone:

[0015] , ;

[0016] in, =1 indicates that task i is assigned to drone j; otherwise... =0; This refers to the total number of drones; It is the total number of tasks;

[0017] The upper limit constraint on the single-machine energy consumption is:

[0018] Each drone consumes a certain amount of energy while performing a mission, but the energy consumption shall not exceed the drone's maximum battery capacity.

[0019] , ;

[0020] in, It is the energy consumption of task i. This is the maximum battery capacity of drone j;

[0021] The task priority timing constraint is as follows:

[0022] High-priority tasks should be completed earlier than low-priority tasks.

[0023] if ;

[0024] in, It is the priority of task i. It is the priority of task i+1. This indicates that task i has a higher priority than task i+1; For the moment when task i is executed, This refers to the moment when task i+1 is executed;

[0025] The formula for the objective function is:

[0026] minf = + + ;

[0027] in, , , These are all weighting coefficients for the target.

[0028] In step (2), the improvement of the inertia weight and learning factor of the particle swarm algorithm specifically refers to:

[0029] Improved inertia weight for:

[0030] = ;

[0031] in, This indicates the maximum number of iterations. Indicates the current iteration number. This represents the maximum value of the inertia weight. This represents the difference between the maximum and minimum values ​​of the inertia weights;

[0032] The improvement of the learning factor specifically refers to:

[0033] Improved individual learning factor for:

[0034] = - ;

[0035] Improved social learning factor for:

[0036] = + ;

[0037] In the formula, This represents the maximum degree of trust a particle has in its own experience. This represents the minimum degree to which a particle adopts the collective experience; The effective interval for the learning factor.

[0038] Step (3) specifically refers to:

[0039] The velocity formula is as follows:

[0040] = + + ;

[0041] in, and All are random numbers in the range [0,1]. and These are the individual optimal position and the global optimal position, respectively. It is the velocity of the particle after the (t+1)th iteration. It is the velocity of the particle after the t-th iteration. It is the position of the particle after the t-th iteration;

[0042] By incorporating the golden sine algorithm into the particle swarm optimization algorithm, an improved position update formula is obtained:

[0043] = - ;

[0044] = a +b ;

[0045] = a +b ;

[0046] in, It is the step size direction factor, with a value range of (0, 2). ); Let be the position disturbance factor, in (0, Take the value within ) and These are the coefficients derived from the golden ratio. It is the golden ratio, with a value of a and b define the scope of the golden ratio search, where a is the dynamic lower bound factor and b is the static upper bound factor. It is the position of the particle after the (t+1)th iteration.

[0047] Step (4) specifically refers to: when the number of iterations reaches the upper limit, returning the task allocation solution corresponding to the globally optimal particle; otherwise, continuing to update the particle state and optimize the solution; and determining whether the current number of iterations has reached the set maximum number of iterations. If the maximum number of iterations has not been reached, the current iteration count is incremented by 1, the learning factor and inertia weight are recalculated, and the particle's velocity and position are updated.

[0048] Another object of the present invention is to provide an electronic device comprising:

[0049] Processor; and

[0050] A memory storing computer program instructions that, when executed by the processor, cause the processor to perform the particle swarm optimization-based multi-UAV task allocation method for clean energy stations as described above.

[0051] The present invention also provides a computer-readable storage medium having stored thereon computer program instructions, which, when executed by a processor, cause the processor to perform the particle swarm optimization-based multi-UAV task allocation method for clean energy power stations as described above.

[0052] As can be seen from the above technical solution, the beneficial effects of the present invention are as follows: First, nonlinear dynamic collaborative improvement of inertial weight and learning factor: the inertial weight adopts a nonlinear adaptive decreasing strategy, maintaining a large value in the early stage of iteration to enhance the global traversal capability of the UAV formation, enabling efficient search of various sub-regions and task combinations in the wide-area scenario of clean energy stations, demonstrating strong global optimization potential; in the later stage of iteration, it decays rapidly to strengthen local fine search and ensure rapid convergence to the optimal cost combination; in terms of learning factor, the individual learning factor decreases nonlinearly from a high initial value, and the social learning factor increases nonlinearly from a low initial value, so that the particles in the early stage of the algorithm focus on their own experience to explore the wide area, avoiding premature convergence; In the later stages, the group information weight gradually becomes dominant, accelerating the aggregation towards the global optimum. This improvement effectively balances individual cognition and group sharing, significantly improving the convergence speed and solution quality in complex inspection tasks, and providing core support for large-scale collaborative needs. Second, the position update mechanism is reconstructed by integrating the golden sine algorithm to achieve a dynamic balance between global and local search. The periodic fluctuation characteristics of the sine function are used to nonlinearly modulate the particle movement trajectory, enabling the UAV to both approach the optimal area and oscillate in the opposite direction to escape local traps. This enhances the adaptability to complex terrains of clean energy stations and effectively meets the actual needs of large-scale inspection tasks for dynamic collaboration and efficient scheduling while ensuring global optimization accuracy. Attached Figure Description

[0053] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0054] like Figure 1 As shown, a method for multi-UAV task allocation at clean energy power stations based on particle swarm optimization includes the following sequential steps:

[0055] (1) Obtain regional data information of clean energy power stations and divide the inspection areas;

[0056] (2) Construct a multi-UAV task allocation model for clean energy stations based on the divided inspection areas; integrate the golden sine algorithm into the particle swarm algorithm, and improve the inertial weight and learning factor of the particle swarm algorithm;

[0057] (3) Calculate the fitness value, i.e. the objective function value, based on the objective function, update the individual optimal position of the particle, determine the global optimal position, and adjust the particle's velocity and position based on the individual optimal position and the global optimal position to optimize the objective function;

[0058] (4) Set the total number of iterations. Each iteration updates the particle's fitness value, individual optimal position, global optimal position, and adjusts the particle's speed and position. When the number of iterations reaches the upper limit, return the task allocation solution corresponding to the globally optimal particle; otherwise, continue updating the particle state and optimizing the solution. The iteration includes updating the particle's fitness value, individual optimal position, global optimal position, and adjusting the particle's speed and position. The iteration steps are: fitness value, individual optimal position, global optimal position, and adjusting the particle's speed and position.

[0059] Step (1) specifically refers to: obtaining the geographical location information of the overall inspection area of ​​the clean energy station, performing data modeling on the overall inspection area, and using the modeling data to divide the overall inspection area into sub-regions to obtain the boundaries and main features of each inspection area. The main features include task attributes, spatial geometric attributes, and environmental constraint attributes. The task attributes include equipment type, load requirements, and defect density. The spatial geometric attributes include area, aspect ratio, and terrain undulation. The environmental constraint attributes include electromagnetic interference, wind speed, and take-off and landing conditions.

[0060] In step (2), the clean energy station multi-UAV task allocation model includes constraints and objective functions; the constraints include one task per UAV, single UAV energy consumption upper limit constraint, and task priority timing constraint.

[0061] The one-task-one-machine constraint is as follows:

[0062] Each task can only be assigned to one drone, each particle's dimension can only take one value, and each task can only be executed by one drone:

[0063] , ;

[0064] in, =1 indicates that task i is assigned to drone j; otherwise... =0; This refers to the total number of drones; It is the total number of tasks;

[0065] The upper limit constraint on the single-machine energy consumption is:

[0066] Each drone consumes a certain amount of energy while performing a mission, but the energy consumption shall not exceed the drone's maximum battery capacity.

[0067] , ;

[0068] in, It is the energy consumption of task i. This is the maximum battery capacity of drone j;

[0069] The task priority timing constraint is as follows:

[0070] If tasks have priority requirements, higher priority tasks should be completed first, and the completion time for higher priority tasks should be earlier than that for lower priority tasks.

[0071] if ;

[0072] in, It is the priority of task i. It is the priority of task i+1. This indicates that task i has a higher priority than task i+1; For the moment when task i is executed, This refers to the moment when task i+1 is executed;

[0073] The objective function design considers multiple factors, such as task completion time, energy consumption, and task priority. Task completion time is the maximum time to complete a task, and the objective is to minimize this time. The energy consumption of the UAV during task execution can be calculated through the relationship between task allocation and energy consumption. For tasks with priorities, higher-priority tasks need to be completed as early as possible. Priority weighting can be represented by the product of task completion time and task priority.

[0074] The formula for the objective function is:

[0075] minf = + + ;

[0076] in, , , These are all weighting coefficients for the target.

[0077] In step (2), the improvement of the inertia weight and learning factor of the particle swarm algorithm specifically refers to:

[0078] Improved inertia weight for:

[0079] = ;

[0080] in, This indicates the maximum number of iterations. Indicates the current iteration number. This represents the maximum value of the inertia weight. This represents the difference between the maximum and minimum values ​​of the inertia weights;

[0081] The improvement of the learning factor specifically refers to:

[0082] Improvement of the Individual Learning Factor: Since the global optimal solution is not yet stable and may be far from the final expected solution, the individual learning ability of particles becomes particularly important. Therefore, the individual learning factor is usually set larger, while the social learning factor is set smaller, to encourage particles to rely more on their own experience, thereby exploring the solution space more effectively. Improved Individual Learning Factor for:

[0083] = - ;

[0084] Improvement of the social learning factor: As iterations proceed, the performance of the current optimal solution gradually improves, and particles begin to reduce their reliance on their own learning, instead drawing more attention to the direction of the current optimal solution. The value gradually decreases, while Increasing the value of the social learning factor allows particles to better utilize global information, accelerating convergence to a better solution. (Improved social learning factor) for:

[0085] = + ;

[0086] In the formula, This represents the maximum degree of trust a particle has in its own experience. This represents the minimum degree to which a particle adopts the collective experience; The effective interval for the learning factor.

[0087] Step (3) specifically refers to:

[0088] The velocity formula is as follows:

[0089] = + + ;

[0090] in, and All are random numbers in the range [0,1]. and These are the individual optimal position and the global optimal position, respectively. It is the velocity of the particle after the (t+1)th iteration. It is the velocity of the particle after the t-th iteration. It is the position of the particle after the t-th iteration;

[0091] Integrating the golden sine wave algorithm into the particle swarm optimization algorithm significantly enhances its search capability and accelerates convergence. The improved position update formula is obtained as follows:

[0092] = - ;

[0093] = a +b ;

[0094] = a +b ;

[0095] in, It is the step size direction factor, with a value range of (0, 2). ); Let be the position disturbance factor, in (0, Take the value within ) and These are the coefficients derived from the golden ratio. It is the golden ratio, with a value of a and b define the scope of the golden ratio search, where a is the dynamic lower bound factor and b is the static upper bound factor. It is the position of the particle after the (t+1)th iteration.

[0096] Step (4) specifically refers to: when the number of iterations reaches the upper limit, returning the task allocation solution corresponding to the globally optimal particle; otherwise, continuing to update the particle state and optimize the solution; and determining whether the current number of iterations has reached the set maximum number of iterations. If the maximum number of iterations has not been reached, the current iteration count is incremented by 1, the learning factor and inertia weight are recalculated, and the particle's velocity and position are updated.

[0097] In summary, this invention employs a nonlinear dynamic collaborative improvement of inertia weights and learning factors. The inertia weights utilize a nonlinear adaptive decreasing strategy, maintaining a large value in the early stages of iteration to enhance the global traversal capability of UAV formations and broadly search various sub-regions and task combinations. In the later stages of iteration, the weights rapidly decrease to strengthen local refinement and ensure rapid convergence to the optimal cost combination. The position update mechanism is reconstructed by integrating the golden sine algorithm, achieving a dynamic balance between global and local searches.

[0098] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for multi-UAV task allocation at clean energy power stations based on particle swarm optimization, characterized in that: The method includes the following steps in sequence: (1) Obtain regional data information of clean energy power stations and divide the inspection areas; (2) Construct a multi-UAV task allocation model for clean energy stations based on the divided inspection areas; integrate the golden sine algorithm into the particle swarm algorithm, and improve the inertial weight and learning factor of the particle swarm algorithm; (3) Calculate the fitness value based on the objective function, update the individual optimal position of the particle, determine the global optimal position, and adjust the particle's velocity and position based on the individual optimal position and the global optimal position to optimize the objective function; (4) Set the total number of iterations. Each iteration will update the particle's fitness value, individual optimal position, global optimal position, and adjust the particle's speed and position. When the number of iterations reaches the upper limit, return the task allocation solution corresponding to the global optimal particle. Otherwise, continue to update the particle state and optimize the solution.

2. The method for multi-UAV task allocation at clean energy power stations based on particle swarm optimization as described in claim 1, characterized in that: Step (1) specifically refers to: obtaining the geographical location information of the overall inspection area of ​​the clean energy station, performing data modeling on the overall inspection area, and using the modeling data to divide the overall inspection area into sub-regions to obtain the boundaries and main features of each inspection area. The main features include task attributes, spatial geometric attributes, and environmental constraint attributes. The task attributes include equipment type, load requirements, and defect density. The spatial geometric attributes include area, aspect ratio, and terrain undulation. The environmental constraint attributes include electromagnetic interference, wind speed, and take-off and landing conditions.

3. The method for multi-UAV task allocation at clean energy power stations based on particle swarm optimization as described in claim 1, characterized in that: In step (2), the clean energy station multi-UAV task allocation model includes constraints and objective functions; the constraints include one task per UAV, single UAV energy consumption upper limit constraint, and task priority timing constraint. The one-task-one-machine constraint is as follows: Each task can only be assigned to one drone, each particle's dimension can only take one value, and each task can only be executed by one drone: , ； in, =1 indicates that task i is assigned to drone j; otherwise... =0; This refers to the total number of drones; It is the total number of tasks; The upper limit constraint on the single-machine energy consumption is: Each drone consumes a certain amount of energy while performing a mission, but the energy consumption shall not exceed the drone's maximum battery capacity. ， ； in, It is the energy consumption of task i. This is the maximum battery capacity of drone j; The task priority timing constraint is as follows: High-priority tasks should be completed earlier than low-priority tasks. ,if ; in, It is the priority of task i. It is the priority of task i+1. This indicates that task i has a higher priority than task i+1; For the moment when task i is executed, This refers to the moment when task i+1 is executed; The formula for the objective function is: minf = + + ; in, , , These are all weighting coefficients for the target.

4. The method for multi-UAV task allocation at clean energy power stations based on particle swarm optimization as described in claim 1, characterized in that: In step (2), the improvement of the inertia weight and learning factor of the particle swarm algorithm specifically refers to: Improved inertia weight for: = ； in, This indicates the maximum number of iterations. Indicates the current iteration number. This represents the maximum value of the inertia weight. This represents the difference between the maximum and minimum values ​​of the inertia weights; The improvement of the learning factor specifically refers to: Improved individual learning factor for: = - ； Improved social learning factor for: = + ； In the formula, This represents the maximum degree of trust a particle has in its own experience. This represents the minimum degree to which a particle adopts the collective experience; The effective interval for the learning factor.

5. The method for multi-UAV task allocation at clean energy power stations based on particle swarm optimization as described in claim 1, characterized in that: Step (3) specifically refers to: The velocity formula is as follows: = + + ； in, and All are random numbers in the range [0,1]. and These are the individual optimal position and the global optimal position, respectively. It is the velocity of the particle after the (t+1)th iteration. It is the velocity of the particle after the t-th iteration. It is the position of the particle after the t-th iteration; By incorporating the golden sine algorithm into the particle swarm optimization algorithm, an improved position update formula is obtained: = - ； = a +b ; = a +b ; in, It is the step size direction factor, with a value range of (0, 2). ); Let be the position disturbance factor, in (0, Take the value within ) and These are the coefficients derived from the golden ratio. It is the golden ratio, with a value of a and b define the scope of the golden ratio search, where a is the dynamic lower bound factor and b is the static upper bound factor. It is the position of the particle after the (t+1)th iteration.

6. The method for multi-UAV task allocation at clean energy power stations based on particle swarm optimization as described in claim 1, characterized in that: Step (4) specifically refers to: when the number of iterations reaches the upper limit, returning the task allocation solution corresponding to the globally optimal particle; otherwise, continuing to update the particle state and optimize the solution; and determining whether the current number of iterations has reached the set maximum number of iterations. If the maximum number of iterations has not been reached, the current iteration count is incremented by 1, the learning factor and inertia weight are recalculated, and the particle's velocity and position are updated.

7. An electronic device, comprising: processor; as well as A memory storing computer program instructions that, when executed by the processor, cause the processor to perform the particle swarm optimization-based multi-UAV task allocation method for clean energy stations as described in any one of claims 1-6.

8. A computer-readable storage medium having stored thereon computer program instructions, which, when executed by a processor, cause the processor to perform the particle swarm optimization method for multi-UAV task allocation at clean energy power stations as described in any one of claims 1-6.

Images