Multi-unmanned aerial vehicle task allocation method based on random differentiation creative search algorithm

Abstract

The invention discloses a multi-unmanned-aerial-vehicle task allocation method based on a random differentiation creative search algorithm, and the method comprises the steps: carrying out the modeling of a task allocation scene according to the actual environment condition of agriculture, determining a constraint condition and an objective function, and building a multi-unmanned-aerial-vehicle task allocation model for farmland plant protection; and then completing population initialization based on a random walk strategy, then establishing an iteration model based on an optimization team performance model, carrying out loop iteration on the optimization process until an iteration ending condition is met, solving to obtain the optimal allocation of the multiple unmanned aerial vehicles to the task sub-regions to be operated, and outputting the allocation scheme. And multi-unmanned aerial vehicle farmland plant protection task allocation is realized. According to the method, optimization is carried out in an algorithm initialization stage by using a random walk strategy, the global search capability is enhanced, an optimal task allocation scheme is rapidly obtained through iterative solution, the method has obvious advantages in global search and solution speed, and an innovative solution is provided for efficient allocation of multiple unmanned aerial vehicles in complex plant protection tasks.

Description

Technical Field

[0001] The present invention belongs to the technical field of multi-UAV task allocation, and in particular relates to a multi-UAV task allocation method based on a random differentiated creative search algorithm. Background Art

[0002] Drones have the characteristics of strong autonomy, high flexibility, low cost and good portability. They can autonomously navigate and avoid obstacles, and can be quickly deployed to perform diverse tasks, such as agricultural spraying, disaster relief and environmental monitoring. Compared with a single drone, a multi-drone system has the advantages of collaborative operation, efficient coverage, flexible response and enhanced reliability. By working together, multiple drones can quickly cover a large area and improve the efficiency of data collection and task execution. In the field of multi-drone task allocation, the application of optimization algorithms has become a hot topic of research. These algorithms aim to determine the best solution for task allocation of each drone in a multi-drone system through precise calculations to achieve specific goals, such as minimizing the total flight time, maximizing the efficiency of task completion, and balancing the load of each drone. Classical optimization algorithms, such as genetic algorithms, particle swarm optimization algorithms and ant colony algorithms, can effectively solve complex problems and find optimal or near-optimal solutions with their excellent global search capabilities.

[0003] However, although these optimization algorithms have shown significant advantages in multi-UAV task allocation, they still face a series of challenges. First, the computational complexity of the algorithm grows exponentially with the increase in the number of UAVs and tasks, which places higher demands on computing power. Secondly, multi-UAV task allocation needs to consider a variety of constraints, such as flight time, energy consumption, capability size and task scenario, etc. These constraints are intertwined, increasing the complexity of task allocation. In addition, the optimization algorithm is prone to fall into a local optimal solution rather than a global optimal solution during the search process, which may lead to a less than ideal task allocation scheme, reduce system efficiency and effect, especially in complex and multi-peak search spaces. This problem is particularly prominent. In addition, most of the current multi-UAV task allocation methods are mainly solved in combat mission planning and target search, and are less used in plant protection, and do not take into account the actual environmental factors of plant protection and constraints such as UAV capabilities. Therefore, according to the constraints of task scenarios and UAV capabilities, the present invention proposes a differential creative search optimization algorithm based on a random walk strategy to solve a better task allocation scheme, reduce resource waste, and improve task efficiency. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a multi-UAV task allocation method based on a random differentiated creative search algorithm, which draws on the foraging strategies of animals in nature, achieves a balance between global and local search through a combination of long jumps and short steps, and reduces the risk of falling into local optimality.

[0005] The technical solution adopted by the present invention is: a multi-UAV task allocation method based on a random differentiation creative search algorithm, and the specific steps are as follows:

[0006] S1. Model the task allocation scenario according to the actual agricultural environment, establish the constraints and objective functions, and build a multi-UAV task allocation model for farmland plant protection;

[0007] S2. Initialize the population based on the random walk strategy, i.e. set the initialization parameters of the differentiated creative search algorithm;

[0008] The initialization parameters include: mission area, UAV capability, amount of medicine that the UAV can carry, amount of medicine required for each unit of light-grade pest and disease area, degree of pest and disease, population size, maximum number of iterations, number of high performers, and probability constant.

[0009] Here, the population size is the number of candidate solutions.

[0010] S3, establishing an iterative model based on the optimized team performance model, that is, simulating the performance of three different members in the team to establish an optimized iterative method, and iterating the optimization process in a loop until the iteration end condition is met, and then going to step S4;

[0011] The three different members in the simulated team include: high performers, average performers, and low performers.

[0012] Among them, high performers are used to create different solutions and explore different areas to prevent premature identification of suboptimal solutions; average performers are used to refine solutions in promising areas; low performers are used to increase team diversity, promote exploration, and increase the pool of ideas.

[0013] The optimization process is iterated cyclically based on the iterative model, and firstly, initial parameters of the optimization problem and iteration termination conditions are set.

[0014] Among them, the initial parameters of the optimization problem include: focusing on the speed of new knowledge acquisition qKR, social impact factor, and differentiated knowledge acquisition process; the iteration end condition is to meet the maximum number of iterations.

[0015] After one iteration, the local optimal solution, i.e., the individual position, is updated, and it is determined whether the maximum number of iterations is met. If not, step S3 is repeated for loop iteration, and the global optimal solution, i.e., the optimal position, obtained in each iteration in the solution space is recorded after the loop starts, until the iteration end condition is met, and then go to step S4.

[0016] S4. Solve to obtain the optimal allocation of multiple UAVs to the sub-areas of the task, output this allocation plan, and realize the allocation of multiple UAVs for farmland plant protection tasks.

[0017] Furthermore, the step S1 is specifically as follows:

[0018] S11. Model the task allocation scenario according to the actual agricultural environment;

[0019] Assume there are m waiting task sub-areas A={A1,...,A m}, n drones set U = {U1,...,U n}, each sub-region has different sizes, using S = {S1,...,S m} represents the size of each area. The severity of pests and diseases in each area varies. L = {1, 1.5, 2} represents the severity level of the area, 1 for mild severity, 1.5 for moderate severity, and 2 for severe severity;

[0020] Define x ij For UAV i Assigned to the pending tasks sub-area A j The plant protection tasks performed, where i∈n, j∈m, and each UAV U i Can be assigned to only one task sub-area;

[0021] Define y j The number of drones assigned to each task sub-area to be operated, where j∈m, and the sum of the number of drones assigned to each task sub-area to be operated is equal to the total number of drones planned;

[0022] Using T j Indicates the sub-area A of the tasks to be done j Total operation time;

[0023] Define Q j is the drug demand of the jth task sub-area to be operated, q i is the drug carrying capacity of the i-th drone. The number of drones allocated to each task sub-area to be operated must meet the drug requirement of the area.

[0024] Define k to represent the amount of pesticide required for each unit of minor pest and disease sub-area to be operated. Combined with the plot size and the severity of the pest and disease, calculate the comprehensive pesticide demand for each sub-area to be operated.

[0025] Definition C i is the unit area processing capacity of the i-th UAV in the task sub-area with minor pests and diseases. The processing capacity of the UAV for areas with different pests and diseases levels is different.

[0026] S12. According to the principle of minimizing the total operation time, establish the objective function and its constraints, and establish a multi-UAV task allocation model for farmland plant protection;

[0027]

[0028] Among them, Z represents the total time for all drones to complete the plant protection operation, L j Indicates the sub-area A of the pending tasks j Pest and disease levels.

[0029] Furthermore, the step S2 is specifically as follows:

[0030] S21. Define the size of the initial task sub-area S, the severity level of pests and diseases L, the unmanned processing capacity C and the amount of medicine carried q, the amount of medicine k required for each unit of light-level pest and disease area, the population size NP, and the maximum number of population iterations NFE max ;

[0031] S22, initializing population position;

[0032] In the differential creative search optimization algorithm, the optimization process starts with a set of candidate solutions randomly generated between the upper and lower bounds of the optimization problem.

[0033] x i,d =LB d +U(0,1)×(UB d -LB d )

[0034] Among them, x i,d Represents x i The element at the dth position, x i Indicates drone U d The task sub-region assigned to the dth position. LB and UB represent the lower and upper bounds of the solution space, U(0,1) represents the uniform distribution on the interval (0,1), and LB d and UB d They represent the lower and upper bounds of the dth dimension specified by the optimization problem. The initial population X is expressed in matrix form as follows:

[0035]

[0036] Here, NP represents the number of candidate solutions, and D represents the dimension of the optimization problem.

[0037] S23, reinitialize the population position using the random walk strategy;

[0038] For the population X initialized in step S22, the random walk strategy is used to reinitialize the population, and the expression is as follows:

[0039]

[0040] Among them, Xt represents the current position, α represents the step size scaling factor, represents the Hadamard product, β represents the stability parameter of the Levy distribution, and Levy(β) represents the generated Levy step size. u~N(0,σ 2 ) indicates that the random variable u has a mean of 0 and a variance of σ 2 v~N(0,1) means that the random variable v follows a standard normal distribution with mean 0 and variance 1. σ represents a scaling factor related to the parameters of the Levy distribution. Γ represents the gamma function.

[0041] S24. Obtain the number of high performers, which is 20% of the total population size and is at least 6, and then set the probability constant Pc based on the actual situation.

[0042] Furthermore, the step S3 is specifically as follows:

[0043] S31, start iteration, update the initial parameters required for the individuals in the iterative population, and calculate the social impact factor;

[0044] The initial parameters required for updating the iterative population individuals are to set the initial parameters of the optimization problem, including: paying attention to the speed of new knowledge acquisition qKR, social influence factors, and differentiated knowledge acquisition process.

[0045] Among them, in a team environment, the social cognition of an individual will be influenced by his peers, reflecting the degree to which the social dynamics of the team influences his personal views, that is, the social influence factor λ t , the calculation expression is as follows:

[0046]

[0047] Among them, λ t A high coefficient value indicates that the individual is more susceptible to peer pressure and social influence, while a low coefficient value indicates that the individual is more independent; NFE t Indicates the number of iterations of the current function, NFE max Indicates the maximum number of function iterations.

[0048] Then, every individual in the population is traversed. Each individual conducts divergent or convergent thinking based on its own identity, and increases team diversity. Before conducting divergent or convergent thinking and increasing team diversity, each individual first acquires differentiated knowledge.

[0049] Among them, differentiated knowledge acquisition DKA is aimed at the different potentials of human beings to acquire knowledge, and focuses on the speed of new knowledge acquisition qKR. i,t _qkr represents the quantitative knowledge acquisition rate of the individual at the tth iteration, and the expression is as follows:

[0050]

[0051] in, represents the individual at the tth iteration The coefficient value is calculated as follows:

[0052]

[0053] Among them, R i,t represents the rank of the i-th individual at the beginning of the t-th iteration.

[0054] The effect of the process of differentiated knowledge acquisition on each individual is expressed as follows:

[0055] j rand =randint(1,D)

[0056]

[0057] Among them, v i,d Represents the test element v i,t The element at position d, v i,t represents the trial element v at the tth iteration i . η i,t _qkr represents the quantitative knowledge acquisition rate of the individual at the tth iteration. rand represents a randomly selected integer from 1 to D.

[0058] S32, judging whether the population individual is a high performer, if so, acquiring differentiated knowledge, and then executing divergent thinking behavior, after which step S35 is entered, if not, step S33 is entered;

[0059] When the individual is a high performer, he first acquires differentiated knowledge and then performs divergent thinking, that is, injecting the concept of divergent thinking into the experimental vector v i,d In the generation of , the divergent thinking strategy expression is as follows:

[0060] v i,d =x r1,d +Lk(α,σ)

[0061] Among them, x r1,d represents the dth position of an individual randomly selected from {1,2,...,NP}, and x r1 ≠x i ≠x best , x i represents the current individual, x bestrepresents the best individual at present. Lk(α,σ) represents an additional term from the Linnik distribution, which is used to encapsulate imagination. α and σ represent the control parameters of the Linnik distribution, and α=0.618, σ=0.05.

[0062] S33, determine whether the individual in the population is a low performer, if so, acquire differentiated knowledge, and then perform the operation of increasing team diversity and enter step S35, otherwise the individual is a general performer, and enter step S34;

[0063] When this iteration traverses to the last individual in the population, if the generated random number is less than the probability constant Pc, the individual in the population is a low performer. First, differentiated knowledge is acquired and then the operation of increasing team diversity is performed. The expression for generating new members is as follows:

[0064] v i,d =LB+U(0,1)×(UB-LB)

[0065] Among them, U(0,1) represents the uniform distribution on the interval (0,1), LB and UB represent the lower bound and upper bound specified by the optimization problem.

[0066] S34, if the individual is a general performer, first acquire differentiated knowledge, then perform the operation of convergent thinking and enter step S35;

[0067] When the execution conditions of high performers and low performers are not met, they are general performers. They first acquire differentiated knowledge and then perform thinking convergence behavior. The thinking convergence strategy expression is as follows:

[0068] v i,d =w×x best,d +λ t ×(x r2,d -x i,d )+ω i,t ×(x r1,d -x i,d )

[0069] Among them, x best,d represents the dth position of the best performing individual in the current iteration. w represents the cognitive weight of the best performing individual and its value is 1. r1,d represents the dth position of an individual randomly selected from {1,2,...,NP}, and x r1 ≠x i ≠x best , they represent three different individuals, x i Represents the current individual. r2,d represents the dth position of a random individual selected from {ngs+1,...,NP}, and x r2 ≠xr1 ≠x i ≠x best , they represent four different individuals. The coefficient ω represents the learning intensity state of the individual. i,t It represents the efficiency value of the individual at the tth iteration, and the expression is as follows:

[0070] ω i,t ~U(0,1)

[0071] S35. Implement boundary handling mechanisms to ensure that the ideas generated remain within the scope of feasibility and reality;

[0072] If the idea proposed by an individual exceeds the actual limit of the dth position, the boundary processing mechanism will correct the value to the closest maximum or minimum value within the range allowed by the data set. The calculation formula is as follows:

[0073]

[0074] Among them, v i,d Represents the test element v i,t The element at position d, v i,t represents the trial element v at the tth iteration i LB d and UB d denote the lower and upper bounds of the d-th dimension specified by the optimization problem.

[0075] S36, update the individual position, and then determine whether it is the maximum number of iterations. If not, return to step S31 to perform the next iteration until the maximum number of iterations is met, end the iteration loop, and enter step S4.

[0076] Beneficial effects of the present invention: The method of the present invention first models the task allocation scenario according to the actual agricultural environment, establishes the constraints and their objective functions, establishes a multi-UAV task allocation model for farmland plant protection, and then completes the population initialization based on the random walk strategy, and then establishes an iterative model based on the optimization team performance model, and iterates the optimization process until the iteration end condition is met, and the optimal allocation of the multi-UAV to the task sub-area is solved, and the allocation scheme is output to realize the multi-UAV farmland plant protection task allocation. The method of the present invention uses the random walk strategy to optimize in the algorithm initialization stage, enhances the global search capability, and quickly obtains the optimal task allocation scheme through iterative solution, which has obvious advantages in global search and solution speed, and provides an innovative solution for the efficient allocation of multiple UAVs in complex plant protection tasks. The method of the present invention is specifically aimed at the field of agricultural plant protection. By constructing a task allocation model that comprehensively considers the level of pests and diseases, the size of the plot and the performance of the UAV, the resource allocation of the UAV is optimized, the efficiency of the plant protection operation is improved and the resource consumption is reduced. The improved difference creative search optimization algorithm based on the random walk strategy proposed by the method of the present invention draws on the foraging strategy of animals in nature, achieves a balance between global and local searches through the combination of long jumps and short steps, and reduces the risk of falling into local optimality. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 The present invention is a flowchart of a multi-UAV task allocation method based on a random differentiated creative search algorithm.

[0078] Figure 2 It is an iterative flow chart of the iterative model in an embodiment of the present invention.

[0079] Figure 3 It is a convergence curve diagram of the random differentiated creative search algorithm of the present invention in an embodiment of the present invention.

[0080] Figure 4 It is a comparative convergence curve diagram of the differentiated creative search algorithm and the random differentiated creative search algorithm in the embodiment of the present invention. DETAILED DESCRIPTION

[0081] The method of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0082] like Figure 1 As shown, a flow chart of a multi-UAV task allocation method based on a randomized differentiated creative search algorithm (DCS-levy) of the present invention, the specific steps are as follows:

[0083] S1. Model the task allocation scenario according to the actual agricultural environment, establish the constraints and objective functions, and build a multi-UAV task allocation model for farmland plant protection;

[0084] S2. Initialize the population based on the random walk strategy, i.e. set the initialization parameters of the differentiated creative search algorithm (DCS);

[0085] The initialization parameters include: mission area, UAV capability, amount of medicine that the UAV can carry, amount of medicine required for each unit of light-grade pest and disease area, degree of pest and disease, population size, maximum number of iterations, number of high performers (ngs), and probability constant.

[0086] Here, the population size is the number of candidate solutions.

[0087] S3, establishing an iterative model based on the optimized team performance model, that is, simulating the performance of three different members in the team to establish an optimized iterative method, and iterating the optimization process in a loop until the iteration end condition is met, and then going to step S4;

[0088] The three different members in the simulated team include: high performers, average performers, and low performers.

[0089] Among them, high performers are used to create different solutions and explore different areas to prevent premature identification of suboptimal solutions; average performers are used to refine solutions in promising areas; low performers are used to increase team diversity, promote exploration, and increase the pool of ideas.

[0090] The optimization process is iterated cyclically based on the iterative model, and firstly, initial parameters of the optimization problem and iteration termination conditions are set.

[0091] Among them, the initial parameters of the optimization problem include: focusing on the speed of new knowledge acquisition (qKR), social impact factors, and differentiated knowledge acquisition process; the iteration end condition is to meet the maximum number of iterations.

[0092] After one iteration, the local optimal solution, i.e., the individual position, is updated, and it is determined whether the maximum number of iterations is met. If not, step S3 is repeated for loop iteration, and the global optimal solution, i.e., the optimal position, obtained in each iteration in the solution space is recorded after the loop starts, until the iteration end condition is met, and then go to step S4.

[0093] S4. Solve to obtain the optimal allocation of multiple UAVs to the sub-areas of the task, output this allocation plan, and realize the allocation of multiple UAVs for farmland plant protection tasks.

[0094] In this embodiment, the step S1 is specifically as follows:

[0095] S11. Model the task allocation scenario according to the actual agricultural environment;

[0096] In this embodiment, m=4 waiting task sub-areas A={A1,...,A4}, n=10 drone sets U={U1,...,U 10}, each sub-region has different sizes, S = {100, 200, 150, 250} represents the size of each region, and the severity of pests and diseases in each region is different, L = {1, 1.5, 1.5, 2} represents the severity level of the region, 1 is a mild degree, 1.5 is a medium degree, and 2 is a severe degree;

[0097] Define x ij For UAV i Assigned to the pending tasks sub-area A j The plant protection tasks performed, where i∈n, j∈m, and each UAV U i Can be assigned to only one task sub-area;

[0098] Define y j The number of drones assigned to each task sub-area to be operated, where j∈m, and the sum of the number of drones assigned to each task sub-area to be operated is equal to the total number of drones planned;

[0099] Using T j Indicates the sub-area A of the tasks to be done j Total operation time;

[0100] Define Q j is the drug demand of the jth task sub-area to be operated, q i is the drug carrying capacity of the i-th drone. The number of drones allocated to each task sub-area to be operated must meet the drug requirement of the area.

[0101] Define k to represent the amount of pesticide required for each unit of minor pest and disease sub-area to be operated. Combined with the plot size and the severity of the pest and disease, calculate the comprehensive pesticide demand for each sub-area to be operated.

[0102] Definition C i is the unit area processing capacity of the i-th UAV in the task sub-area with minor pests and diseases. The processing capacity of the UAV for areas with different pests and diseases levels is different.

[0103] S12. According to the principle of minimizing the total operation time, establish the objective function and its constraints, and establish a multi-UAV task allocation model for farmland plant protection;

[0104]

[0105] Among them, Z represents the total time for all drones to complete the plant protection operation, L j Indicates the sub-area A of the tasks to be done jPest and disease levels.

[0106] In this embodiment, step S2 is specifically as follows:

[0107] S21. Define the size S of the initial task sub-area to be operated, the severity level of pests and diseases L, the unmanned processing capacity C = {5, 6, 5.5, 6.2, 6, 5.8, 6.1, 5.9, 6, 5.7} and the amount of medicine carried q = {10, 12, 15, 10, 14, 16, 13, 11, 14, 15}, the amount of medicine required for each unit of light pest and disease area k = 0.1, the population size NP = 30, and the maximum number of population iterations NFE max =500;

[0108] S22, initializing population position;

[0109] In the differential creative search optimization algorithm, the optimization process starts with a set of candidate solutions randomly generated between the upper and lower bounds of the optimization problem.

[0110] x i,d =LB d +U(0,1)×(UB d -LB d )

[0111] Among them, x i,d Represents x i The element at the dth position (dimension), x i Indicates drone U d The task sub-region assigned to the dth position. LB and UB represent the lower and upper bounds of the solution space, U(0,1) represents the uniform distribution on the interval (0,1), and LB d and UB d They represent the lower and upper bounds of the dth dimension specified by the optimization problem. The initial population X is expressed in matrix form as follows:

[0112]

[0113] Here, NP represents the number of candidate solutions (called the population size), and D represents the dimension of the optimization problem.

[0114] S23, reinitialize the population position using the random walk strategy;

[0115] For the population X initialized in step S22, the random walk strategy is used to reinitialize the population, and the expression is as follows:

[0116]

[0117] Among them, X t represents the current position, α represents the step scaling factor, represents the Hadamard product (i.e., the corresponding elements are multiplied), β represents the stability parameter of the Levy distribution, which controls the distribution characteristics of the step length, and Levy(β) represents the generated Levy step length. u~N(0,σ 2 ) indicates that the random variable u has a mean of 0 and a variance of σ 2 Normal distribution. v~N(0,1) means that the random variable v follows a standard normal distribution with mean 0 and variance 1. σ represents a scaling factor related to the parameters of the Levy distribution. Γ represents the gamma function, which extends the factorial function to the continuous domain.

[0118] S24. Obtain the number of high performers, which is 20% of the total population size and is at least 6, and then set the probability constant Pc=0.5 according to the actual situation.

[0119] like Figure 2 As shown, in this embodiment, step S3 is specifically as follows:

[0120] S31, start iteration, update the initial parameters required for the individuals in the iterative population, and calculate the social impact factor;

[0121] The initial parameters required for updating the iterative population individuals are to set the initial parameters of the optimization problem, including: focusing on the speed of new knowledge acquisition (qKR), social influence factors, and differentiated knowledge acquisition process.

[0122] Among them, in a team environment, the social cognition of an individual will be influenced by his peers, reflecting the degree to which the social dynamics of the team influences his personal views, that is, the social influence factor λ t , the calculation expression is as follows:

[0123]

[0124] Among them, λ t A high coefficient value indicates that the individual is more susceptible to peer pressure and social influence, while a low coefficient value indicates that the individual is more independent; NFE t Indicates the number of iterations of the current function, NFE max Indicates the maximum number of function iterations.

[0125] Then, every individual in the population is traversed. Each individual conducts divergent or convergent thinking based on its own identity, and increases team diversity. Before conducting divergent or convergent thinking and increasing team diversity, each individual first acquires differentiated knowledge.

[0126] Among them, differentiated knowledge acquisition (DKA) is aimed at the different potentials of human beings to acquire knowledge, and focuses on the speed of new knowledge acquisition (qKR). i,t _qkr represents the quantitative knowledge acquisition rate of the individual at the tth iteration, and the expression is as follows:

[0127]

[0128] in, represents the individual at the tth iteration The coefficient value is calculated as follows:

[0129]

[0130] Among them, R i,t represents the rank (order) of the i-th individual at the beginning of the t-th iteration.

[0131] The effect of the process of differentiated knowledge acquisition on each individual is expressed as follows:

[0132] j rand =randint(1,D)

[0133]

[0134] Among them, v i,d Represents the test element v i,t The element at position d (dimension), v i,t represents the trial element v at the tth iteration i . η i,t _qkr represents the quantitative knowledge acquisition rate of the individual at the tth iteration. rand represents a randomly selected integer from 1 to D.

[0135] S32, judging whether the population individual is a high performer, if so, acquiring differentiated knowledge, and then executing divergent thinking behavior, after which step S35 is entered, if not, step S33 is entered;

[0136] High performers generate creative ideas. Divergent thinking, which involves generating diverse, creative ideas and solutions to problems, is a key characteristic of high performers and is at the heart of creativity.

[0137] When the individual is a high performer, he first acquires differentiated knowledge and then performs divergent thinking, that is, injecting the concept of divergent thinking into the experimental vector v i,d In the generation of , the divergent thinking strategy expression is as follows:

[0138] v i,d =x r1,d +Lk(α,σ)

[0139] Among them, x r1,d represents the dth position of an individual randomly selected from {1,2,...,NP}, and x r1 ≠x i ≠xbest , x i represents the current individual, x best represents the best individual at present. Lk(α,σ) represents an additional term from the Linnik distribution, which is used to encapsulate imagination. α and σ represent the control parameters of the Linnik distribution, and α=0.618, σ=0.05.

[0140] S33, determine whether the individual in the population is a low performer, if so, acquire differentiated knowledge, and then perform the operation of increasing team diversity and enter step S35, otherwise the individual is a general performer, and enter step S34;

[0141] Low performers improve team diversity. Teams with a changing membership generate more diverse ideas, replacing poor performers with new members.

[0142] When this iteration traverses to the last individual in the population, if the generated random number is less than the probability constant Pc, the individual in the population is a low performer. First, differentiated knowledge is acquired and then the operation of increasing team diversity is performed. The expression for generating new members is as follows:

[0143] v i,d =LB+U(0,1)×(UB-LB)

[0144] Among them, U(0,1) represents the uniform distribution on the interval (0,1), LB and UB represent the lower bound and upper bound specified by the optimization problem.

[0145] S34, if the individual is a general performer, first acquire differentiated knowledge, then perform the operation of convergent thinking and enter step S35;

[0146] Average performers engage in convergent thinking behaviors, refining ideas into solutions.

[0147] When the execution conditions of high performers and low performers are not met, they are general performers, who first acquire differentiated knowledge and then perform convergent thinking behavior. It relies on the knowledge base of the best performer and combines the random contributions from two different team members brought by the current individual. Therefore, it promotes the diversity and convergence of the generated solutions. The expression of the convergent thinking strategy is as follows:

[0148] v i,d =w×x best,d +λ t ×(x r2,d -x i,d )+ω i,t ×(x r1,d -x i,d )

[0149] Among them, xbest,d represents the dth position of the best performing individual in the current iteration. w represents the cognitive weight of the best performing individual and its value is 1. r1,d represents the dth position of an individual randomly selected from {1,2,...,NP}, and x r1 ≠x i ≠x best , they represent three different individuals, x i Represents the current individual. r2,d represents the dth position of a random individual selected from {ngs+1,...,NP}, and x r2 ≠x r1 ≠x i ≠x best , which represent four different individuals.

[0150] The coefficient ω represents the individual's learning intensity state.

[0151] It takes into account an individual's potential to acquire knowledge, which may be influenced by factors such as mood, interest level and environment. This coefficient is like a fascinating mimic, reflecting the inherent complexity of interpersonal communication and collaboration. It encapsulates the nature of the mysterious human mind, where emotions and thoughts change like a kaleidoscope.

[0152] ω i,t It represents the efficiency value of the individual at the tth iteration, and the expression is as follows:

[0153] ω i,t ~U(0,1)

[0154] S35. Implement boundary handling mechanisms to ensure that the ideas generated remain within the scope of feasibility and reality;

[0155] If the idea proposed by an individual exceeds the actual limit of the dth position, the boundary processing mechanism will correct the value to the closest maximum or minimum value within the range allowed by the data set. The calculation formula is as follows:

[0156]

[0157] Among them, v i,d Represents the test element v i,t The element at position d (dimension), v i,t represents the trial element v at the tth iteration i LB d and UB d denote the lower and upper bounds of the d-th dimension specified by the optimization problem.

[0158] S36, update the individual position, and then determine whether it is the maximum number of iterations. If not, return to step S31 to perform the next iteration until the maximum number of iterations is met, end the iteration loop, and enter step S4.

[0159] In this embodiment, if the constraint condition established in step S12 is not satisfied, the fitness value is and The value is 6. The value is determined according to the actual situation.

[0160] In this embodiment, the total number of iterations of the optimization model is NFE max =500, the initial population size NP = 30, and the final task allocation scheme is as follows:

[0161] According to the task scenario and parameter settings of this embodiment, the calculated optimal fitness value is 3.913043, and the multi-UAV task allocation solution Best_solusion expression is as follows:

[0162] Best_solusion=[4 1 4 4 4 2 3 3 4 2]

[0163] The optimal assignment matrix assignment_matrix is ​​(sub-area × number of drones), expressed as follows:

[0164]

[0165] The fitness value convergence curve of the improved algorithm of the method of the present invention is calculated as follows: Figure 3 As shown in Figure 2, the convergence curve of the algorithm before improvement (DCS) is compared with the convergence curve of the algorithm after improvement (DCS-levy). Figure 4 As shown, it can be seen that the fitness value is lower after the improvement, and it is not easy to fall into the local optimal solution compared to before the improvement, and the solution speed is faster. The simulation results confirm that the method of the present invention can achieve the optimal allocation of multi-UAV tasks.

[0166] In summary, the method of the present invention provides an innovative solution for the efficient and energy-saving allocation of multiple UAVs in agricultural plant protection tasks, which has significant practical application value and theoretical research significance. Compared with the existing task allocation method, the method of the present invention adopts a differential creative search optimization algorithm, which is a more novel and excellent optimization strategy compared with the existing particle swarm algorithm and genetic algorithm. Applying it to the field of agricultural plant protection not only broadens its application scope, but also brings innovative solutions to this field; the method of the present invention is designed based on an improved strategy of random walk. This strategy quickly explores a broad search space in the early stage of the algorithm through the long jump characteristics, effectively improves the global search ability, and reduces the possibility of falling into the local optimal solution. Combining the long jump and short step movement strategies, the algorithm quickly conducts extensive exploration in the early stage and conducts fine search in the later stage, thereby accelerating the overall convergence speed; the method of the present invention is more efficient in solving complex multi-UAV task allocation problems, can adapt to a variety of application scenarios, and has the potential for promotion and application in other fields.

[0167] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims

1. A multi-UAV task allocation method based on randomized differentiated creative search algorithm, the specific steps are as follows: S1. Model the task allocation scenario according to the actual agricultural environment, establish the constraints and objective functions, and build a multi-UAV task allocation model for farmland plant protection; S2. Initialize the population based on the random walk strategy, i.e. set the initialization parameters of the differentiated creative search algorithm; The initialization parameters include: Mission area, drone capability, amount of medicine the drone can carry, amount of medicine required per unit area with minor pests and diseases, degree of pests and diseases, population size, maximum number of iterations, number of high performers, and probability constant; Among them, the population size is the number of candidate solutions; S3, establishing an iterative model based on the optimized team performance model, that is, simulating the performance of three different members in the team to establish an optimized iterative method, and iterating the optimization process in a loop until the iteration end condition is met, and then going to step S4; The three different types of members in the simulated team include: high performers, average performers, and low performers; Among them, high performers are used to create different solutions and explore different areas to prevent premature identification of suboptimal solutions; average performers are used to refine solutions in promising areas; low performers are used to increase team diversity, promote exploration, and increase the idea pool; The optimization process is iterated cyclically based on the iterative model, firstly, the initial parameters of the optimization problem and the iteration end conditions are set; Among them, the initial parameters of the optimization problem include: focusing on the speed of new knowledge acquisition qKR, social impact factor, and differentiated knowledge acquisition process; the iteration end condition is to meet the maximum number of iterations; After one iteration, the local optimal solution, i.e., the individual position, is updated, and it is determined whether the maximum number of iterations is met. If not, step S3 is repeated for loop iteration, and the global optimal solution, i.e., the optimal position, obtained in each iteration in the solution space is recorded after the loop starts, until the iteration end condition is met, and then step S4 is turned to; S4. Solve to obtain the optimal allocation of multiple UAVs to the sub-areas of the task, output this allocation plan, and realize the allocation of multiple UAVs for farmland plant protection tasks.

2. The multi-UAV task allocation method based on random differential creative search algorithm according to claim 1 is characterized in that: The step S1 is specifically as follows: S11. Model the task allocation scenario according to the actual agricultural environment; Assume there are m waiting task sub-areas A={A1,...,A m }, n drones set U = {U1,...,U n }, each sub-region has different sizes, using S = {S1,...,S m } represents the size of each area. The severity of pests and diseases in each area varies. L = {1, 1.5, 2} represents the severity level of the area, 1 for mild severity, 1.5 for moderate severity, and 2 for severe severity; Define x ij For UAV i Assigned to the pending tasks sub-area A j The plant protection tasks performed, where i∈n, j∈m, and each UAV U i Can be assigned to only one task sub-area; Define y j The number of drones assigned to each task sub-area to be operated, where j∈m, and the sum of the number of drones assigned to each task sub-area to be operated is equal to the total number of drones planned; Using T j Indicates the sub-area A of the tasks to be done j Total operation time; Define Q j is the drug demand of the jth task sub-area to be operated, q i is the drug carrying capacity of the i-th drone. The number of drones assigned to each task sub-area to be operated must meet the drug requirement of the area; Define k to represent the amount of pesticide required for each unit of minor pest and disease sub-area to be operated. Combined with the plot size and the severity of the pest and disease, calculate the comprehensive pesticide demand for each sub-area to be operated. Definition C i is the unit area processing capacity of the i-th UAV in the task sub-area with minor pests and diseases. The processing capacity of the UAV for areas with different pests and diseases levels is different; S12. According to the principle of minimizing the total operation time, establish the objective function and its constraints, and establish a multi-UAV task allocation model for farmland plant protection; Among them, Z represents the total time for all drones to complete the plant protection operation, L j Indicates the sub-area A of the tasks to be done j Pest and disease levels.

3. The multi-UAV task allocation method based on random differentiation creative search algorithm according to claim 1 is characterized in that: The step S2 is specifically as follows: S21. Define the size of the initial task sub-area S, the severity level of pests and diseases L, the unmanned processing capacity C and the amount of medicine carried q, the amount of medicine k required for each unit of light-level pest and disease area, the population size NP, and the maximum number of population iterations NFE max ; S22, initializing population position; In the differential creative search optimization algorithm, the optimization process starts with a set of candidate solutions randomly generated between the upper and lower bounds of the optimization problem; x i,d =LB d +U(0,1)×(UB d -LB d ) Among them, x i,d Represents x i The element at the dth position, x i Indicates drone U d The task sub-region assigned to the d-th position; LB and UB represent the lower and upper bounds of the solution space, U(0,1) represents the uniform distribution on the interval (0,1), and LB d and UB d They represent the lower and upper bounds of the dth dimension specified by the optimization problem respectively; the initial population X is expressed in matrix form, as follows: Where NP represents the number of candidate solutions, and D represents the dimension of the optimization problem; S23, reinitialize the population position using the random walk strategy; For the population X initialized in step S22, the random walk strategy is used to reinitialize the population, and the expression is as follows: Among them, X t represents the current position, α represents the step scaling factor, represents the Hadamard product, β represents the stability parameter of the Levy distribution, Levy(β) represents the generated Levy step size; u~N(0,σ 2 ) indicates that the random variable u has a mean of 0 and a variance of σ 2 Normal distribution; v~N(0,1) means that the random variable v obeys the standard normal distribution with mean 0 and variance 1; σ represents the scaling factor related to the parameters of the Levy distribution; Γ represents the gamma function; S24. Obtain the number of high performers, which is 20% of the total population size and is at least 6, and then set the probability constant Pc based on the actual situation.

4. The multi-UAV task allocation method based on random differentiation creative search algorithm according to claim 1 is characterized in that: The step S3 is specifically as follows: S31, start iteration, update the initial parameters required for the individuals in the iterative population, and calculate the social impact factor; The initial parameters required for updating the iterative population individuals are to set the initial parameters of the optimization problem, including: focusing on the speed of new knowledge acquisition qKR, social influence factors, and differentiated knowledge acquisition process; Among them, in a team environment, the social cognition of an individual will be influenced by his peers, reflecting the degree to which the social dynamics of the team influences his personal views, that is, the social influence factor λ t , the calculation expression is as follows: Among them, λ t A high coefficient value indicates that the individual is more susceptible to peer pressure and social influence, while a low coefficient value indicates that the individual is more independent; NFE t Indicates the number of iterations of the current function, NFE max Indicates the maximum number of function iterations; Then, every individual in the population is traversed. Each individual performs divergent or convergent thinking based on its own identity, and increases team diversity. Before performing divergent or convergent thinking and increasing team diversity, each individual first acquires differentiated knowledge. Among them, differentiated knowledge acquisition DKA is aimed at the different potentials of human beings to acquire knowledge, and focuses on the speed of new knowledge acquisition qKR; η i,t _qkr represents the quantitative knowledge acquisition rate of the individual at the tth iteration, and the expression is as follows: in, It represents the individual at the tth iteration The coefficient value is calculated as follows: Among them, R i,t represents the rank of the i-th individual at the beginning of the t-th iteration; The effect of the process of differentiated knowledge acquisition on each individual is expressed as follows: j rand =randint(1,D) Among them, v i,d Represents the test element v i,t The element at position d, v i,t represents the trial element v at the tth iteration i ; η i,t _qkr represents the quantitative knowledge acquisition rate of the individual at the tth iteration; j rand represents a randomly selected integer from 1 to D; S32, judging whether the population individual is a high performer, if so, acquiring differentiated knowledge, and then executing divergent thinking behavior, after which step S35 is entered, if not, step S33 is entered; When the individual is a high performer, he first acquires differentiated knowledge and then performs divergent thinking, that is, injecting the concept of divergent thinking into the experimental vector v i,d In the generation of , the divergent thinking strategy expression is as follows: v i,d =x r1,d +P(α,σ) Among them, x r1,d represents the dth position of an individual randomly selected from {1,2,...,NP}, and x r1 ≠x i ≠x best , x i represents the current individual, x best represents the best performing individual at present; Lk(α,σ) represents an additional term from the Linnik distribution, which is used to encapsulate imagination, α and σ represent the control parameters of the Linnik distribution, and α=0.618, σ=0.05; S33, determine whether the individual in the population is a low performer, if so, acquire differentiated knowledge, and then perform the operation of increasing team diversity and enter step S35, otherwise the individual is a general performer, and enter step S34; When this iteration traverses to the last individual in the population, if the generated random number is less than the probability constant Pc, the individual in the population is a low performer. First, differentiated knowledge is acquired and then the operation of increasing team diversity is performed. The expression for generating new members is as follows: v i,d =LB+U(0,1)×(UB-LB) Among them, U(0,1) represents the uniform distribution on the interval (0,1), LB and UB represent the lower and upper bounds specified by the optimization problem; S34, if the individual is a general performer, first acquire differentiated knowledge, then perform the operation of convergent thinking and enter step S35; When the execution conditions of high performers and low performers are not met, they are general performers. They first acquire differentiated knowledge and then perform thinking convergence behavior. The thinking convergence strategy expression is as follows: v i,d =w×x best,d +λ t ×(x r2,d -x i,d )+ω i,t ×(x r1,d -x i,d ) Among them, x best,d represents the dth position of the best performing individual in the current iteration; w represents the best cognitive weight, and its value is 1; x r1,d represents the dth position of an individual randomly selected from {1,2,...,NP}, and x r1 ≠x i ≠x best , they represent three different individuals, x i represents the current individual; x r2,d represents the dth position of a random individual selected from {ngs+1,...,NP}, and x r2 ≠x r1 ≠x i ≠x best , they represent four different individuals; the coefficient ω represents the learning intensity state of the individual; ω i,t It represents the efficiency value of the individual at the tth iteration, and the expression is as follows: oh i,t ~U(0,1) S35. Implement boundary handling mechanisms to ensure that the ideas generated remain within the scope of feasibility and reality; If the idea proposed by an individual exceeds the actual limit of the dth position, the boundary processing mechanism will correct the value to the closest maximum or minimum value within the range allowed by the data set. The calculation formula is as follows: Among them, v i,d Represents the test element v i,t The element at position d, v i,t represents the trial element v at the tth iteration i LB d and UB d denote the lower and upper bounds of the dth dimension specified by the optimization problem respectively; S36, update the individual position, and then determine whether it is the maximum number of iterations. If not, return to step S31 to perform the next iteration until the maximum number of iterations is met, end the iteration loop, and enter step S4.

Images