UAV Target Allocation Method, Device and Equipment Based on Diversified Enhanced Genetic Algorithm

Abstract

The present application provides a method, device and equipment for UAV target allocation based on a diverse enhanced genetic algorithm, including: selecting a first parent individual and a second parent individual from a first population; wherein, the first population contains multiple parent individuals, exchanging at least one first corresponding relationship at the same position in the first parent individual and the second parent individual to obtain a first offspring individual, determining the first fitness of multiple parent individuals and the first offspring individual respectively based on the weight values corresponding to each detection point to be detected in the multiple parent individuals and the first offspring individual and the detection and recognition rate of the UAV for the detection point to be detected, determining the target individual corresponding to the maximum value in the first fitness from the first offspring individual and the multiple parent individuals, and allocating UAVs to each detection point to be detected based on the first corresponding relationship included in the target individual, which improves the matching degree between the UAV and the detection point to be detected to a certain extent.

Description

Technical Field

[0001] The present application relates to the technical field of unmanned aerial vehicles, and in particular to an unmanned aerial vehicle target assignment method, device and equipment based on a diverse enhanced genetic algorithm. Background Art

[0002] At present, due to the characteristic of not requiring a manned operation, unmanned aerial vehicles have been widely used in recent years. Among them, Unmanned Aerial Vehicle Target Assignment (UTA) is a key part of the command and control of unmanned aerial vehicles, which involves how to assign a series of targets to available unmanned aerial vehicles to complete a given task in an optimized manner. The UTA task needs to improve the efficiency of task execution, reduce costs, and increase the success rate of target execution on the premise of meeting task requirements and constraints.

[0003] In related technologies, when the scale of the unmanned aerial vehicle assignment task is small, techniques such as dynamic programming, branch and bound, and Lagrangian relaxation can be used for accurate solution; when the scale of the unmanned aerial vehicle assignment task is large, evolutionary algorithms and heuristic rules are used to obtain approximate solutions.

[0004] However, in practice, the matching degree between the points to be detected and the actually assigned unmanned aerial vehicles in the above methods is relatively low. Summary of the Invention

[0005] In view of the above problems, embodiments of the present application provide an unmanned aerial vehicle target assignment method, device, electronic device and readable storage medium based on a diverse enhanced genetic algorithm, so as to overcome the above problems or at least partially solve the above problems.

[0006] In a first aspect of embodiments of the present application, an unmanned aerial vehicle target assignment method based on a diverse enhanced genetic algorithm is provided, and the method includes:

[0007] Select a first parent individual and a second parent individual from a first population; wherein, the first population includes a plurality of parent individuals; wherein, each parent individual includes a first correspondence relationship between a plurality of points to be detected and a plurality of unmanned aerial vehicles, and the first correspondence relationships included in different parent individuals are different;

[0008] Exchange at least one first correspondence relationship at the same position in the first parent individual and the second parent individual to obtain a first offspring individual;

[0009] Based on the weight values respectively corresponding to each point to be detected in the plurality of parent individuals and the first offspring individual and the detection and recognition rate of the unmanned aerial vehicle for the point to be detected, determine the first fitness of the plurality of parent individuals and the first offspring individual respectively;

[0010] Determine a target individual corresponding to the maximum value in the first fitness from the first-generation individuals and the multiple parental individuals;

[0011] Based on the first correspondence relationship included in the target individual, allocate the drones to the respective points to be detected.

[0012] Optionally, the selecting the first parental individual and the second parental individual from the first population includes:

[0013] Generate a global information matrix of the first population based on each parental individual included in the first population and the second fitness respectively corresponding to each parental individual; wherein, the rows of the global information matrix correspond to the drones one by one, the columns of the global information matrix correspond to the points to be detected one by one, and the first value at each position in the global information matrix is determined based on the second fitness respectively corresponding to each parental individual;

[0014] Determine the first maximum value of the first values included in each column of the global information matrix from the global information matrix;

[0015] Based on the row number and column number of the first maximum value in the global information matrix, determine a second correspondence relationship between the multiple points to be detected and the multiple drones;

[0016] Generate the first parental individual based on the second correspondence relationship;

[0017] Randomly select a second parental individual from the first population.

[0018] Optionally, the determining the second correspondence relationship between the multiple points to be detected and the multiple drones based on the row number and column number of the first maximum value in the global information matrix includes:

[0019] In the case where there are second maximum values with the same corresponding column numbers in the first maximum values, based on the row numbers and column numbers of the remaining maximum values except the second maximum values in the first maximum values, determine a third correspondence relationship between the multiple points to be detected and the multiple drones;

[0020] Determine the third maximum value of the first value from the remaining rows and remaining columns except the row numbers and column numbers corresponding to the remaining maximum values in the global information matrix;

[0021] Based on the row number and column number of the third maximum value in the global information matrix, determine a fourth correspondence relationship between the multiple points to be detected and the multiple drones;

[0022] Integrate the third corresponding relationship and the fourth corresponding relationship to obtain the second corresponding relationship between the multiple points to be detected and the multiple drones.

[0023] Optionally, the selecting the first parent individual and the second parent individual from the first population includes:

[0024] Generate a first data sequence based on a preset chaotic mapping;

[0025] Arbitrarily select the total number of points to be detected of second data from the first data included in the first data sequence;

[0026] Arrange the first order of the second data in the first data sequence in ascending order to obtain a first order sequence; wherein, the value in the first order sequence represents the first order of the second data in the first data sequence;

[0027] Arrange the second data in descending order to obtain a second data sequence of the second data;

[0028] Based on the corresponding relationship between the second data and each value in the first order sequence, respectively adjust each value in the first order sequence to the same position of the second data in the second data sequence to obtain a second order sequence;

[0029] Select a first parent individual corresponding to the first corresponding relationship that matches the second corresponding relationship included in the second order sequence from the first population;

[0030] Randomly select a second parent individual different from the first parent individual from the first population.

[0031] Optionally, the swapping at least one same-position first corresponding relationship between the first parent individual and the second parent individual to obtain a first offspring individual includes:

[0032] Randomly generate a first value;

[0033] In the case where the first value is greater than a first threshold, swap at least one same-position first corresponding relationship between the first parent individual and the second parent individual to obtain a first offspring individual;

[0034] In the case where the first value is less than or equal to the first threshold, swap at least one same-position first corresponding relationship between the first parent individual and the second parent individual to obtain a second offspring individual, and mutate the second offspring individual to obtain a first offspring individual.

[0035] Optionally, the mutating the second offspring individual to obtain a first offspring individual includes:

[0036] When the first value is less than the second threshold, mutate the second offspring individual based on the Levy flight mutation strategy to obtain the first offspring individual; wherein, the second threshold is less than the first threshold.

[0037] When the first value is greater than or equal to the second threshold and less than the first threshold, randomly select a plurality of first unmanned aerial vehicles (UAVs) from the second offspring individual, and randomly exchange the positions of the plurality of first UAVs in the second offspring individual to obtain the first offspring individual.

[0038] Optionally, determining the target individual corresponding to the maximum value in the first fitness from the first offspring individual and the plurality of parent individuals includes:

[0039] Add the first offspring individual to the first population, and delete the first individual with the lowest first fitness from the first population to obtain the iterated first population; wherein, the number of the first individuals is the same as the number of the first offspring individuals.

[0040] Re - execute the step of selecting the first parent individual and the second parent individual from the first population.

[0041] When the number of iterations of the iterated first population reaches the preset number of iterations, determine the target individual corresponding to the maximum value in the first fitness from the iterated first population.

[0042] Optionally, determining the target individual corresponding to the maximum value in the first fitness from the first offspring individual and the plurality of parent individuals includes:

[0043] Screen the second individuals from the first offspring individual and the plurality of parent individuals; wherein, the UAVs corresponding to the first corresponding relationships included in the second individuals are pairwise different.

[0044] Determine the target individual corresponding to the maximum value in the first fitness from the second individuals.

[0045] In a second aspect, an embodiment of the present application provides a UAV target allocation device based on a diverse enhanced genetic algorithm, and the device includes:

[0046] A selection module, configured to select a first parent individual and a second parent individual from a first population; wherein, the first population includes a plurality of parent individuals; wherein, each parent individual includes a first corresponding relationship between a plurality of detection points and a plurality of UAVs, and the first corresponding relationships included in different parent individuals are different.

[0047] An exchange module, configured to exchange at least one first corresponding relationship at the same position between the first parental individual and the second parental individual to obtain a first offspring individual;

[0048] A first determination module, configured to respectively determine first fitnesses of the multiple parental individuals and the first offspring individual based on weight values respectively corresponding to each detection point to be detected among the multiple parental individuals and the first offspring individual and detection recognition rates of the UAVs for the detection points to be detected;

[0049] A second determination module, configured to determine a target individual corresponding to the maximum value in the first fitnesses from the first offspring individual and the multiple parental individuals;

[0050] An allocation module, configured to allocate the UAVs to each detection point to be detected based on the first corresponding relationship included in the target individual.

[0051] Optionally, the selection module includes:

[0052] A first generation sub-module, configured to generate a global information matrix of the first population based on each parental individual included in the first population and second fitnesses respectively corresponding to the parental individuals; wherein, rows of the global information matrix correspond to the UAVs one by one, columns of the global information matrix correspond to the detection points to be detected one by one, and a first value at each position in the global information matrix is determined based on the second fitnesses respectively corresponding to the parental individuals;

[0053] A first determination sub-module, configured to determine a first maximum value of the first values included in each column of the global information matrix from the global information matrix;

[0054] A second determination sub-module, configured to determine a second corresponding relationship between the multiple detection points to be detected and the multiple UAVs based on the row number and column number of the first maximum value in the global information matrix;

[0055] A second generation sub-module, configured to generate a first parental individual based on the second corresponding relationship;

[0056] A first selection sub-module, configured to randomly select a second parental individual from the first population.

[0057] Optionally, the second determination sub-module includes:

[0058] A first determination unit, configured to, when there are second maximum values with the same corresponding column numbers in the first maximum values, determine a third corresponding relationship between the multiple detection points to be detected and the multiple UAVs based on the row numbers and column numbers of the remaining maximum values except the second maximum values in the first maximum values in the global information matrix;

[0059] A second determination unit, configured to determine a third maximum value of the first numerical value from the remaining rows and remaining columns in the global information matrix except for the row number and column number corresponding to the remaining maximum value;

[0060] A third determination unit, configured to determine a fourth correspondence between the multiple points to be detected and the multiple drones based on the row number and column number of the third maximum value in the global information matrix;

[0061] An integration unit, configured to integrate the third correspondence and the fourth correspondence to obtain a second correspondence between the multiple points to be detected and the multiple drones.

[0062] Optionally, the selection module includes:

[0063] A third generation sub-module, configured to generate a first data sequence based on a preset chaotic mapping;

[0064] A second selection sub-module, configured to arbitrarily select a total number of second data equal to the number of points to be detected from the first data included in the first data sequence;

[0065] A first sorting sub-module, configured to sort the first order of the second data in the first data sequence in ascending order to obtain a first order sequence; wherein, the numerical value in the first order sequence represents the first order of the second data in the first data sequence;

[0066] A second sorting sub-module, configured to sort the second data in descending order to obtain a second data sequence of the second data;

[0067] An adjustment sub-module, configured to adjust each numerical value in the first order sequence to the same position of the second data in the second data sequence based on the correspondence between the second data and each numerical value in the first order sequence, to obtain a second order sequence;

[0068] A third selection sub-module, configured to select a first parental individual corresponding to a first correspondence that matches the second correspondence included in the second order sequence from a first population;

[0069] A fourth selection sub-module, configured to randomly select a second parental individual different from the first parental individual from the first population.

[0070] Optionally, the exchange module includes:

[0071] A second generation sub-module, configured to randomly generate a first numerical value;

[0072] The first exchange sub-module is used to exchange at least one first corresponding relationship at the same position between the first parent individual and the second parent individual to obtain a first offspring individual when the first value is greater than a first threshold value;

[0073] The exchange mutation sub-module is used to exchange at least one first corresponding relationship at the same position between the first parent individual and the second parent individual to obtain a second offspring individual when the first value is less than or equal to the first threshold value, and mutate the second offspring individual to obtain a first offspring individual.

[0074] Optionally, the exchange mutation sub-module includes:

[0075] The first mutation unit is used to mutate the second offspring individual based on the Lévy flight mutation strategy to obtain a first offspring individual when the first value is less than a second threshold value; wherein, the second threshold value is less than the first threshold value;

[0076] The second mutation unit is used to randomly select a plurality of first unmanned aerial vehicles from the second offspring individual and randomly exchange the positions of the plurality of first unmanned aerial vehicles in the second offspring individual to obtain a first offspring individual when the first value is greater than or equal to the second threshold value and less than the first threshold value.

[0077] Optionally, the second determination module includes:

[0078] The iteration sub-module is used to add the first offspring individual into the first population and delete the first individual with the lowest first fitness from the first population to obtain an iterated first population; wherein, the number of the first individuals is the same as the number of the first offspring individuals;

[0079] The step execution sub-module is used to re-execute the step of selecting the first parent individual and the second parent individual from the first population;

[0080] The first determination sub-module is used to determine a target individual corresponding to the maximum value in the first fitness from the iterated first population when the number of iterations of the iterated first population reaches a preset number of times.

[0081] Optionally, the second determination module includes:

[0082] The screening sub-module is used to screen a second individual from the first offspring individual and the plurality of parent individuals; wherein, the unmanned aerial vehicles corresponding to the first corresponding relationships included in the second individual are pairwise different;

[0083] The second determination sub-module is used to determine a target individual corresponding to the maximum value in the first fitness from the second individual.

[0084] In a third aspect, an embodiment of the present application provides an electronic device, including a memory, a processor, and a computer program stored on the memory. The processor executes the computer program to implement the drone target allocation method based on the diverse enhanced genetic algorithm as described in any one of the above.

[0085] In a fourth aspect, an embodiment of the present application provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the drone target allocation method based on the diverse enhanced genetic algorithm as described in any one of the above is implemented.

[0086] Specific beneficial effects are as follows:

[0087] In the embodiment of the present application, a first parent individual and a second parent individual are selected from a first population; wherein, the first population includes multiple parent individuals; wherein, each parent individual includes a first correspondence between multiple points to be detected and multiple drones, and the first correspondences included in different parent individuals are different. At least one first correspondence at the same position in the first parent individual and the second parent individual is exchanged to obtain a first offspring individual. Based on the weight values respectively corresponding to each point to be detected in multiple parent individuals and the first offspring individual and the detection and recognition rate of the drone for the point to be detected, the first fitness of multiple parent individuals and the first offspring individual is respectively determined. The target individual corresponding to the maximum value in the first fitness is determined from the first offspring individual and multiple parent individuals. Based on the first correspondence included in the target individual, drones are allocated to each point to be detected. The method of individual optimization can be carried out through population iteration to obtain the target individual with the maximum first fitness, and drones are allocated to each point to be detected according to the first correspondence between the drones and the points to be detected included in the target individual, which improves the matching degree between the drones and the points to be detected to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0089] Figure 1 is a schematic flowchart of a drone target allocation method based on a diverse enhanced genetic algorithm provided by an embodiment of the present application;

[0090] Figure 2 is a schematic flowchart of another drone target allocation method based on a diverse enhanced genetic algorithm provided by an embodiment of the present application;

[0091] Figure 3 It is a schematic diagram of a parental individual and a global information matrix provided by an embodiment of the present application;

[0092] Figure 4 It is another schematic diagram of a parental individual and a global information matrix provided by an embodiment of the present application;

[0093] Figure 5 It is a schematic diagram of the simulation of Levy flight provided by an embodiment of the present application;

[0094] Figure 6 It is a flowchart of an algorithm of a diversity-enhanced genetic algorithm provided by an embodiment of the present application;

[0095] Figure 7 It is a logic block diagram of an unmanned aerial vehicle target allocation device based on a diversity-enhanced genetic algorithm provided by an embodiment of the present application;

[0096] Figure 8 It is a schematic diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0097] The exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings in the embodiments of the present application. Although the exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present application can be more thoroughly understood and the scope of the present application can be fully conveyed to those skilled in the art.

[0098] Refer to Figure 1 , Figure 1 It is a schematic flowchart of a method for allocating unmanned aerial vehicle targets based on a diversity-enhanced genetic algorithm provided by an embodiment of the present application. The method includes:

[0099] Step 101: Select a first parental individual and a second parental individual from a first population; wherein, the first population includes a plurality of parental individuals; wherein, the parental individual includes a first correspondence between a plurality of points to be detected and a plurality of unmanned aerial vehicles, and the first correspondences included in different parental individuals are different.

[0100] In the embodiment of the present application, a first correspondence between each point to be detected and an unmanned aerial vehicle can be established, and based on this first correspondence, parental individuals and a first population that can be iterated can be constructed. Wherein, the number of points to be detected can be the same as the number of unmanned aerial vehicles.

[0101] For example, if there are five detection points to be detected in the target scenario, namely a1, a2, a3, a4, a5, and five drones, namely b1, b2, b3, b4, b5, then the five detection points to be detected and the five drones can be associated one by one to establish a first correspondence, such as a1~b2. If vectors are used to represent parental individuals, and the element positions in the vector respectively correspond to the five detection points to be detected, then the parental individual can be represented in the form of Equation 1 as follows:

[0102] a[i] = {b j1 , j2 , j3 , j4 , j5} (Equation 1)

[0103] In Equation 1, a[i] represents the parental individual. The five element positions in a[i] can respectively correspond to the detection points to be detected "a1, a2, a3, a4, a5" from front to back. j1, j2, j3, j4, j5 can respectively correspond to integers in the interval [1, 5], and j1, j2, j3, j4, j5 can be the same or different from each other. b represents the drone. Thus, b j1 , b j2 , b j3 , b j4 , b j5 can respectively correspond to one of the drones "b1, b2, b3, b4, b5". Among them, b j1 is located in the first position of a[i], indicating the first correspondence between the drone b j1 and the detection point a1. Exemplarily, one parental individual can be represented as a[1] = {2, 4, 5, 3, 1}, where "2" is located in the first position of a[1], indicating the first correspondence between the drone b2 and the detection point a1.

[0104] In the embodiments of the present application, after establishing a sufficient number of parental individuals, the parental individuals can be integrated to form the first population, denoted as A x = {a x [i]}, where x represents the iteration number of the first population. Thus, the first parental individual and the second parental individual can be selected from the first population. Among them, when selecting the first parental individual and the second parental individual, a random selection method can be used, or a roulette wheel method can be used to select based on the fitness of the parental individuals. It should be noted that if the first population has not been iterated, the fitness of each parental individual can be calculated when generating the parental individuals.

[0105] Step 102, exchange at least one first correspondence at the same position in the first parental individual and the second parental individual to obtain the first offspring individual.

[0106] In an embodiment of the present application, at least one first corresponding relationship at the same position in the first parent individual and the second parent individual can be exchanged to obtain a first offspring individual.

[0107] For example, if the first parent individual is a[1] = {2, 4, 5, 3, 1} and the second parent individual is a[2] = {4, 5, 4, 3, 1}, by exchanging "a[1][1] = 2" and "a[2][1] = 4", the first offspring individuals c[1] = {4, 4, 5, 3, 1} and c[2] = {2, 5, 4, 3, 1} can be obtained; by exchanging "a[1][2] = 4" and "a[2][2] = 5", the first offspring individuals c[1] = {2, 5, 5, 3, 1} and c[2] = {4, 4, 4, 3, 1} can be obtained; or the above two exchanges can be performed simultaneously to obtain the first offspring individuals c[1] = {4, 5, 5, 3, 1} and c[2] = {2, 4, 4, 3, 1}.

[0108] Step 103: Based on the weight values corresponding to each detection point to be detected in the multiple parent individuals and the first offspring individual, and the detection recognition rate of the drone for the detection point to be detected, respectively determine the first fitness of the multiple parent individuals and the first offspring individual.

[0109] In an embodiment of the present application, each detection point to be detected can correspond to a fixed weight value, indicating the detection value of the detection point. The larger the weight value, the higher the detection value of the detection point. The weight value of the detection point to be detected can usually be calibrated manually. In addition, for the detection point to be detected, different drones can also have different detection recognition rates. The detection recognition rate can be given by the manufacturer of the detection device carried on the drone, or determined through experiments, and the ratio of the number of successful detections of the sample detection point to the total number of detections is used as the detection recognition rate. In this way, each first corresponding relationship included in the parent individual and the first offspring individual can correspond to a weight value of the detection point to be detected and the detection recognition rate of the drone for the detection point to be detected, so that the first fitness of the multiple parent individuals and the first offspring individual can be calculated according to the weight value of the detection point to be detected and the detection recognition rate of the drone for the detection point to be detected, as shown in Equation 2 below:

[0110]

[0111] In Equation 2, f represents the first fitness, i represents the position of the first corresponding relationship in the multiple parent individuals and the first offspring individual, v i represents the weight value of the detection point to be detected, and P ij represents the detection recognition rate of drone j for detection point i.

[0112] Step 104: Determine the target individual corresponding to the maximum value in the first fitness value from the first offspring individual and the multiple parent individuals.

[0113] In the embodiment of the present application, after calculating the first fitness values of multiple parent individuals and the first offspring individual, the maximum value of the first fitness value can be determined first, and then the target individual corresponding to the maximum value of the first fitness value is determined from the first offspring individual and the multiple parent individuals. For example, if the first offspring individual corresponds to the maximum value of the first fitness value, the first offspring individual corresponding to the maximum value of the first fitness value can be determined as the target individual; if the parent individual corresponds to the maximum value of the first fitness value, the parent individual corresponding to the maximum value of the first fitness value can be determined as the target individual.

[0114] Step 105: Allocate the drones to each detection point based on the first correspondence relationship included in the target individual.

[0115] In the embodiment of the present application, the number of the first correspondence relationships included in the target individual is the same as the number of detection points. Therefore, the drones can be allocated to each detection point according to the first correspondence relationship included in the target individual.

[0116] Continuing with the above example, if the target individual is c[3] = {4, 5, 2, 3, 1}, the drone b4 can be allocated to the detection point a1, the drone b5 can be allocated to the detection point a2, the drone b2 can be allocated to the detection point a3, the drone b3 can be allocated to the detection point a4, and the drone b1 can be allocated to the detection point a5.

[0117] In the embodiment of the present application, by selecting the first parent individual and the second parent individual from the first population; wherein, the first population includes multiple parent individuals; wherein, the parent individual includes the first correspondence relationship between multiple detection points and multiple drones, and the first correspondence relationships included in different parent individuals are different, exchanging at least one first correspondence relationship at the same position in the first parent individual and the second parent individual to obtain the first offspring individual, determining the first fitness values of the multiple parent individuals and the first offspring individual respectively based on the weight values corresponding to each detection point in the multiple parent individuals and the first offspring individual and the detection and recognition rate of the drone for the detection point, determining the target individual corresponding to the maximum value in the first fitness value from the first offspring individual and the multiple parent individuals, and allocating the drones to each detection point based on the first correspondence relationship included in the target individual, the individual optimization can be carried out by the method of population iteration to obtain the target individual with the maximum first fitness value, and the drones are allocated to each detection point according to the first correspondence relationship between the drones and the detection points included in the target individual, which improves the matching degree between the drones and the detection points to a certain extent.

[0118] Refer to Figure 2, Figure 2 It is a schematic flowchart of another UAV target allocation method provided by an embodiment of the present application. This method may include:

[0119] Step 201: Select a first parent individual and a second parent individual from the first population; wherein, the first population includes multiple parent individuals; wherein, each parent individual includes a first correspondence between multiple detection points and multiple UAVs, and the first correspondences included in different parent individuals are different.

[0120] In the embodiments of the present application, the implementation content of this step can refer to the embodiment content of step 101, which will not be elaborated here.

[0121] Optionally, step 201 may include the following sub-steps:

[0122] Sub-step 20111: Generate a global information matrix of the first population based on each parent individual included in the first population and the second fitness corresponding to each parent individual respectively; wherein, the rows of the global information matrix correspond to UAVs one by one, the columns of the global information matrix correspond to detection points one by one, and the first value at each position in the global information matrix is determined based on the second fitness corresponding to each parent individual respectively.

[0123] In the embodiments of the present application, a global information matrix of the first population can be generated according to each parent individual included in the first population and the second fitness corresponding to each parent individual respectively. The calculation method of the second fitness can refer to the calculation method of the first fitness, which will not be elaborated here. Among them, the rows of the global information matrix correspond to UAVs one by one, the columns of the global information matrix correspond to detection points one by one, and the first value at each position in the global information matrix is determined based on the second fitness corresponding to each parent individual respectively. Specifically, the row number of the global information matrix can correspond to the UAV, the column number of the global information matrix can correspond to the detection point, and the first value at each position of the global information matrix can be obtained by adding the second fitness of each parent individual corresponding to this position.

[0124] For example, referring to Figure 3 , Figure 3 a schematic diagram of a parent individual and a global information matrix is given. The figure includes two parent individuals, namely a[1] = {5, 3, 1, 4, 2} and a[2] = {1, 2, 3, 4, 5}, where the second fitness of a[1] is 10 and the second fitness of a[2] is 5, then the generated global information matrix is Figure 3The 5×5 matrix shown on the right. Among them, the maximum number of rows of the matrix is equal to the number of drones, and the maximum number of columns of the matrix is equal to the number of points to be detected. Since the rows of the global information matrix correspond to the drones and the columns of the global information matrix correspond to the points to be detected, for the parental individual a[1], the second fitness value "10" is placed in the 5th row and 1st column of the global information matrix, indicating that drone b5 corresponds to the point to be detected a1, that is, the correspondence between the drone represented by the element "5" in the first position of the parental individual a[1] and the point to be detected. Similarly, the second fitness value "10" can be filled in the 3rd row and 2nd column, the 1st row and 3rd column, the 4th row and 4th column, and the 2nd row and 5th column respectively. Similar to the parental individual a[1], the second fitness value "5" of the parental individual a[2] can be filled in the 1st row and 1st column, the 2nd row and 2nd column, the 3rd row and 3rd column, the 4th row and 4th column, and the 5th row and 5th column of the global information matrix respectively. In this way, the global information matrix formed by the parental individuals a[1] and a[2] can be constructed.

[0125] Sub-step 20112, determine the first maximum value of the first numerical values included in each column of the global information matrix from the global information matrix.

[0126] In the embodiment of the present application, after obtaining the global information matrix corresponding to the first population, the first maximum value of the first numerical values included in each column of the global information matrix can be determined in the global information matrix.

[0127] Continuing with the above example, for Figure 3 the global information matrix, the first maximum value of the first numerical values included in the first column is 10, the first maximum value of the first numerical values included in the second column is 10, the first maximum value of the first numerical values included in the third column is 10, the first maximum value of the first numerical values included in the fourth column is 15, and the first maximum value of the first numerical values included in the fifth column is 10.

[0128] Sub-step 20113, determine the second correspondence between the multiple points to be detected and the multiple drones based on the row number and column number of the first maximum value in the global information matrix.

[0129] In the embodiment of the present application, the second correspondence between the multiple points to be detected and the multiple drones can be determined according to the row number and column number of the first maximum value in the global information matrix.

[0130] Continuing with the above example, the row numbers and column numbers of the first maximum values in the global information matrix are the 5th row and the 1st column, the 3rd row and the 2nd column, the 1st row and the 3rd column, the 4th row and the 4th column, and the 2nd row and the 5th column respectively. Then the second correspondence may include: drone b5 corresponds to the point to be detected a1, drone b3 corresponds to the point to be detected a2, drone b1 corresponds to the point to be detected a3, drone b4 corresponds to the point to be detected a4, and drone b2 corresponds to the point to be detected a5.

[0131] Optionally, sub-step 20113 may include the following sub-steps:

[0132] Sub-step A1, when there are second maximum values with the same corresponding column numbers in the first maximum values, based on the row numbers and column numbers of the remaining maximum values in the global information matrix except the second maximum values in the first maximum values, determine the third correspondence between the multiple points to be detected and the multiple drones.

[0133] In the embodiments of the present application, if there are second maximum values with the same corresponding row numbers in the first maximum values, then the third correspondence between the multiple points to be detected and the multiple drones may be determined first according to the row numbers and column numbers of the remaining maximum values in the global information matrix except the second maximum values in the first maximum values.

[0134] For example, as Figure 4 shown, Figure 4 is a schematic diagram of another parent individual and the global information matrix provided by the embodiments of the present application. The figure includes three parent individuals a[1] = {5, 3, 1, 4, 2}, a[2] = {1, 2, 3, 4, 5} and a[3] = {5, 4, 3, 2, 1}. The global information matrix formed by the above three parent individuals is as shown in the 5×5 matrix in Figure 4 In this global information matrix, there are the same second maximum values "10" in the second column and the fifth column. Therefore, the above two second maximum values can be deleted from the first maximum values to obtain the remaining maximum values "20" and "15". Among them, the remaining maximum value "20" is in the 5th row and the 1st column of the global information matrix, and the remaining maximum value "15" is respectively in the 3rd row and the 3rd column and the 4th row and the 4th column of the global information matrix. Thus, the third correspondence can be obtained as follows: drone b5 corresponds to the point to be detected a1, drone b3 corresponds to the point to be detected a3, and drone b4 corresponds to the point to be detected a4.

[0135] Sub-step A2, determine the third maximum value of the first value from the remaining rows and remaining columns in the global information matrix except the row numbers and column numbers corresponding to the remaining maximum values.

[0136] In an embodiment of the present application, the third maximum value of the first numerical value may be determined in the remaining rows and remaining columns of the global information matrix except for the row number and column number corresponding to the remaining maximum value.

[0137] Continuing with the above example, in Figure 4 the global information matrix, the 5th row, 3rd row, 4th row, and the 1st column, 3rd column, and 4th column all correspond to the remaining maximum value. Therefore, the remaining rows of the global information matrix are the 1st row and the 2nd row, and the remaining columns are the 2nd column and the 5th column. Among the positions corresponding to the above remaining rows and remaining columns, the third maximum values of the first numerical value can be determined as "5" and "10".

[0138] Sub-step A3: Based on the row number and column number of the third maximum value in the global information matrix, determine the fourth correspondence between the multiple points to be detected and the multiple drones.

[0139] In an embodiment of the present application, the fourth correspondence between the multiple points to be detected and the multiple drones may be determined according to the row number and column number of the third maximum value in the global information matrix. Among them, if the third maximum value does not correspond to the same column number, the fourth correspondence can be directly determined; if the third maximum value corresponds to the same column number, the column corresponding to the third maximum value can be suspended, first determine the fourth correspondence of other columns, and then determine the fourth correspondence of the suspended column.

[0140] Continuing with the above example, Figure 4 in, the third maximum value "10" included in the 5th column corresponds to the same column number (that is, there are multiple third maximum values in this column), then the fourth correspondence corresponding to the 5th column may not be determined first. At this time, the remaining column is the 2nd column. Then, the third maximum value "5" can be determined in the 1st row and the 2nd row of the 2nd column. Since the third maximum value "5" is located in the 2nd row and the 2nd column, the following fourth correspondence can be obtained: drone b2 corresponds to the point to be detected a2. At this time, since the 2nd row has been used, for the 5th column, it can be determined that the third maximum value "10" is located in the 1st row and the 5th column, so the following fourth correspondence can be obtained: drone b1 corresponds to the point to be detected a5.

[0141] Sub-step A4: Integrate the third correspondence and the fourth correspondence to obtain the second correspondence between the multiple points to be detected and the multiple drones.

[0142] In an embodiment of the present application, the third correspondence and the fourth correspondence may be integrated, so that the second correspondence between the multiple points to be detected and the multiple drones can be obtained.

[0143] Continuing with the above example, the third correspondence includes: drone b5 corresponds to detection point a1, drone b3 corresponds to detection point a3, and drone b4 corresponds to detection point a4. The fourth correspondence includes: drone b2 corresponds to detection point a2, and drone b1 corresponds to detection point a5. After integrating the third correspondence and the fourth correspondence, the second correspondence can be obtained as follows: drone b5 corresponds to detection point a1, drone b2 corresponds to detection point a2, drone b3 corresponds to detection point a3, drone b4 corresponds to detection point a4, and drone b1 corresponds to detection point a5.

[0144] In the embodiments of the present application, when there is a second maximum value with the same number of corresponding columns in the first maximum value, based on the number of rows and columns of the remaining maximum values except the second maximum value in the first maximum value in the global information matrix, the third correspondence between multiple detection points and multiple drones is determined. From the remaining rows and remaining columns in the global information matrix except the rows and columns corresponding to the remaining maximum values, the third maximum value of the first value is determined. Based on the number of rows and columns of the third maximum value in the global information matrix, the fourth correspondence between multiple detection points and multiple drones is determined. By integrating the third correspondence and the fourth correspondence, the second correspondence between multiple detection points and multiple drones is obtained. When the columns in the global information matrix contain the maximum values of the same first value, the third correspondence without the maximum values of the same first value can be preferentially determined, and then the fourth correspondence with the maximum values of the same first value can be determined. Finally, by integrating the third correspondence and the fourth correspondence to obtain the second correspondence, it can be ensured that each detection point in the second correspondence corresponds to a different drone, which fully improves the diversity and richness of the second correspondence.

[0145] Sub-step 20114: Generate the first parental individual based on the second correspondence.

[0146] In the embodiments of the present application, the first parental individual can be generated according to the second correspondence between multiple detection points and multiple drones.

[0147] Continuing with sub-step 20113 and Figure 3 the example, the generated parental individual is as Figure 3 shown directly below the global information matrix in the figure, which is a[4] = {5, 3, 1, 4, 2}.

[0148] Continuing with sub-step A4 and Figure 4 the example, the generated parental individual is as Figure 4 shown directly below the global information matrix in the figure, which is a[4] = {5, 2, 3, 4, 1}.

[0149] Sub-step 20115: Randomly select a second parental individual from the first population.

[0150] In an embodiment of the present application, another parental individual can be randomly selected from the first population as the second parental individual. Among them, the second parental individual and the first parental individual can be different.

[0151] In an embodiment of the present application, by generating a global information matrix of the first population based on each parental individual included in the first population and the second fitness corresponding to each parental individual respectively; wherein, the rows of the global information matrix correspond to the drones one by one, the columns of the global information matrix correspond to the points to be detected one by one, the first value at each position in the global information matrix is determined based on the second fitness corresponding to each parental individual respectively, the first maximum value of the first values included in each column of the global information matrix is determined from the global information matrix, based on the row number and column number of the first maximum value in the global information matrix, a second correspondence relationship between multiple points to be detected and multiple drones is determined, based on the second correspondence relationship, a first parental individual is generated, and a second parental individual is randomly selected from the first population. The method of obtaining the first parental individual by using the global information matrix and obtaining the second parental individual by random selection can improve the diversity and richness of the first parental individual, thereby avoiding to a certain extent that the individuals in the population fall into local optimality and improving the reliability of the finally obtained target individual to a certain extent.

[0152] Sub-step 20121: Generate a first data sequence based on a preset chaotic mapping.

[0153] In an embodiment of the present application, a preset chaotic mapping can be established in advance, and multiple data are generated through this chaotic mapping to form a first data sequence. The preset chaotic mapping is shown in Equation 3 below:

[0154] X n+1 = X n (1 - X n ) (Equation 3)

[0155] In Equation 3, X n ∈(0, 1) represents the nth generated data, which can retain two significant figures, and μ ∈ [0, 4] is a preset coefficient. Among them, X1 can be generated by a random method.

[0156] Sub-step 20122: Randomly select the total number of points to be detected of second data from the first data included in the first data sequence.

[0157] In an embodiment of the present application, the number of first data included in the first data sequence is usually greater than the number of points to be detected. In this way, multiple second data can be selected from the first data included in the first data sequence, and the number of second data is made the same as the total number of points to be detected. When selecting the second data, a random selection method can be adopted, or a uniform sampling or other sampling methods can be used for selection.

[0158] For example, if the number of points to be detected is 5, the selected second data can be [0.45, 0.60, 0.55, 0.72, 0.34].

[0159] Sub-step 20123: Arrange the first order of the second data in the first data sequence in ascending order to obtain a first order sequence; wherein, the values in the first order sequence represent the first order of the second data in the first data sequence.

[0160] In an embodiment of the present application, the first order of the second data in the first data sequence can be arranged in ascending order to obtain a first order sequence. Obviously, among the second data, the first order of the second data located at the forefront of the first data sequence is 1, and the first order of the second data located at the end of the first data sequence is equal to the number of points to be detected.

[0161] Continuing with the above example, the obtained first order sequence is [1, 2, 3, 4, 5].

[0162] Sub-step 20124: Arrange the second data in descending order to obtain a second data sequence of the second data.

[0163] In an embodiment of the present application, the second data can be arranged in descending order to obtain a second data sequence of the second data.

[0164] Continuing with the above example, if the second data is [0.45, 0.60, 0.55, 0.72, 0.34], after arranging the second data in descending order, the obtained second data sequence is [0.72, 0.60, 0.55, 0.42, 0.34].

[0165] Sub-step 20125: Based on the correspondence between the second data and each value in the first order sequence, adjust each value in the first order sequence to the same position of the second data in the second data sequence to obtain a second order sequence.

[0166] In an embodiment of the present application, the second data and each value in the first-order sequence may have a corresponding relationship. According to this corresponding relationship, each value in the first-order sequence can be respectively adjusted to the same position of the second data in the second data sequence, so that a second-order sequence can be obtained.

[0167] Continuing with the above example, "0.45" in the second data can correspond to the value "1" in the first-order sequence (being in the same position), and "0.72" in the second data can correspond to the value "4" in the first-order sequence. After sorting in descending order, the second data "0.45" is in the fourth position in the second data sequence. Therefore, the value "1" in the first-order sequence can also be adjusted to the fourth position in the first-order sequence; the second data "0.72" is in the first position in the second data sequence. Therefore, the value "4" in the first-order sequence can be adjusted to the first position in the first-order sequence. After adjusting all the values in the first-order sequence corresponding to the second data whose positions have changed, a second-order sequence can be obtained. In the above example, only the positions of two second data have changed. Therefore, the obtained second-order sequence is [4, 2, 3, 1, 5].

[0168] Sub-step 20126: Select the first parental individual corresponding to the first corresponding relationship that matches the second corresponding relationship included in the second-order sequence from the first population.

[0169] In an embodiment of the present application, the first parental individual corresponding to the first corresponding relationship that matches the second corresponding relationship included in the second-order sequence can be selected from the first population. Among them, the second corresponding relationship and the first corresponding relationship being matched may mean that the second corresponding relationship included in the second-order sequence is the same as the first corresponding relationship included in the first parental individual in terms of data representation form.

[0170] Continuing with the above example, if the second-order sequence is [4, 2, 3, 1, 5], the second corresponding relationship may include the following corresponding relationships: the value "4" corresponds to the sequence position "1", the value "2" corresponds to the sequence position "2", the value "3" corresponds to the sequence position "3", the value "1" corresponds to the sequence position "4", and the value "5" corresponds to the sequence position "5". The first corresponding relationship that matches the second corresponding relationship may include the following corresponding relationships: UAV b4 corresponds to detection point a1, UAV b2 corresponds to detection point a2, UAV b3 corresponds to detection point a3, UAV b1 corresponds to detection point a4, and UAV b5 corresponds to detection point a5. In this way, the first parental individual selected according to the second-order sequence can be [4, 2, 3, 1, 5], which is the same as the second-order sequence in terms of data representation form.

[0171] Sub-step 20127: Randomly select a second parental individual different from the first parental individual from the first population.

[0172] In the embodiments of the present application, a second parental individual different from the first parental individual can be randomly selected from the first population. This random selection can be carried out through some algorithms, such as the roulette wheel algorithm for random selection, or can be arbitrarily selected from the first population.

[0173] In the embodiments of the present application, by generating a first data sequence based on a preset chaotic mapping, arbitrarily selecting a total number of detection points of second data from the first data included in the first data sequence, arranging the first order of the second data in the first data sequence in ascending order to obtain a first order sequence; where the value in the first order sequence represents the first order of the second data in the first data sequence, arranging the second data in descending order to obtain a second data sequence of the second data, based on the corresponding relationship between the second data and each value in the first order sequence, respectively adjusting each value in the first order sequence to the same position of the second data in the second data sequence to obtain a second order sequence, selecting the first parental individual corresponding to the first corresponding relationship that matches the second corresponding relationship included in the second order sequence from the first population, randomly selecting a second parental individual different from the first parental individual from the first population, a method of chaotic mapping can be adopted to generate a data sequence, a second order sequence of the first order can be obtained by rearranging the data sequence, and the first parental individual can be screened from the first population through the second order sequence, and at the same time, a method of random selection can be adopted to select the second parental individual, improving the diversity of the first parental individual and can, to a certain extent, improve the reliability of the finally obtained target individual.

[0174] Step 202: Randomly generate a first value.

[0175] In the embodiments of the present application, a method of generating a random number can be adopted to randomly generate a first value.

[0176] Step 203: In the case where the first value is greater than the first threshold, exchange at least one first corresponding relationship at the same position between the first parental individual and the second parental individual to obtain a first filial individual.

[0177] In the embodiments of the present application, the first threshold can be preset and the first threshold is within the value range of the first value. If the first value is greater than the first threshold, then at least one first corresponding relationship at the same position between the first parental individual and the second parental individual can be exchanged to obtain a first filial individual. The specific implementation content of exchanging the first corresponding relationship can refer to the embodiment content of step 102 and will not be elaborated here.

[0178] Step 204, when the first value is less than or equal to the first threshold, exchange at least one first corresponding relationship at the same position in the first parent individual and the second parent individual to obtain a second offspring individual, and mutate the second offspring individual to obtain a first offspring individual.

[0179] In an embodiment of the present application, if the first value is less than or equal to the first threshold, at least one first corresponding relationship at the same position in the first parent individual and the second parent individual can be exchanged first to obtain a second offspring individual, and then the second offspring individual can be mutated to obtain a first offspring individual. When mutating, two of the first corresponding relationships can be selected for exchange, or more than two of the first corresponding relationships can be selected for exchange, or one or more of the values can be changed to other values for mutation.

[0180] For example, if the second offspring individual is d[2] = {2, 5, 4, 3, 1}, the first offspring individual obtained after mutation can be c[2] = {2, 4, 5, 3, 1}, or c[2] = {2, 3, 4, 3, 1}.

[0181] Optionally, step 204 may include the following sub-steps:

[0182] Sub-step 2041, when the first value is less than the second threshold, mutate the second offspring individual based on the Levy flight mutation strategy to obtain a first offspring individual; where the second threshold is less than the first threshold.

[0183] In an embodiment of the present application, Levy flight refers to a random walk in which the probability distribution of the step size is a heavy-tailed distribution, that is, there is a relatively high probability of taking a large step during the random walk. The expression of the Levy flight mutation strategy is as shown in Equation 4 below:

[0184]

[0185] In Equation 4, L(λ) is the step size of the Levy flight strategy, u and v are random variables in independent and identically distributed standard normal distributions, α is the characteristic exponent of the Levy distribution, and its value range is (1, 2), and λ is the step size control parameter of the Levy flight, and its value range interval is [1, 2]. Figure 5 It is a simulation schematic diagram of a Levy flight provided by an embodiment of the present application.

[0186] In an embodiment of the present application, when the first value is less than the second threshold, the second offspring individual can be mutated according to the algorithm shown in Equation 4 to obtain a first offspring individual. Where the second threshold is less than the first threshold.

[0187] Sub-step 2042: When the first value is greater than or equal to the second threshold and less than the first threshold, randomly select multiple first drones from the second offspring individuals, and randomly exchange the positions of the multiple first drones in the second offspring individuals to obtain the first offspring individual.

[0188] In an embodiment of the present application, if the first value is greater than or equal to the second threshold and less than the first threshold, multiple first drones can be randomly selected from the second offspring individuals, and the positions of the multiple first drones in the second offspring individuals can be randomly exchanged, so as to obtain the first offspring individual.

[0189] For example, if the second offspring individual is d[2] = {2, 5, 4, 3, 1}, the drones b5, b4, and b3 in the second, third, and fourth positions can be selected as the first drones, and the positions of b5, b4, and b3 in d[2] can be randomly exchanged to obtain the first offspring individual c[2] = {2, 4, 3, 5, 1} or c[2] = {2, 3, 5, 4, 1}.

[0190] In an embodiment of the present application, when the first value is less than the second threshold, the second offspring individual is mutated based on the Lévy flight mutation strategy to obtain the first offspring individual; wherein the second threshold is less than the first threshold. When the first value is greater than or equal to the second threshold and less than the first threshold, multiple first drones are randomly selected from the second offspring individuals, and the positions of the multiple first drones in the second offspring individuals are randomly exchanged to obtain the first offspring individual. Two different mutation strategies can be adopted according to the size relationship between the randomly generated first value, the second threshold, and the first threshold to mutate the second offspring individual to obtain the first individual, further improving the diversity and richness of the first individual, and further improving the reliability of the finally obtained target individual.

[0191] In an embodiment of the present application, by randomly generating a first value, when the first value is greater than the first threshold, at least one first corresponding relationship at the same position in the first parent individual and the second parent individual is exchanged to obtain the first offspring individual. When the first value is less than or equal to the first threshold, at least one first corresponding relationship at the same position in the first parent individual and the second parent individual is exchanged to obtain the second offspring individual, and the second offspring individual is mutated to obtain the first offspring individual. The first corresponding relationships included in the first parent individual and the second parent individual can be exchanged and mutated according to the size relationship between the randomly generated first value and the first threshold to obtain the first offspring individual, improving the diversity and richness of the first offspring individual, and improving the reliability of the finally obtained target individual.

[0192] Step 205: Based on the weight values corresponding to each point to be detected in the multiple parent individuals and the first offspring individual, and the detection and recognition rate of the drone for the points to be detected, respectively determine the first fitness of the multiple parent individuals and the first offspring individual.

[0193] In the embodiments of the present application, the implementation content of this step can refer to the embodiment content of step 103, which will not be elaborated here.

[0194] Step 206: Determine the target individual corresponding to the maximum value in the first fitness from the first offspring individual and the multiple parent individuals.

[0195] In the embodiments of the present application, the implementation content of this step can refer to the embodiment content of step 104, which will not be elaborated here.

[0196] Optionally, step 206 may include the following sub-steps:

[0197] Sub-step 2061: Add the first offspring individual to the first population, and delete the first individual with the lowest first fitness from the first population to obtain the iterated first population; where the number of the first individuals is the same as the number of the first offspring individuals.

[0198] In the embodiments of the present application, the first offspring individual can be added to the first population, and the first individual with the lowest first fitness can be deleted from the first population, so that the iterated first population can be obtained, and the number of individuals in the first population can be maintained unchanged. When deleting the first individual, the number of the first offspring individuals can be determined first, and the first individual with the same number as the first offspring individuals can be deleted from the first population.

[0199] Sub-step 2062: Re-execute the step of selecting the first parent individual and the second parent individual from the first population.

[0200] In the embodiments of the present application, after the iteration of the first population is completed, the step of "selecting the first parent individual and the second parent individual from the first population" can be re-executed to continue the iteration of the first population.

[0201] Sub-step 2063: In the case that the number of iterations of the iterated first population reaches a preset number, determine the target individual corresponding to the maximum value in the first fitness from the iterated first population.

[0202] In the embodiments of the present application, the number of iterations of the first population can be set as the preset number. In the case that the number of iterations of the iterated first population reaches the preset number, the target individual corresponding to the maximum value of the first fitness can be determined from the iterated first population.

[0203] In an embodiment of the present application, the first population after iteration is obtained by adding the first offspring individuals into the first population and deleting the first individual with the lowest first fitness from the first population; wherein, the number of the first individuals is the same as the number of the first offspring individuals, and the step of selecting the first parent individual and the second parent individual from the first population is re-executed. When the number of iterations of the first population after iteration reaches a preset number, the target individual corresponding to the maximum value in the first fitness is determined from the first population after iteration. The first population can be iterated according to the first offspring individuals, and the target individual can be determined from the first population after iteration when the number of iterations reaches the preset number, further improving the reliability of the target individual.

[0204] Sub-step 2064: Screening to obtain a second individual from the first offspring individuals and the multiple parent individuals; wherein, the drones corresponding to the first correspondence relationships included in the second individual are pairwise different.

[0205] In an embodiment of the present application, multiple second individuals can be first screened from the first offspring individuals and the multiple parent individuals. Among them, the drones corresponding to the first correspondence relationships included in the second individual are pairwise different.

[0206] For example, if there is an individual {1, 2, 3, 3, 4} among the first offspring individuals and the multiple parent individuals, then there is the same drone b4 among the drones corresponding to the first correspondence relationships included in this individual. Therefore, this individual cannot become a second individual; if there is an individual {5, 4, 2, 3, 1}, then the drones corresponding to the first correspondence relationships included in this individual are pairwise different (that is, any two selected drones are different). Therefore, this individual can be used as a second individual.

[0207] Sub-step 2065: Determining the target individual corresponding to the maximum value in the first fitness from the second individual.

[0208] In an embodiment of the present application, after screening to obtain the second individual, the target individual corresponding to the maximum value of the first fitness can be determined among the second individuals.

[0209] In an embodiment of the present application, by screening to obtain a second individual from the first offspring individuals and the multiple parent individuals; wherein, the drones corresponding to the first correspondence relationships included in the first individual are pairwise different, and determining the target individual corresponding to the maximum value in the first fitness from the second individual, it can be ensured that the drones corresponding to the first correspondence relationships included in the finally obtained target individual are pairwise different, improving the usability of the drone allocation method. When detecting a point to be detected by a drone based on the first correspondence relationship included in the target individual, the detection efficiency can be improved, and at the same time, logical errors of the drone during detection can be avoided.

[0210] Step 207: Based on the first correspondence included in the target individual, allocate the drones to the respective points to be detected.

[0211] In the embodiment of the present application, the implementation content of this step can refer to the embodiment content of step 105, which will not be elaborated here.

[0212] Refer to Figure 6 , Figure 6 is a flowchart of an algorithm for a diverse enhanced genetic algorithm provided by an embodiment of the present application. In the figure, the process can start first, and then the first population can be initialized, setting the number of iterations of the population, the number of individuals included in the population, and the data structure of the individuals, etc. After that, the fitness of each individual included in the initialized first population can be calculated, and the individual corresponding to the maximum fitness is determined as the optimal solution. Then, the parent individuals (male parents) can be obtained from the first population, and the positions of the first correspondences included in the parent individuals are exchanged. Among them, the first correspondence refers to the correspondence between the points to be detected and the drones. After the exchange of the first correspondences between the parent individuals is completed, it can be determined whether the current situation meets a preset condition. Among them, the preset condition refers to whether the generated random number is greater than the first threshold. If the generated random number is less than or equal to the first threshold, no mutation is performed, and it is directly determined whether the current number of iterations reaches the maximum number of iterations; if the generated random number is greater than the first threshold, after the exchange of the first correspondences between the parent individuals, the obtained offspring individuals after the exchange can be mutated, and after the mutation, it is determined whether the current number of iterations reaches the maximum number of iterations. If the maximum number of iterations is not reached, the steps of evaluating the chromosome fitness and updating the optimal solution are returned. If the maximum number of iterations is reached, the fitness of all the obtained individuals is calculated, and the target individual with the maximum fitness is output as the optimal solution. After the optimal solution is output, the process can end.

[0213] Refer to Figure 7 , Figure 7 is a logic block diagram of a drone target allocation device based on a diverse enhanced genetic algorithm provided by an embodiment of the present application. The drone target allocation device 700 based on the diverse enhanced genetic algorithm may include:

[0214] A selection module 701, configured to select a first male parent individual and a second male parent individual from the first population; wherein, the first population includes a plurality of male parent individuals; wherein, each male parent individual includes a first correspondence between a plurality of points to be detected and a plurality of drones, and the first correspondences included in different male parent individuals are different;

[0215] An exchange module 702, configured to exchange at least one same-position first correspondence between the first male parent individual and the second male parent individual to obtain a first offspring individual;

[0216] The first determination module 703 is configured to respectively determine the first fitness of the multiple parental individuals and the first offspring individual based on the weight values corresponding to the respective detection points to be detected in the multiple parental individuals and the first offspring individual and the detection and recognition rate of the UAV for the detection points to be detected.

[0217] The second determination module 704 is configured to determine a target individual corresponding to the maximum value in the first fitness from the first offspring individual and the multiple parental individuals.

[0218] The allocation module 705 is configured to allocate the UAVs to the respective detection points based on the first correspondence relationship included in the target individual.

[0219] Optionally, the selection module 701 includes:

[0220] The first generation sub-module is configured to generate a global information matrix of the first population based on each parental individual included in the first population and the second fitness corresponding to each parental individual; wherein, the rows of the global information matrix correspond to the UAVs one by one, the columns of the global information matrix correspond to the detection points to be detected one by one, and the first value at each position in the global information matrix is determined based on the second fitness corresponding to each parental individual.

[0221] The first determination sub-module is configured to determine the first maximum value of the first values included in each column of the global information matrix from the global information matrix.

[0222] The second determination sub-module is configured to determine a second correspondence relationship between the multiple detection points and the multiple UAVs based on the row number and column number of the first maximum value in the global information matrix.

[0223] The second generation sub-module is configured to generate a first parental individual based on the second correspondence relationship.

[0224] The first selection sub-module is configured to randomly select a second parental individual from the first population.

[0225] Optionally, the second determination sub-module includes:

[0226] The first determination unit is configured to, in the case where there are second maximum values with the same corresponding column numbers in the first maximum values, determine a third correspondence relationship between the multiple detection points and the multiple UAVs based on the row numbers and column numbers of the remaining maximum values except the second maximum values in the first maximum values.

[0227] A second determination unit, configured to determine a third maximum value of the first numerical value from the remaining rows and remaining columns of the global information matrix except for the row number and column number corresponding to the remaining maximum value;

[0228] A third determination unit, configured to determine a fourth correspondence between the multiple points to be detected and the multiple unmanned aerial vehicles based on the row number and column number of the third maximum value in the global information matrix;

[0229] An integration unit, configured to integrate the third correspondence and the fourth correspondence to obtain a second correspondence between the multiple points to be detected and the multiple unmanned aerial vehicles.

[0230] Optionally, the selection module 701 includes:

[0231] A third generation sub-module, configured to generate a first data sequence based on a preset chaotic mapping;

[0232] A second selection sub-module, configured to arbitrarily select a total number of points to be detected of second data from the first data included in the first data sequence;

[0233] A first sorting sub-module, configured to sort the first order of the second data in the first data sequence in ascending order to obtain a first order sequence; wherein, the numerical value in the first order sequence represents the first order of the second data in the first data sequence;

[0234] A second sorting sub-module, configured to sort the second data in descending order to obtain a second data sequence of the second data;

[0235] An adjustment sub-module, configured to adjust each numerical value in the first order sequence to the same position of the second data in the second data sequence respectively based on the correspondence between the second data and each numerical value in the first order sequence, to obtain a second order sequence;

[0236] A third selection sub-module, configured to select a first parental individual corresponding to a first correspondence that matches the second correspondence included in the second order sequence from a first population;

[0237] A fourth selection sub-module, configured to randomly select a second parental individual different from the first parental individual from the first population.

[0238] Optionally, the exchange module 702 includes:

[0239] A second generation sub-module, configured to randomly generate a first numerical value;

[0240] The first exchange sub-module is used to exchange at least one first corresponding relationship at the same position between the first parent individual and the second parent individual to obtain a first offspring individual when the first value is greater than the first threshold;

[0241] The exchange mutation sub-module is used to exchange at least one first corresponding relationship at the same position between the first parent individual and the second parent individual to obtain a second offspring individual when the first value is less than or equal to the first threshold, and mutate the second offspring individual to obtain a first offspring individual.

[0242] Optionally, the exchange mutation sub-module includes:

[0243] The first mutation unit is used to mutate the second offspring individual based on the Lévy flight mutation strategy to obtain a first offspring individual when the first value is less than the second threshold; wherein, the second threshold is less than the first threshold;

[0244] The second mutation unit is used to randomly select a plurality of first drones from the second offspring individual and randomly exchange the positions of the plurality of first drones in the second offspring individual to obtain a first offspring individual when the first value is greater than or equal to the second threshold and less than the first threshold.

[0245] Optionally, the second determination module 704 includes:

[0246] The iteration sub-module is used to add the first offspring individual into the first population and delete the first individual with the lowest first fitness from the first population to obtain the iterated first population; wherein, the number of the first individuals is the same as the number of the first offspring individuals;

[0247] The step execution sub-module is used to re-execute the step of selecting the first parent individual and the second parent individual from the first population;

[0248] The first determination sub-module is used to determine the target individual corresponding to the maximum value in the first fitness from the iterated first population when the number of iterations of the iterated first population reaches the preset number of iterations.

[0249] Optionally, the second determination module 704 includes:

[0250] The screening sub-module is used to screen and obtain a second individual from the first offspring individual and the plurality of parent individuals; wherein, the drones corresponding to the first corresponding relationships included in the second individual are pairwise different;

[0251] A second determination sub-module, configured to determine a target individual corresponding to the maximum value in the first fitness from the second population of individuals.

[0252] The UAV allocation device based on enhanced genetics in the embodiments of the present application may be an electronic device or a component in an electronic device, such as an integrated circuit or a chip. The electronic device may be a terminal or other devices other than terminals. Exemplarily, the electronic device may be a GPU BOX, a mobile phone, a tablet computer, a laptop computer, a handheld computer, a vehicle-mounted electronic device, a Mobile Internet Device (MID), an augmented reality (AR) / virtual reality (VR) device, a robot, a wearable device, an ultra-mobile personal computer (UMPC), a netbook, or a personal digital assistant (PDA), etc. It may also be a server, a Network Attached Storage (NAS), a personal computer (PC), a television (TV), a teller machine, or a self-service machine, etc. The embodiments of the present application do not make specific limitations.

[0253] The UAV allocation device based on enhanced genetics in the embodiments of the present application may be a device with an operating system. The operating system may be an Android operating system, a Linux, a Windows operating system, etc., or other possible operating systems. The embodiments of the present application do not make specific limitations.

[0254] The UAV allocation device based on enhanced genetics provided in the embodiments of the present application can implement Figures 1 to 3 each process implemented by the method embodiments. To avoid repetition, it will not be elaborated here.

[0255] The embodiments of the present application provide an electronic device. Refer to Figure 8 , the electronic device 80 includes: a processor 801, a memory 802, and a computer program 8021 stored on the memory 802 and executable on the processor 801. When the processor 801 executes the program, it implements the UAV target allocation method based on the diverse enhanced genetic algorithm in the foregoing embodiments.

[0256] The embodiments of the present application also provide a computer-readable storage medium, on which computer programs / instructions are stored. When the computer programs / instructions are executed by a processor, the steps in the UAV target allocation method based on the diverse enhanced genetic algorithm disclosed in the embodiments of the present application are implemented.

[0257] The embodiments of the present application also provide a computer program product. When the computer program product runs on an electronic device, the steps in the UAV target allocation method based on the diverse enhanced genetic algorithm disclosed in the embodiments of the present application are implemented when the processor executes.

[0258] The embodiments in this specification are all described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other.

[0259] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of methods, devices, electronic devices, and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the processes and / or blocks in the flowchart and / or block diagram can also be implemented. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal devices generate a device for implementing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or multiple blocks.

[0260] These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing terminal device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured product including an instruction device, and the instruction device implements the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or multiple blocks.

[0261] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable terminal device provide steps for implementing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or multiple blocks.

[0262] Although the preferred embodiments of the embodiments of the present application have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concepts. Therefore, the appended claims are intended to be construed as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present application.

[0263] Finally, it should also be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or terminal device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or terminal device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, article or terminal device comprising the element.

[0264] The above has introduced in detail a resource decoupling system, execution method and device for deep learning applications provided by the present application. Specific examples are used in this text to elaborate on the principle and implementation manner of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A UAV target assignment method based on a diverse enhanced genetic algorithm, characterized in that, The method includes: Select a first parent individual and a second parent individual from a first population; wherein, the first population includes multiple parent individuals; wherein, each parent individual includes a first correspondence relationship between multiple detection points to be detected and multiple unmanned aerial vehicles, and the first correspondence relationships included in different parent individuals are different; Exchange the first correspondence relationships at at least one same position in the first parent individual and the second parent individual to obtain a first offspring individual; Based on the weight values respectively corresponding to each detection point to be detected in the multiple parent individuals and the first offspring individual and the detection and recognition rate of the unmanned aerial vehicle for the detection point to be detected, respectively determine the first fitness of the multiple parent individuals and the first offspring individual; Determine a target individual corresponding to the maximum value in the first fitness from the first offspring individual and the multiple parent individuals; Based on the first correspondence relationship included in the target individual, allocate the unmanned aerial vehicles to each detection point to be detected; The selecting the first parent individual and the second parent individual from the first population includes: Generate a global information matrix of the first population based on each parent individual included in the first population and the second fitness respectively corresponding to each parent individual; wherein, the rows of the global information matrix correspond to the unmanned aerial vehicles one by one, the columns of the global information matrix correspond to the detection points to be detected one by one, and the first value at each position in the global information matrix is determined based on the second fitness respectively corresponding to each parent individual; wherein, the calculation method of the second fitness is the same as the calculation method of the first fitness; Determine the first maximum value of the first values included in each column of the global information matrix from the global information matrix; Based on the row number and column number of the first maximum value in the global information matrix, determine a second correspondence relationship between the multiple detection points to be detected and the multiple unmanned aerial vehicles; Generate a first parent individual based on the second correspondence relationship; Randomly select a second parent individual from the first population.

2. A method for unmanned aerial vehicle target assignment based on a diverse enhanced genetic algorithm, characterized in that The method includes: Select a first parent individual and a second parent individual from a first population; wherein, the first population includes multiple parent individuals; wherein, each parent individual includes a first correspondence relationship between multiple detection points to be detected and multiple unmanned aerial vehicles, and the first correspondence relationships included in different parent individuals are different; Exchange the first correspondence relationships at at least one same position in the first parent individual and the second parent individual to obtain a first offspring individual; Based on the weight values respectively corresponding to each detection point to be detected in the multiple parent individuals and the first offspring individual and the detection and recognition rate of the unmanned aerial vehicle for the detection point to be detected, respectively determine the first fitness of the multiple parent individuals and the first offspring individual; Determine a target individual corresponding to the maximum value in the first fitness from the first offspring individual and the multiple parent individuals; Based on the first correspondence relationship included in the target individual, allocate the unmanned aerial vehicles to each detection point to be detected; The selecting the first parent individual and the second parent individual from the first population includes: Generate a first data sequence based on a preset chaotic mapping; Arbitrarily select a total number of second data points to be detected from the first data included in the first data sequence; Arrange the first order of the second data in the first data sequence in ascending order to obtain a first order sequence; wherein, the values in the first order sequence represent the first order of the second data in the first data sequence; Arrange the second data in descending order to obtain a second data sequence of the second data; Based on the correspondence between the second data and each value in the first order sequence, adjust each value in the first order sequence to the same position of the second data in the second data sequence to obtain a second order sequence; Select a first parental individual corresponding to the first correspondence that matches the second correspondence included in the second order sequence from the first population; Randomly select a second parental individual different from the first parental individual from the first population.

3. The method according to claim 1, wherein The determining the second correspondence between the multiple points to be detected and the multiple UAVs based on the number of rows and columns of the first maximum value in the global information matrix includes: In the case where there is a second maximum value with the same number of corresponding columns in the first maximum value, based on the number of rows and columns of the remaining maximum values except the second maximum value in the first maximum value in the global information matrix, determine a third correspondence between the multiple points to be detected and the multiple UAVs; Determine a third maximum value of the first value from the remaining rows and remaining columns in the global information matrix except for the rows and columns corresponding to the remaining maximum values; Based on the number of rows and columns of the third maximum value in the global information matrix, determine a fourth correspondence between the multiple points to be detected and the multiple UAVs; Integrate the third correspondence and the fourth correspondence to obtain the second correspondence between the multiple points to be detected and the multiple UAVs.

4. The method according to claim 1 or 2, characterized in that, The obtaining a first offspring individual by swapping at least one first correspondence at the same position between the first parental individual and the second parental individual includes: Randomly generate a first value; In the case where the first value is greater than a first threshold, swap at least one first correspondence at the same position between the first parental individual and the second parental individual to obtain a first offspring individual; In the case where the first value is less than or equal to the first threshold, swap at least one first correspondence at the same position between the first parental individual and the second parental individual to obtain a second offspring individual, and mutate the second offspring individual to obtain a first offspring individual.

5. The method according to claim 4, characterized in that, The mutating the second offspring individual to obtain a first offspring individual includes: In the case where the first value is less than a second threshold, mutate the second offspring individual based on a Lévy flight mutation strategy to obtain a first offspring individual; wherein, the second threshold is less than the first threshold; When the first value is greater than or equal to the second threshold and less than the first threshold, randomly select multiple first drones from the second offspring individuals, and randomly exchange the positions of the multiple first drones in the second offspring individuals to obtain a first offspring individual.

6. The method according to claim 1 or 2, characterized in that The determining the target individual corresponding to the maximum value in the first fitness from the first offspring individual and the multiple parent individuals includes: Adding the first offspring individual into the first population, and deleting the individual with the lowest first fitness from the first population to obtain the iterated first population; wherein, the number of the individuals is the same as the number of the first offspring individuals; Re-executing the step of selecting a first parent individual and a second parent individual from the first population; When the number of iterations of the iterated first population reaches a preset number, determining the target individual corresponding to the maximum value in the first fitness from the iterated first population.

7. The method according to claim 1 or 2, characterized in that, The determining the target individual corresponding to the maximum value in the first fitness from the first offspring individual and the multiple parent individuals includes: Screening to obtain a second individual from the first offspring individual and the multiple parent individuals; wherein, the drones corresponding to the first corresponding relationships included in the second individual are pairwise different; Determining the target individual corresponding to the maximum value in the first fitness from the second individual.

8. An unmanned aerial vehicle target allocation device based on a diverse enhanced genetic algorithm, characterized in that, The device includes: A selection module, configured to select a first parent individual and a second parent individual from a first population; wherein, the first population includes multiple parent individuals; wherein, each parent individual includes a first corresponding relationship between multiple points to be detected and multiple drones, and the first corresponding relationships included in different parent individuals are different; An exchange module, configured to exchange at least one first corresponding relationship at the same position in the first parent individual and the second parent individual to obtain a first offspring individual; A first determination module, configured to respectively determine the first fitness of the multiple parent individuals and the first offspring individual based on the weight values respectively corresponding to each point to be detected in the multiple parent individuals and the first offspring individual and the detection and recognition rate of the drones for the points to be detected; A second determination module, configured to determine the target individual corresponding to the maximum value in the first fitness from the first offspring individual and the multiple parent individuals; An allocation module, configured to allocate the drones to each point to be detected based on the first corresponding relationship included in the target individual; Wherein, the selecting a first parent individual and a second parent individual from the first population includes: Generating a global information matrix of the first population based on each parent individual included in the first population and the second fitness respectively corresponding to each parent individual; wherein, the rows of the global information matrix correspond to the drones one by one, the columns of the global information matrix correspond to the points to be detected one by one, and the first value at each position in the global information matrix is determined based on the second fitness respectively corresponding to each parent individual; wherein, the calculation method of the second fitness is the same as the calculation method of the first fitness; Determine a first maximum value of the first numerical values included in each column of the global information matrix from the global information matrix; Based on the number of rows and columns of the first maximum value in the global information matrix, determine a second correspondence between the multiple points to be detected and the multiple drones; Generate a first parental individual based on the second correspondence; Randomly select a second parental individual from the first population.

9. An unmanned aerial vehicle target allocation device based on a diverse enhanced genetic algorithm, characterized in that, The device includes: A selection module for selecting a first parental individual and a second parental individual from a first population; wherein, the first population includes multiple parental individuals; wherein, each parental individual includes a first correspondence between multiple points to be detected and multiple drones, and the first correspondences included in different parental individuals are different; An exchange module for exchanging the first correspondences at at least one same position in the first parental individual and the second parental individual to obtain a first offspring individual; A first determination module for respectively determining first fitness values of the multiple parental individuals and the first offspring individual based on the weight values respectively corresponding to each point to be detected in the multiple parental individuals and the first offspring individual and the detection and recognition rate of the drone for the point to be detected; A second determination module for determining a target individual corresponding to the maximum value in the first fitness values from the first offspring individual and the multiple parental individuals; An allocation module for allocating the drones to the respective points to be detected based on the first correspondence included in the target individual; Wherein, the selecting the first parental individual and the second parental individual from the first population includes: Generating a first data sequence based on a preset chaotic mapping; Arbitrarily selecting a total number of points to be detected of second data from the first data included in the first data sequence; Arranging the first order of the second data in the first data sequence in ascending order to obtain a first order sequence; wherein, the values in the first order sequence represent the first order of the second data in the first data sequence; Arranging the second data in descending order to obtain a second data sequence of the second data; Based on the correspondence between the second data and the respective values in the first order sequence, respectively adjust the respective values in the first order sequence to the same positions of the second data in the second data sequence to obtain a second order sequence; Select a first parental individual corresponding to the first correspondence that matches the second correspondence included in the second order sequence from the first population; Randomly select a second parental individual different from the first parental individual from the first population.

10. An electronic device, comprising a memory, a processor, and a computer program stored on the memory, characterized in that, The processor executes the computer program to implement the drone target allocation method based on the diverse enhanced genetic algorithm as described in any one of claims 1-7.

Images