Sensor network target distribution method and device

Abstract

The invention provides a sensor network target distribution method and device, and the method comprises the steps: obtaining the threat weight of a monitoring target according to the initial weight of the monitoring target in a sensor network; obtaining a sensor target monitoring efficiency function based on a sensor target sensing probability model and the threat weight of the monitoring target; and based on the sensor target monitoring efficiency function and the sensor utilization rate function, taking a randomly generated target allocation scheme as a giant panda initial population, obtaining a sensor network target allocation optimization result through a giant panda optimization algorithm, and performing target allocation on the sensor network based on the sensor network target allocation optimization result. Therefore, the multi-target distribution requirement of an automatic control device for the sensor network is effectively met, the monitoring precision can be improved, the sensor utilization efficiency is improved, and the energy consumption of the sensor network can be reduced.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to a sensor network target allocation method and device. Background Art

[0002] Wireless sensor networks are used in industrial automation to monitor and control production processes. They can be used to monitor equipment status, temperature, pressure, vibration and other parameters, and perform remote monitoring and fault diagnosis, which can provide the necessary control mechanism for automatic control. However, since sensor nodes are in a distributed environment, communication between sensor nodes may face security risks, resulting in target allocation errors or information transmission interruption, which in turn affects the performance of the entire automatic control device. According to statistics, a series of failures between sensor nodes have caused more than 90% of resource waste. Therefore, optimizing the target allocation of sensor networks has become a key factor in improving the utilization efficiency of sensor networks and ensuring device automation.

[0003] The main feature of sensor networks is that they have a distributed perception and centralized control structure. Each sensor node is installed in a different location and is responsible for measuring the intensity of the area where it is located. These nodes transmit the collected data to the central control node or cloud platform through wired or wireless communication for centralized management and analysis. The advantage of sensor networks is that they can provide more comprehensive and accurate information and achieve more refined control. Through the collaborative work of multiple sensor nodes, a comprehensive perception of the situation in the entire area can be achieved, including changes in different locations, directions and time. This helps the sensor network better understand the distribution characteristics of the environment and accurately perceive changes in intensity. The target allocation problem of sensor networks usually involves multiple targets. How to improve the utilization efficiency of sensors and reduce the energy consumption of sensor networks are all issues that need to be considered. For example, greedy algorithms and Hungarian algorithms can only solve single-target allocation problems and can no longer meet the needs of automatic control devices to solve multi-target allocation problems in sensor networks. Summary of the invention

[0004] The present invention provides a sensor network target allocation method and device, which are used to solve the defect that only single target allocation problem can be solved in the prior art, and realize the multi-target allocation requirement of the sensor network solved by the automatic control device.

[0005] The present invention provides a sensor network target allocation method, comprising the following steps: According to the initial weight of the monitoring target in the sensor network, the threat weight of the monitoring target is obtained; Based on the sensor target perception probability model and the threat weight of the monitoring target, a sensor target monitoring effectiveness function is obtained; Based on the sensor target monitoring effectiveness function and the sensor utilization rate function, a randomly generated target allocation scheme is used as the giant panda initial population, a sensor network target allocation optimization result is obtained through a giant panda optimization algorithm, and a target allocation is performed on the sensor network based on the sensor network target allocation optimization result.

[0006] According to a sensor network target allocation method provided by the present invention, a sensor network target allocation optimization result is obtained by using a giant panda optimization algorithm, including: Calculating the rank of each allocation scheme in the initial population of giant pandas based on the sensor target monitoring effectiveness function and the sensor utilization rate function; Performing non-inferiority stratification on each allocation scheme based on the rank of each allocation scheme, and calculating the fitness value of each allocation scheme; When the iteration termination condition is met, all the first non-inferior layers of the initial population of giant pandas are deconstructed into a Pareto optimal solution set and used as the sensor network target allocation optimization result; when the iteration termination condition is not met, the initial population of giant pandas is updated, the rank of each allocation scheme in the updated giant panda population is calculated, each allocation scheme in the updated giant panda population is non-inferior stratified, and the fitness value of each allocation scheme in the updated giant panda population is calculated, and the steps of updating the giant panda population, calculating the rank of each allocation scheme in the updated giant panda population, non-inferior stratified each allocation scheme in the updated giant panda population, and calculating the fitness value of each allocation scheme in the updated giant panda population are repeated until the iteration termination condition is met, and all the first non-inferior layers of the updated giant panda population are deconstructed into a Pareto optimal solution set and used as the sensor network target allocation optimization result.

[0007] According to a sensor network target allocation method provided by the present invention, the initial giant panda population and the updated giant panda population are updated according to the following method: Randomly select any two individuals in the current giant panda population, perform differential mutation operations, and generate new mutant individuals; Comparing the fitness of the variant individual with the arbitrary two individuals, retaining the individual with higher fitness, and obtaining an offspring giant panda population; The current giant panda population and the offspring giant panda population are merged, and an updated giant panda population having the same size as the initial giant panda population is obtained by screening in a Pareto-dominant manner.

[0008] According to a sensor network target allocation method provided by the present invention, individuals of the initial population of giant pandas are generated according to the following method: In the arbitrary target allocation scheme, for any monitoring target, determining the allocation of the arbitrary monitoring target to each sensor; Based on the order of all monitoring targets, the allocation status of each monitoring target is connected in series through binary coding to form individuals in the initial population of giant pandas.

[0009] According to a sensor network target allocation method provided by the present invention, the sensor target monitoring effectiveness function is determined according to the following formula:

[0010] in, Represents the sensor target monitoring effectiveness value of the entire sensor network; Indicates the number of monitoring targets; Indicates the number of sensors; Indicates monitoring target Threat weight value of Indicates sensor For monitoring targets The perceived probability of The value is 1 or 0, indicating that the sensor Is it assigned to the monitoring target? .

[0011] According to a sensor network target allocation method provided by the present invention, the threat weight of the monitoring target is determined according to the following formula:

[0012] in, Indicates monitoring target Threat weight value of Indicates monitoring target Distance to the sensor; Indicates monitoring target speed; Indicates monitoring target The initial weight value of , and They are the weight coefficients for adjusting the influence of distance, speed and initial weight respectively.

[0013] The present invention also provides a sensor network target allocation device, comprising the following modules: A weight module is used to obtain the threat weight of the monitoring target according to the initial weight of the monitoring target in the sensor network; An effectiveness module, used to obtain a sensor target monitoring effectiveness function based on a sensor target perception probability model and a threat weight of the monitoring target; The optimization module is used to obtain the sensor network target allocation optimization result through the giant panda optimization algorithm based on the sensor target monitoring effectiveness function and the sensor utilization rate function, and to perform target allocation on the sensor network based on the sensor network target allocation optimization result, using the randomly generated target allocation scheme as the giant panda initial population.

[0014] The present invention also provides an electronic device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the above-mentioned sensor network target allocation methods is implemented.

[0015] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the sensor network target allocation method described in any one of the above methods is implemented.

[0016] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the sensor network target allocation method described above is implemented.

[0017] The sensor network target allocation method and device provided by the present invention meet the monitoring accuracy requirements through the threat weights of the monitoring targets in the sensor network, increase the sensor utilization function, save sensor resources, and thus realize reasonable and effective resource allocation. The giant panda optimization algorithm is also used until the sensor network target allocation optimization result is obtained to realize the target allocation of the sensor network, effectively meeting the needs of the automatic control device to solve the multi-target allocation of the sensor network, which can not only improve the monitoring accuracy and sensor utilization efficiency, but also reduce the energy consumption of the sensor network. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 It is a flow chart of the sensor network target allocation method provided by the present invention.

[0020] Figure 2 It is a flow chart of an embodiment of the sensor network target allocation method provided by the present invention.

[0021] Figure 3 It is a sub-process diagram of step S310 provided by the present invention.

[0022] Figure 4It is a schematic diagram of the coding of the sensor network target allocation scheme provided by the present invention.

[0023] Figure 5 It is a Pareto solution set diagram of the sensor target allocation scheme provided by the present invention.

[0024] Figure 6 It is a structural schematic diagram of the sensor network target allocation device provided by the present invention.

[0025] Figure 7 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] Figure 1 FIG. 1 is a flow chart of the sensor network target allocation method provided by the present invention, such as Figure 1 As shown, the method comprises the following steps: Step 100: Obtain the threat weight of the monitoring target according to the initial weight of the monitoring target in the sensor network.

[0028] Specifically, the initial weight values ​​and multiple characteristic parameters (such as distance and speed) of each monitoring target in the sensor network can be collected first. The initial weight value refers to the basic importance given to a target without considering other factors (such as distance and speed). This weight value is usually pre-set according to the characteristics of the target itself or application requirements, and is used to indicate the priority or threat level of the target in the entire system. Each characteristic parameter can be evaluated to determine their importance to each monitoring target.

[0029] In some implementations, in order to calculate the threat weight, the obtained target importance (initial weight value and multiple feature parameters) may be normalized.

[0030] For example, n targets are extracted from the sensor network to be monitored, and characteristic parameters are extracted from these targets. For each target, common fuzzy logic methods or neural network methods can be used to estimate the importance of the target, which can be expressed as (i=1,2,...,n).

[0031] Optionally, the threat weight of the monitoring target can be determined according to the following formula:

[0032] in, Indicates monitoring target Threat weight value of Indicates monitoring target Distance to the sensor; Indicates monitoring target speed; Indicates monitoring target The initial weight value of , and They are the weight coefficients for adjusting the influence of distance, speed and initial weight respectively.

[0033] Step 101: Based on the sensor target perception probability model and the threat weight of the monitoring target, a sensor target monitoring effectiveness function is obtained.

[0034] Specifically, the sensor target perception probability model is used to describe the probability that a sensor can effectively monitor a target. In the target allocation problem of sensor networks, the matching degree between each sensor and the target can be evaluated. Indicates sensor For monitoring targets The perceived probability.

[0035] The sensor target monitoring effectiveness function can be obtained based on the sensor target perception probability model and the threat weight of the monitored target.

[0036] Optionally, the sensor target monitoring effectiveness function is determined according to the following formula:

[0037] in, Represents the sensor target monitoring effectiveness value of the entire sensor network; Indicates the number of monitoring targets; Indicates the number of sensors; Indicates monitoring target Threat weight value of Indicates sensor For monitoring targets The perceived probability of The value is 1 or 0, indicating that the sensor Is it assigned to the monitoring target? .

[0038] Step 102: Based on the sensor target monitoring effectiveness function and the sensor utilization rate function, a randomly generated target allocation scheme is used as the giant panda initial population, a sensor network target allocation optimization result is obtained through a giant panda optimization algorithm, and a target allocation is performed on the sensor network based on the sensor network target allocation optimization result.

[0039] Specifically, a sensor usage function may be generated according to the usage of the sensor. The sensor usage function may be determined according to the following formula:

[0040] in, Indicates the sensor usage rate, Indicates the number of batches of monitoring targets assigned to each sensor, express The maximum number of monitoring target batches per sensor.

[0041] The number of batches of monitoring targets assigned to each sensor It can be determined according to the following formula:

[0042] in, The value is 1 or 0, indicating that the sensor Is it assigned to the monitoring target? .

[0043] In order to perform multi-objective optimization, the sensor target monitoring effectiveness function and the sensor utilization rate function can be standardized to obtain a constrained multi-objective optimization model.

[0044] The constraint objective function of the constrained multi-objective optimization model can be determined according to the following formula:

[0045] in, and They represent the standardized monitoring effectiveness function and sensor utilization function respectively.

[0046] After obtaining the constrained multi-objective optimization model, the randomly generated target allocation scheme can be used as the initial population of giant pandas, and the giant panda optimization algorithm (POA) can be used to perform optimization search to obtain the sensor network target allocation optimization result.

[0047] The Giant Panda Optimization Algorithm is a relatively new metaheuristic algorithm inspired by the behavior and ecological characteristics of giant pandas. Its basic idea is to regard the problem space as an ecosystem containing multiple giant panda individuals. Each giant panda individual represents a potential solution and has a certain fitness evaluation value. The algorithm performs optimization search by simulating the giant panda's behavior of finding food and habitat.

[0048] During the iteration of the algorithm, individual giant pandas move and interact based on their current fitness evaluation values ​​and neighborhood information. Giant pandas can improve their own solutions through information exchange and learning, and gradually find better solutions. The algorithm continuously updates and adjusts the position and fitness evaluation values ​​of individuals to achieve the optimal solution to the optimization problem. Compared with Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), and Genetic Algorithm (GA), POA has better adaptability when dealing with dynamic environments.

[0049] Optionally, the target allocation scheme of the sensor network can be binary encoded, and a set of initial allocation schemes can be randomly generated to form an initial population of giant pandas.

[0050] Among them, individuals of the initial population of giant pandas can be generated in the following ways: In the arbitrary target allocation scheme, for any monitoring target, the allocation of the arbitrary monitoring target to each sensor is determined; Based on the order of all monitoring targets, the allocation status of each monitoring target is linked together through binary coding to form the individuals in the initial population of giant pandas.

[0051] Specifically, a binary number may be used to represent the allocation status of each target, wherein each bit represents a sensor, a value of 1 on the bit indicates that the sensor is allocated to the target, and a value of 0 indicates that the sensor is not allocated.

[0052] In a certain target allocation scheme, for any monitoring target, the allocation of the monitoring target to each sensor can be determined, thereby obtaining the allocation of each detection target to each sensor. Then, according to the order of all monitoring targets, the allocation of each monitoring target can be connected in series through binary coding to form individuals in the initial population of giant pandas. Furthermore, the initial population of giant pandas formed by all target allocation schemes can be integrated.

[0053] In some implementations, the specific encoding scheme may be as follows: Assume that there are n targets to be assigned and m available sensor nodes. Use binary 0-1 encoding to represent the assignment scheme of targets and sensors. Each assignment scheme consists of sensor node numbers arranged in target order, where each small segment represents the assignment result of a batch of targets. The length of the encoding scheme is n*m.

[0054] For example, n is 5, m is 3, the target number is: 12345, the sensor number is: ABC, and a coding scheme for the population is 011100001110101. The above coding scheme represents a target allocation scheme, in which target 1 is monitored by sensor B and sensor C, target 2 is monitored by sensor A, target 3 is only monitored by sensor C, target 4 is monitored by sensor A and sensor B, and target 5 is monitored by sensor A and sensor C.

[0055] For sensor network target allocation based on the giant panda optimization algorithm, the individual characteristics of the initial population will have an important impact on the optimization effect of the subsequent algorithm. The initial population with a good diversity of allocation schemes has a uniform distribution of individuals and a high degree of discreteness, which helps to improve the convergence speed and solution accuracy of the algorithm, reduce the probability of the algorithm falling into local extreme values, and effectively alleviate the premature phenomenon of the algorithm.

[0056] In an embodiment of the present invention, an initial population much larger than the required group size can be generated, and allocation scheme individuals suitable for the required size can be randomly selected therefrom. Then, the selected initial population is evaluated and the best allocation scheme individual is selected, and then it is continuously iterated until the sensor network target allocation optimization result is determined.

[0057] Preferably, the initial population size of giant pandas is between 100 and 200.

[0058] After obtaining the sensor network target allocation optimization result through the giant panda optimization algorithm, the sensor network can be targeted according to the sensor network target allocation optimization result.

[0059] The sensor network target allocation method provided by the present invention meets the monitoring accuracy requirement through the threat weight of the monitoring target in the sensor network, increases the sensor utilization function, saves sensor resources, and thus realizes reasonable and effective resource allocation. It also adopts the giant panda optimization algorithm until the sensor network target allocation optimization result is obtained to realize the target allocation of the sensor network, effectively meets the automatic control device to solve the multi-target allocation demand of the sensor network, can not only improve the monitoring accuracy and sensor utilization efficiency, but also reduce the energy consumption of the sensor network.

[0060] According to a sensor network target allocation method provided by the present invention, a sensor network target allocation optimization result is obtained by using a giant panda optimization algorithm, including: Based on the sensor target monitoring effectiveness function and the sensor utilization rate function, the rank of each allocation scheme in the initial population of giant pandas is calculated; Perform non-inferiority stratification on each allocation scheme based on its rank, and calculate the fitness value of each allocation scheme; When the iteration termination condition is met, all the first non-inferior layers of the initial population of giant pandas are deconstructed into a Pareto optimal solution set and used as the target allocation optimization result of the sensor network; when the iteration termination condition is not met, the initial population of giant pandas is updated, the rank of each allocation scheme in the updated giant panda population is calculated, each allocation scheme in the updated giant panda population is non-inferior stratified, and the fitness value of each allocation scheme in the updated giant panda population is calculated. The steps of updating the giant panda population, calculating the rank of each allocation scheme in the updated giant panda population, non-inferior stratified each allocation scheme in the updated giant panda population, and calculating the fitness value of each allocation scheme in the updated giant panda population are repeated until the iteration termination condition is met, and all the first non-inferior layers of the updated giant panda population are deconstructed into a Pareto optimal solution set and used as the target allocation optimization result of the sensor network.

[0061] Specifically, in the process of iterating and screening the initial population using the giant panda optimization algorithm, the sensor target monitoring effectiveness function value and sensor utilization function value of each allocation scheme can be calculated according to the constrained multi-objective optimization model. Using these values, a coordinate system can be established to calculate the rank of each allocation scheme, and non-inferiority stratification and the fitness of the allocation scheme can be calculated according to the size of the rank.

[0062] For multi-objective optimization problems, the Pareto optimization principle is considered when updating the giant panda population. The non-dominated sorting method is used to divide the population into different levels according to the performance of each individual. The first non-dominated level contains all the optimal solutions.

[0063] For individuals in the non-inferior layer, if they have advantages in multiple objectives, they will be retained in the next generation population first. Therefore, the parent allocation scheme population of the first (optimal) non-inferior layer, that is, the Pareto front solution set, can be recorded.

[0064] For example, based on the execution results of the giant panda optimization algorithm, the following hierarchical structure and rank values were obtained in the plane of the constrained multi-objective optimization model: The first non-dominated layer (Pareto front): This layer consists of five individual allocation schemes, denoted as individuals A1, A2, A3, A4, and A5. These individuals show optimal solutions in terms of both coverage and energy consumption, and no other individuals can completely dominate them. Their rank values are all 1. The second non-dominated layer: This layer consists of six individual allocation schemes, denoted as individuals B1, B2, B3, B4, B5, and B6. These individuals still show good performance in terms of both coverage and energy consumption. Each individual may be slightly lacking in one objective, but achieves better results in the other objective. Their rank values are all 2. The third non-dominated layer: This layer consists of four individual allocation schemes, denoted as individuals C1, C2, C3, and C4. These individuals show relatively balanced performance in terms of both coverage and energy consumption. Each individual has certain advantages, but may be slightly disadvantaged in one objective. Their rank values are all 3.

[0065] The non-dominated hierarchical division of the Pareto set can effectively perform non-dominated hierarchical division on all individual allocation schemes in the solution space to distinguish their superiority and inferiority relationships and provide a basis for selecting a set of solutions with advantages.

[0066] Then, it can be judged whether the initial population of the giant panda satisfies the iteration termination condition.

[0067] If the termination condition is satisfied, all the solutions in the first non-dominated layer of the initial population of the giant panda constitute the Pareto optimal solution set, that is, the optimization result of the target allocation of the sensor network.

[0068] Suppose there is a multi-objective optimization problem with the number of objective functions being m. For any two solutions x and y, a domination relationship can be defined to compare their objective function values: For all i ∈ {1, 2,..., m}, there is ≤ , that is, each objective function value of solution x is less than or equal to the corresponding objective function value of solution y; There is at least one j ∈ {1, 2,..., m} such that < , that is, the value of solution x in at least one objective function is strictly less than the value of solution y in the corresponding objective function; According to the definition of the domination relationship, the Pareto optimal solution set can be strictly defined as the set of all solutions that are not dominated by other solutions: Pareto optimal solution set = {x|∀y ∈ solution space, if y < x, then y does not belong to the Pareto optimal solution set}.

[0069] If the termination condition is not met, it is necessary to update the initial population of giant pandas, continue to calculate the rank of each allocation scheme in the updated giant panda population, perform non-inferiority stratification on each allocation scheme in the updated giant panda population, calculate the fitness value of each allocation scheme in the updated giant panda population, repeat the steps of updating the giant panda population, calculating the rank of each allocation scheme in the updated giant panda population, performing non-inferiority stratification on each allocation scheme in the updated giant panda population, and calculating the fitness value of each allocation scheme in the updated giant panda population until the iteration termination condition is met.

[0070] Optionally, the number of iterations or the convergence threshold of the objective function may be set as the iteration termination condition.

[0071] According to a sensor network target allocation method provided by the present invention, the initial population of giant pandas and the updated population of giant pandas are updated according to the following method: Randomly select any two individuals in the current giant panda population, perform differential mutation operations, and generate new mutant individuals; Compare the fitness of the mutant individual with any two individuals, retain the individual with higher fitness, and obtain the offspring giant panda population; The current giant panda population and the offspring giant panda population are merged, and an updated giant panda population with the same size as the initial giant panda population is screened through the Pareto dominance method.

[0072] Specifically, in the process of updating the giant panda population, any two individuals in the current giant panda population can be randomly selected to perform a differential mutation operation, that is, subtracting the positions of the two individuals, then multiplying them by a differential weight factor, and adding the result to the position of the current individual to generate a new individual position.

[0073] Then, the fitness of the new individuals generated by differential mutation can be compared with that of the original individuals, and the individuals with better fitness can be selected as members of the next generation population, ultimately generating a population of offspring giant pandas of the current giant panda population.

[0074] For example, suppose that sensors need to be allocated to achieve optimal target coverage, and a binary code is used to represent the allocation scheme, where each position represents a sensor and takes a value of 0 or 1, indicating whether the sensor is allocated to the target.

[0075] According to the differential mutation operation in the giant panda optimization algorithm, a distribution scheme code is mutated. Assume there are three individuals (codes): A, B, C.

[0076] The initial individual code is: A=101010; B=011001; C=110011.

[0077] Two individuals (for example, A and B) can be randomly selected to perform a differential mutation operation.

[0078] Subtract the encoding positions of A and B: Δ=AB=(1-0)(0-1)(1-1)(0-0)(1-0)(0-1)=101010-011001=100011

[0079] Select a differential weight factor, such as 0.5, multiply the differential mutation result by the differential weight factor, and add it to the current position of individual A to generate a new individual position: New A = A + (Δ * differential weight factor) = 101010 + (100011 * 0.5) = 101010 + 100001 = 100010

[0080] A new individual code is generated through differential mutation operation, and then the fitness is compared with the original individual, and the individual with better fitness is selected as the member of the next generation population.

[0081] After obtaining the offspring giant panda population of the current giant panda population, the current giant panda population and the offspring giant panda population can be merged, and an updated giant panda population having the same size as the initial giant panda population can be screened through a Pareto-dominant method.

[0082] In the embodiment of the present invention, an updated giant panda population having the same size as the initial giant panda population is selected in a Pareto dominant manner, and the selection is based on Pareto non-inferiority stratification and individual crowding distance.

[0083] The method provided by the present invention is further introduced below through embodiments in specific application scenarios.

[0084] Figure 2 FIG. 1 is a flow chart of an embodiment of a sensor network target allocation method provided by the present invention. Figure 2 As shown, the method embodiment includes the following steps: S110, first, collect the initial weight values ​​of the monitored target, and evaluate these parameters to determine their importance to the target. Then, normalize these initial weight values ​​to obtain the threat weight of the target; For example, n targets are extracted from the sensor network to be monitored, and characteristic parameters of these targets are extracted. For each target, common fuzzy logic methods or neural network methods are used to estimate the importance of the target, which is expressed as (i=1,2,...,n). In order to calculate the threat weight, the obtained target importance is normalized; In step S110, the target importance calculation formula is as follows:

[0085] in, Indicates the distance from the target to the sensor, represents the speed of the target, represents the initial weight value of the target, and α, β, and γ are coefficients for adjusting the influence of distance, speed, and size.

[0086] S210, based on the sensor target perception probability model and the target threat weight, generate a sensor target monitoring effectiveness function. At the same time, generate a sensor usage function according to the sensor usage. In order to perform multi-objective optimization, the sensor target monitoring effectiveness function and the sensor usage function are standardized to obtain a constrained multi-objective optimization model; In step S210, the sensor monitoring performance function is as follows:

[0087] Among them, n means there are n targets, m means there are m sensors, and the target importance of target i is .

[0088] The sensor usage function is as follows:

[0089] in, Indicates the maximum number of monitoring target batches for m sensors.

[0090] The batch number of monitoring targets assigned to each sensor is as follows:

[0091] The constraint objective function is as follows:

[0092] in and They represent the standardized monitoring effectiveness function and sensor utilization function respectively.

[0093] S310, encoding the target allocation scheme of the sensor network, and randomly generating a set of initial allocation schemes to form an initial population of giant pandas; Figure 3 is a schematic diagram of a sub-process of step S310 provided by the present invention, such as Figure 3 As shown, step S310 includes the following sub-steps: S311, encoding multiple sensor target allocation schemes in binary form, using binary numbers to represent the allocation of each target, where each bit represents a sensor, and a value of 1 on the bit indicates that the sensor is allocated to the target, and a value of 0 indicates that the sensor is not allocated; S312, considering whether to assign the target to each sensor one by one according to the order of the targets. For each target, decide whether to assign it to each sensor according to a specific strategy or condition; S313. For each target, its assigned binary code is connected in series, the allocation situation is integrated, and a final allocation plan is formed.

[0094] The specific encoding scheme is as follows: Assume that there are n targets to be assigned and m available sensor nodes. Use binary 0-1 encoding to represent the assignment scheme of targets and sensors. Each assignment scheme consists of sensor node numbers arranged in target order, where each small segment represents the assignment result of a batch of targets. The length of the encoding scheme is n*m. Figure 4 is a schematic diagram of the coding of the sensor network target allocation scheme provided by the present invention, such as Figure 4 As shown, n is 5, m is 3, the target number is: 12345, the sensor number is: ABC, then a 1 allocation scheme of the allocation scheme population is 011100001110101, which represents a firepower allocation scheme. The above coding scheme represents a target allocation scheme, in which target 1 is monitored by sensor B and sensor C, target 2 is monitored by sensor A, target 3 is only monitored by sensor C, target 4 is monitored by sensor A and sensor B, and target 5 is monitored by sensor A and sensor C.

[0095] For the sensor network target allocation based on the giant panda optimization algorithm, the individual characteristics of the initial population will have an important impact on the optimization effect of the later algorithm. The initial population of the allocation scheme with good diversity, whose individual allocation schemes are evenly distributed and have high discreteness, is helpful to improve the convergence speed and solution accuracy of the algorithm, can reduce the probability of the algorithm falling into a local extreme value, and can effectively alleviate the algorithm premature phenomenon. In the given invention embodiment, it is mentioned that an initial population much larger than the required population size is generated, and the allocation scheme individuals suitable for the required scale are randomly selected therefrom. Then, the selected initial population is evaluated and the best allocation scheme individuals are selected. The size of the initial population finally selected is between 100 and 200.

[0096] S410. Use the giant panda optimization algorithm to iterate and screen the initial population. According to the constrained multi-objective optimization model, calculate the sensor target monitoring effectiveness function value and sensor utilization function value of each allocation scheme. Using these values, establish a coordinate system to calculate the rank of each allocation scheme, and perform non-inferiority stratification and calculate the fitness of the allocation scheme according to the size of the rank. Record the parent allocation scheme population of the optimal non-inferiority layer, that is, the Pareto front solution set; Figure 5 is the Pareto solution diagram of the sensor target allocation scheme provided by the present invention, such as Figure 5 As shown in the figure, according to the execution results of the giant panda optimization algorithm, the following hierarchical structure and rank values ​​are obtained in the constrained multi-objective optimization model plane: The first non-inferior layer (Pareto front): This layer consists of five allocation scheme individuals, represented by individuals A1, A2, A3, A4 and A5. These individuals show the optimal solution for both coverage and energy consumption. No other individuals can completely dominate them, and their ranks are all 1; The second non-inferior layer: This layer consists of six allocation scheme individuals, represented by individuals B1, B2, B3, B4, B5 and B6. These individuals still show good performance in both coverage and energy consumption. Each individual may be slightly insufficient in one goal, but achieves better results in another goal. Their ranks are all 2; The third non-inferior layer: This layer consists of four allocation scheme individuals, represented by individuals C1, C2, C3 and C4. These individuals perform relatively evenly in terms of coverage and energy consumption. Each individual has certain advantages, but may be slightly disadvantaged in one goal. Their rank values ​​are all 3.

[0097] The non-inferiority stratification of the Pareto set effectively stratifies all allocation scheme individuals in the solution space to distinguish their superiority and inferiority relationships and provide a basis for selecting a set of solutions with advantages.

[0098] S510, determine whether the initial population meets the iteration termination condition. If the termination condition is met, all the first non-inferior layer solutions in the initial population constitute the Pareto optimal solution set. If the termination condition is not met, proceed to the next step S610; Suppose there is a multi-objective optimization problem with m objective functions. For any two solutions x and y, a dominance relation can be defined to compare their objective function values: For all i∈{1,2,...,m}, we have ≤ , that is, each objective function value of solution x is less than or equal to the corresponding objective function value of solution y; There exists at least one j∈{1,2,...,m} such that < , that is, the value of solution x on at least one objective function is strictly less than the value of solution y on the corresponding objective function; According to the definition of the domination relationship, the Pareto optimal solution set can be strictly defined as the set of all solutions that are not dominated by other solutions: Pareto optimal solution set = {x|∀y ∈ solution space, if y < x, then y does not belong to the Pareto optimal solution set}; The number of iterations or the convergence threshold of the objective function can be used as the iteration termination condition.

[0099] S610. Randomly select two other individuals using the giant panda optimization algorithm and perform differential mutation operations, that is, subtract the positions of the two individuals, then multiply by a differential weight factor, and add the result to the position of the current individual to generate a new individual position. Then compare the fitness of the new individual generated by differential mutation with the original individual, and select the individual with better fitness as a member of the next generation population. Finally, generate a population of offspring allocation schemes for the sensor target allocation scheme; Suppose it is necessary to allocate sensors to achieve optimal target coverage, and use binary encoding to represent the allocation scheme, where each position represents a sensor, and the value is 0 or 1, indicating whether the sensor is allocated to the target.

[0100] An allocation scheme encoding will be mutated according to the differential mutation operation in the giant panda optimization algorithm. Suppose there are three individuals (encodings): A, B, C.

[0101] The initial individual encodings are: A = 101010; B = 011001; C = 110011.

[0102] Randomly select two individuals (for example, A and B) and perform differential mutation operations.

[0103] Subtract the encodings of A and B: Δ = A - B = (1 - 0)(0 - 1)(1 - 1)(0 - 0)(1 - 0)(0 - 1) = 101010 - 011001 = 100011

[0104] Select a differential weight factor, for example, 0.5, multiply the differential mutation result by the differential weight factor, and then add it to the position of the current individual A to generate a new individual position: New A = A + (Δ * differential weight factor) = 101010 + (100011 * 0.5) = 101010 + 100001 = 100010

[0105] A new individual encoding is generated through differential mutation operations, and then the fitness is compared with the original individual, and the individual with better fitness is selected as a member of the next generation population.

[0106] S710, merging the initial population and the offspring allocation scheme population, and selecting therefrom a population of giant pandas with the same allocation scheme size as the initial population; In step S710, a population of allocation schemes having the same size as the initial population is selected in a Pareto-dominant manner, and the selection is based on Pareto non-inferiority stratification and individual crowding distance.

[0107] S810, return to step S510, determine whether the selected allocation scheme population meets the iteration termination condition, until the termination condition is met, and finally obtain the Pareto optimal solution set. Perform target allocation on the sensor network according to the Pareto optimal solution set.

[0108] The sensor network target allocation optimization method provided by the embodiment of the present invention increases the sensor utilization function, saves sensor resources, and realizes reasonable and effective resource allocation under the premise of meeting the monitoring accuracy requirements; adopts the giant panda optimization algorithm until the Pareto optimal solution set is obtained to realize the target allocation of the sensor network, and effectively solves the problems of low accuracy and poor safety performance of the target allocation algorithm in the existing scheme.

[0109] The sensor network target allocation device provided by the present invention is described below. The sensor network target allocation device described below and the sensor network target allocation method described above can be referred to each other.

[0110] Figure 6 is a schematic diagram of the structure of the sensor network target allocation device provided by the present invention, such as Figure 6 As shown, the device includes the following modules: A weight module 600 is used to obtain a threat weight of a monitoring target according to an initial weight of the monitoring target in the sensor network; An effectiveness module 610 is used to obtain a sensor target monitoring effectiveness function based on the sensor target perception probability model and the threat weight of the monitoring target; The optimization module 620 is used to obtain the sensor network target allocation optimization result through the giant panda optimization algorithm based on the sensor target monitoring effectiveness function and the sensor utilization rate function, using the randomly generated target allocation scheme as the giant panda initial population, and perform target allocation on the sensor network based on the sensor network target allocation optimization result.

[0111] According to a sensor network target allocation device provided by the present invention, a sensor network target allocation optimization result is obtained by using a giant panda optimization algorithm, including: Based on the sensor target monitoring effectiveness function and the sensor utilization rate function, the rank of each allocation scheme in the initial population of giant pandas is calculated; Perform non-inferiority stratification on each allocation scheme based on its rank, and calculate the fitness value of each allocation scheme; When the iteration termination condition is met, all the first non-inferior layers of the initial population of giant pandas are deconstructed into a Pareto optimal solution set and used as the target allocation optimization result of the sensor network; when the iteration termination condition is not met, the initial population of giant pandas is updated, the rank of each allocation scheme in the updated giant panda population is calculated, each allocation scheme in the updated giant panda population is non-inferior stratified, and the fitness value of each allocation scheme in the updated giant panda population is calculated. The steps of updating the giant panda population, calculating the rank of each allocation scheme in the updated giant panda population, non-inferior stratified each allocation scheme in the updated giant panda population, and calculating the fitness value of each allocation scheme in the updated giant panda population are repeated until the iteration termination condition is met, and all the first non-inferior layers of the updated giant panda population are deconstructed into a Pareto optimal solution set and used as the target allocation optimization result of the sensor network.

[0112] According to a sensor network target allocation device provided by the present invention, the initial population of giant pandas and the updated population of giant pandas are updated according to the following method: Randomly select any two individuals in the current giant panda population, perform differential mutation operations, and generate new mutant individuals; Compare the fitness of the mutant individual with any two individuals, retain the individual with higher fitness, and obtain the offspring giant panda population; The current giant panda population and the offspring giant panda population are merged, and an updated giant panda population with the same size as the initial giant panda population is screened through the Pareto dominance method.

[0113] According to a sensor network target allocation device provided by the present invention, individuals of the initial population of giant pandas are generated according to the following method: In the arbitrary target allocation scheme, for any monitoring target, the allocation of the arbitrary monitoring target to each sensor is determined; Based on the order of all monitoring targets, the allocation status of each monitoring target is linked together through binary coding to form the individuals in the initial population of giant pandas.

[0114] According to a sensor network target allocation device provided by the present invention, the sensor target monitoring effectiveness function is determined according to the following formula:

[0115] in, Represents the sensor target monitoring effectiveness value of the entire sensor network; Indicates the number of monitoring targets; Indicates the number of sensors; Indicates monitoring target Threat weight value of Indicates sensor For monitoring targets The perceived probability of The value is 1 or 0, indicating that the sensor Is it assigned to the monitoring target? .

[0116] According to a sensor network target allocation device provided by the present invention, the threat weight of the monitored target is determined according to the following formula:

[0117] in, Indicates monitoring target Threat weight value of Indicates monitoring target Distance to the sensor; Indicates monitoring target speed; Indicates monitoring target The initial weight value of , and They are the weight coefficients for adjusting the influence of distance, speed and initial weight respectively.

[0118] Figure 7 is a schematic diagram of the structure of the electronic device provided by the present invention, such as Figure 7 As shown, the electronic device may include: a processor 710, a communication interface 720, a memory 730 and a communication bus 740, wherein the processor 710, the communication interface 720 and the memory 730 communicate with each other through the communication bus 740. The processor 710 may call the logic instructions in the memory 730 to execute the sensor network target allocation method, which includes: According to the initial weight of the monitoring target in the sensor network, the threat weight of the monitoring target is obtained; Based on the sensor target perception probability model and the threat weight of the monitoring target, the sensor target monitoring effectiveness function is obtained; Based on the sensor target monitoring effectiveness function and the sensor utilization rate function, the randomly generated target allocation scheme is used as the giant panda initial population. The giant panda optimization algorithm is used to obtain the sensor network target allocation optimization result, and the sensor network is targeted based on the sensor network target allocation optimization result.

[0119] In addition, the logic instructions in the above-mentioned memory 730 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0120] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the sensor network target allocation method provided by the above methods, the method includes: According to the initial weight of the monitoring target in the sensor network, the threat weight of the monitoring target is obtained; Based on the sensor target perception probability model and the threat weight of the monitoring target, the sensor target monitoring effectiveness function is obtained; Based on the sensor target monitoring effectiveness function and the sensor utilization rate function, the randomly generated target allocation scheme is used as the giant panda initial population. The giant panda optimization algorithm is used to obtain the sensor network target allocation optimization result, and the sensor network is targeted based on the sensor network target allocation optimization result.

[0121] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute the sensor network target allocation method provided by the above methods, the method comprising: According to the initial weight of the monitoring target in the sensor network, the threat weight of the monitoring target is obtained; Based on the sensor target perception probability model and the threat weight of the monitoring target, the sensor target monitoring effectiveness function is obtained; Based on the sensor target monitoring effectiveness function and the sensor utilization rate function, the randomly generated target allocation scheme is used as the giant panda initial population. The giant panda optimization algorithm is used to obtain the sensor network target allocation optimization result, and the sensor network is targeted based on the sensor network target allocation optimization result.

[0122] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0123] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0124] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A sensor network target allocation method, characterized in that: include: According to the initial weight of the monitoring target in the sensor network, the threat weight of the monitoring target is obtained; Based on the sensor target perception probability model and the threat weight of the monitoring target, a sensor target monitoring effectiveness function is obtained; Based on the sensor target monitoring effectiveness function and the sensor utilization rate function, a randomly generated target allocation scheme is used as the giant panda initial population, a sensor network target allocation optimization result is obtained through a giant panda optimization algorithm, and a target allocation is performed on the sensor network based on the sensor network target allocation optimization result.

2. The sensor network target allocation method according to claim 1, characterized in that: The sensor network target allocation optimization results are obtained through the giant panda optimization algorithm, including: Calculating the rank of each allocation scheme in the initial population of giant pandas based on the sensor target monitoring effectiveness function and the sensor utilization rate function; Performing non-inferiority stratification on each allocation scheme based on the rank of each allocation scheme, and calculating the fitness value of each allocation scheme; When the iteration termination condition is met, all the first non-inferior layers of the initial population of giant pandas are deconstructed into a Pareto optimal solution set and used as the sensor network target allocation optimization result; when the iteration termination condition is not met, the initial population of giant pandas is updated, the rank of each allocation scheme in the updated giant panda population is calculated, each allocation scheme in the updated giant panda population is non-inferior stratified, and the fitness value of each allocation scheme in the updated giant panda population is calculated, and the steps of updating the giant panda population, calculating the rank of each allocation scheme in the updated giant panda population, non-inferior stratified each allocation scheme in the updated giant panda population, and calculating the fitness value of each allocation scheme in the updated giant panda population are repeated until the iteration termination condition is met, and all the first non-inferior layers of the updated giant panda population are deconstructed into a Pareto optimal solution set and used as the sensor network target allocation optimization result.

3. The sensor network target allocation method according to claim 2, characterized in that: The initial giant panda population and the updated giant panda population are updated according to the following method: Randomly select any two individuals in the current giant panda population, perform differential mutation operations, and generate new mutant individuals; Comparing the fitness of the variant individual with the arbitrary two individuals, retaining the individual with higher fitness, and obtaining an offspring giant panda population; The current giant panda population and the offspring giant panda population are merged, and an updated giant panda population having the same size as the initial giant panda population is obtained by screening in a Pareto-dominant manner.

4. The sensor network target allocation method according to claim 1, characterized in that: The individuals of the initial population of giant pandas are generated in the following manner: In the arbitrary target allocation scheme, for any monitoring target, determining the allocation of the arbitrary monitoring target to each sensor; Based on the order of all monitoring targets, the allocation status of each monitoring target is connected in series through binary coding to form individuals in the initial population of giant pandas.

5. The sensor network target allocation method according to claim 1, characterized in that: The sensor target monitoring effectiveness function is determined according to the following formula: in, Represents the sensor target monitoring effectiveness value of the entire sensor network; Indicates the number of monitoring targets; Indicates the number of sensors; Indicates monitoring target Threat weight value of Indicates sensor For monitoring targets The perceived probability of The value is 1 or 0, indicating that the sensor Is it assigned to the monitoring target? .

6. The sensor network target allocation method according to claim 1 or 5, characterized in that: The threat weight of the monitoring target is determined according to the following formula: in, Indicates monitoring target Threat weight value of Indicates monitoring target Distance to the sensor; Indicates monitoring target speed; Indicates monitoring target The initial weight value of , and They are the weight coefficients for adjusting the influence of distance, speed and initial weight respectively.

7. A sensor network target allocation device, characterized in that: include: A weight module is used to obtain the threat weight of the monitoring target according to the initial weight of the monitoring target in the sensor network; An effectiveness module, used to obtain a sensor target monitoring effectiveness function based on a sensor target perception probability model and a threat weight of the monitoring target; The optimization module is used to obtain the sensor network target allocation optimization result through the giant panda optimization algorithm based on the sensor target monitoring effectiveness function and the sensor utilization rate function, and to perform target allocation on the sensor network based on the sensor network target allocation optimization result, using the randomly generated target allocation scheme as the giant panda initial population.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the sensor network target allocation method according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the sensor network target allocation method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the sensor network target allocation method according to any one of claims 1 to 6 is implemented.

Images