Data-optimized weight comprehensive weighting target dynamic threat assessment method and system

Abstract

The present disclosure relates to a kind of dynamic threat evaluation methods based on data optimization weight comprehensive weighting target, including for specific target threat, obtain multiple evaluation indexes in target threat, and multiple evaluation indexes are handled, obtain normalized decision matrix;With normalized decision matrix, subjective weight optimization model and objective weight optimization model are constructed, subjective weight optimization model and objective weight optimization model are solved respectively, subjective weight and objective weight are obtained respectively, and subjective weight and objective weight are handled, and comprehensive weight is obtained;Dynamic evaluation is carried out to target threat using comprehensive weight and normalized decision matrix, and the dynamic evaluation result of target threat is obtained.The present disclosure can realize the dynamic ordering of target threat, effectively solve the problem that the generalization ability is not strong and the research is insufficient in the implementation process of existing target threat evaluation, provide strong support for command and decision-making.

Description

TECHNICAL FIELD

[0001] The present disclosure belongs to the technical field of information fusion, and particularly relates to the technical field of target threat assessment, and especially relates to a data-optimized weight-based comprehensive weighting target dynamic threat assessment method and system. BACKGROUND

[0002] Target threat assessment refers to inferring the threat degree of an enemy to a friendly force according to the situation of both sides, so as to provide support for the decision and command of a commander. This is an important basis for the combat decision of a commander and an important research content of information fusion. From the perspective of the assessment object, there are mainly air targets, ground targets, radiation source targets and group targets, etc.; from the perspective of the threat assessment method, there are mainly assessment based on multi-attribute decision theory, assessment based on fuzzy neural network, assessment based on Bayes network, rough set, cloud model theory and interval number method, etc. Since the multi-attribute decision theory comprehensively considers multiple factors in target threat, it can fully reflect the influence of multiple factors on assessment, and has become a research and application hotspot in the field of threat assessment.

[0003] In target threat assessment based on multi-attribute decision theory, four types of problems are mainly concerned: first, the selection of target threat assessment indexes and the representation form of assessment index data. Taking air target threat assessment indexes as an example, they can include target speed, distance, acceleration, azimuth, height, heading, target type, electronic interference and attack intention, etc.; and the representation form of assessment index data has multiple forms such as real number, fuzzy number, interval number and language variable, etc.; second, the rationality of the weight setting of each index. Generally, there are subjective weight method, objective weight method and method combining subjective and objective methods; third, the acquisition method of target threat ranking result. Generally, there are operator weighting method, approximation ideal solution ranking method and dominant decision method, etc.; fourth, the comprehensive consideration of current time and historical time data in dynamic threat assessment.

[0004] However, in the actual implementation process, although the traditional method combining subjective and objective methods comprehensively considers subjective and objective factors, the method for obtaining subjective weight and objective weight is relatively single, the value of data is not fully tapped, and the generalization ability is not strong; and in the current technology, there are more studies on single time static threat assessment, but the research on dynamic threat assessment is not sufficient, and the change of the target threat degree in the complex and changeable battlefield situation cannot be effectively reflected. Therefore, it is necessary to improve one or more problems in the above-mentioned related technical solutions.

[0005] It should be noted that the information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The first aspect of the embodiment of the present disclosure provides a dynamic threat evaluation method based on data optimization weight comprehensive weighting target, which comprises the following steps:

[0007] For a specific target threat, a plurality of evaluation indexes in the target threat are acquired, and the plurality of evaluation indexes are processed to obtain a normalized decision matrix;

[0008] A subjective weight optimization model and an objective weight optimization model are constructed in combination with the normalized decision matrix, the subjective weight optimization model and the objective weight optimization model are solved respectively to obtain a subjective weight and an objective weight, and the subjective weight and the objective weight are processed to obtain a comprehensive weight;

[0009] The target threat is dynamically evaluated by using the comprehensive weight and the normalized decision matrix to obtain a dynamic evaluation result of the target threat.

[0010] In an example embodiment of the present disclosure, the step of acquiring a plurality of evaluation indexes in the target threat and processing the plurality of evaluation indexes to obtain a normalized decision matrix for a specific target threat comprises:

[0011] For a specific target threat, a plurality of evaluation indexes are acquired through a sensor network or a command system;

[0012] A plurality of quantitative evaluation indexes are obtained by quantitatively processing the plurality of evaluation indexes from the aspects of the type, interference ability and damage ability of the target threat;

[0013] A plurality of optimized evaluation indexes are obtained by optimizing the plurality of quantitative evaluation indexes by using a minimum variance method;

[0014] A normalized decision matrix is obtained by normalizing the plurality of optimized evaluation indexes by using a fuzzy membership function.

[0015] In an example embodiment of the present disclosure, in the step of optimizing the plurality of quantitative evaluation indexes by using a minimum variance method to obtain a plurality of optimized evaluation indexes, the expression of the optimized evaluation index comprises:

[0016]

[0017] wherein, s j represents the jth optimized evaluation index, represents the mean value of the jth evaluation index of all target threats, x ij represents the value of the jth evaluation index of the ith target threat; n represents the number of target threats; if s k≈0(≤α),1≤k≤m,α,then x k Can be rejected; x k represents the value of the kth evaluation index;

[0018] In the step of normalizing the plurality of optimization evaluation indexes by using the fuzzy membership function to obtain a normalized decision matrix,

[0019] The expression of the fuzzy membership function comprises:

[0020]

[0021] wherein x s represents the fuzzy membership function; f s represents the evaluation index, m s ≤f s ≤M s , m s represents the minimum value, M s represents the maximum value, and f represents the ideal moderate value;

[0022] The expression of the normalized decision matrix comprises:

[0023] X = (x ij ) n×m (3)

[0024] wherein n x m represents an n-row m-column normalized decision matrix.

[0025] In an example embodiment of the present disclosure, the step of combining the normalized decision matrix to construct a subjective weight optimization model and an objective weight optimization model, respectively solving the subjective weight optimization model and the objective weight optimization model to obtain a subjective weight and an objective weight, and processing the subjective weight and the objective weight to obtain a comprehensive weight comprises:

[0026] constructing the subjective weight optimization model and solving the subjective weight optimization model to obtain the subjective weight;

[0027] constructing the objective weight optimization model according to the total deviation of the target threat and the positive and negative ideal targets and the maximum entropy principle, and solving the objective weight optimization model to obtain the objective weight;

[0028] processing the subjective weight and the objective weight by using a linear weighting method to obtain the comprehensive weight.

[0029] In an example embodiment of the present disclosure, the step of constructing the subjective weight optimization model and solving the subjective weight optimization model to obtain the subjective weight comprises:

[0030] determining an expert group from an expert library, calculating the importance of each expert in the expert group, and the expression of calculating the importance of each expert in the expert group comprising:

[0031]

[0032] wherein b i represents the importance of the i th expert; z i represents the number of times that the i th expert meets the actual threat assessment test; Z represents the total number of threat assessment tests; p represents the number of experts;

[0033] each of the experts respectively subjectively evaluates the evaluation indexes, and respectively calculates the initial index weight of each of the experts by using the analytic network process

[0034] according to the importance of each of the experts and the initial index weight of each of the experts, calculating the preference degree of the expert group to the evaluation indexes, and establishing a subjective weight optimization model by using the Hamming distance;

[0035] the expression of calculating the preference degree of the expert group to the evaluation indexes comprising:

[0036]

[0037] wherein d i represents the preference degree of the expert group to the i th evaluation index; represents the initial index weight of the j th evaluation index of the q th expert; k q represents the importance of the decoupled expert; x ij represents the observation value of the j th evaluation index of the i th target threat; q represents the q th expert;

[0038] the expression of the subjective weight optimization model comprising:

[0039]

[0040] wherein d i represents the preference degree of the expert group to the i th evaluation index; s.t. represents “constrained to”; w j represents the initial index weight of the j th evaluation index; d(X,Y) represents the Hamming distance; F1 represents the subjective weight optimization model; m represents the number of evaluation indexes; n represents the number of target threats; the subjective weight optimization model F1 is minimized to obtain minF1;

[0041] constructing a Lagrange function, and solving the subjective weight optimization model by using the Lagrange conditional extremum method to obtain the subjective weight;

[0042] The expression of the Lagrange function includes:

[0043]

[0044] Wherein, F1 represents a subjective weight optimization model; λ represents a Lagrange multiplier method operator; w j represents the initial index weight of the jth evaluation index;

[0045] The expression of the subjective weight includes:

[0046]

[0047] Wherein, represents the subjective weight; -1 represents the reciprocal; d represents the preference degree.

[0048] In an example embodiment of the present disclosure, the step of constructing the objective weight optimization model according to the total deviation of the target threat and the positive and negative ideal targets, and solving the objective weight optimization model to obtain the objective weight includes:

[0049] Constructing a total deviation scheme of the target threat and the positive and negative ideal targets;

[0050] The expression of the positive and negative ideal targets includes:

[0051]

[0052] The expression of the total deviation of the target threat and the positive and negative ideal targets includes:

[0053]

[0054] Wherein, represents the positive ideal target, represents the negative ideal target,

[0055] The expression of the objective weight optimization model includes:

[0056]

[0057] Wherein, D(w) represents the total deviation, H represents the maximum entropy, μ represents a balance coefficient between multiple condition evaluation indexes, 0≤μ≤1;

[0058] Constructing a Lagrange function and solving the objective weight optimization model by using the Lagrange condition extremum method to obtain the objective weight.

[0059] The expression of the objective weight includes:

[0060]

[0061] In an example embodiment of the present disclosure, in the step of processing the subjective weight and the objective weight by using the linear weighting method to obtain a comprehensive weight,

[0062] The expression of the comprehensive weight includes:

[0063]

[0064] Wherein, α represents a balance factor between the subjective weights, β represents a balance factor between the objective weights, 0≤α≤1, 0≤β≤1; The subjective weight is represented as w The objective weight is represented as w j The comprehensive weight is represented as w

[0065] In an example embodiment of the present disclosure, the step of dynamically evaluating the target threat by using the comprehensive weight and the normalized decision matrix to obtain a dynamic evaluation result of the target threat includes:

[0066] The normalized decision matrix is weighted and processed by using the comprehensive weight to obtain a comprehensive weighted decision matrix;

[0067] The comprehensive weighted decision matrix is processed by using a TOPSIS algorithm or a VIKOR algorithm, and a relative closeness C i ,i=1,…,n or a threat compromise value Q i ,i=1,…,n is used to represent the single-time target threat degree obtained;

[0068] The time series weight is obtained by using an inverse form of Poisson distribution method or a normal cumulative distribution method;

[0069] The single-time target threat degree is weighted by using the time series weight to obtain a comprehensive weighted dynamic threat degree;

[0070] The target threat is sorted according to the relative size of the comprehensive weighted dynamic threat degree, and the sorting result is the dynamic evaluation result of the target threat.

[0071] In an example embodiment of the present disclosure, in the step of weighting and processing the normalized decision matrix by using the comprehensive weight to obtain a comprehensive weighted decision matrix,

[0072] The expression of the comprehensive weighted decision matrix includes:

[0073] V = (v ij ) n×m ,v ij = w j × x ij (14)

[0074] wherein v ij represents the value of the weighted index.

[0075] In the step of calculating the time series weight by using the Poisson distribution method or the normal cumulative distribution method,

[0076] The expression for calculating the time series weight by using the Poisson distribution method comprises:

[0077]

[0078] wherein η k ≥ 0, represents the Poisson distribution parameter;

[0079] The expression for calculating the time series weight by using the normal cumulative distribution method comprises:

[0080]

[0081] wherein F(t k ; μ p , σ p ) represents the normal cumulative distribution function, t represents the duration; t k represents the time weight at the kth time; μ p represents the expectation of the normal distribution, σ p represents the standard deviation of the normal distribution, k = 1, 2, …, p, t > 0.

[0082] In the step of weighting the single-time target threat degree by using the time series weight to obtain the comprehensive weighted dynamic threat degree,

[0083] If the TOPSIS algorithm and the Poisson distribution method are used, the expression for the comprehensive weighted dynamic threat degree comprises:

[0084]

[0085] If the TOPSIS algorithm and the normal cumulative distribution method are used, the expression for the comprehensive weighted dynamic threat degree comprises:

[0086]

[0087] If the VIKOR algorithm and the Poisson distribution method are used, the comprehensive weighted dynamic threat degree expression includes:

[0088]

[0089] If the VIKOR algorithm and the normal cumulative distribution method are used, the comprehensive weighted dynamic threat degree expression includes:

[0090]

[0091] wherein C ik represents the relative proximity at time t k ; Q ik represents the threat compromise value at time t k ; η k represents the time series weight using the Poisson distribution method; t k represents the time weight at the kth time; λ represents the Lagrange multiplier method operator; and p represents the number of experts.

[0092] A second aspect of the embodiments of the present disclosure provides a data-optimized weight-based comprehensive weighted target dynamic threat evaluation system, which comprises an evaluation index processing module, an evaluation index optimized weight module, a target threat dynamic evaluation module, a display module and a storage module, wherein

[0093] the evaluation index processing module is respectively connected with a command system or a sensor network and the evaluation index optimized weight module; the evaluation index processing module is configured to obtain a plurality of evaluation indexes in a target threat, and process the plurality of evaluation indexes to obtain a normalized decision matrix;

[0094] the evaluation index optimized weight module is respectively connected with the evaluation index processing module and the target threat dynamic evaluation module; the evaluation index optimized weight module comprises a subjective weight submodule, an objective weight submodule and a comprehensive weight submodule; the subjective weight submodule is configured to generate a subjective weight; the objective weight submodule is configured to generate an objective weight; and the comprehensive weight submodule is configured to linearly weight process the subjective weight and the objective weight to obtain a comprehensive weight;

[0095] the target threat dynamic evaluation module is configured to dynamically evaluate the target threat using the normalized decision matrix and the comprehensive weight;

[0096] the display module is configured to display information of the evaluation index processing module, the evaluation index optimized weight module and the target threat dynamic evaluation module.

[0097] The storage module is used for storing and counting information of the evaluation index processing module, the evaluation index optimization weight module and the target threat dynamic evaluation module.

[0098] The technical solution provided by the present disclosure can include the following beneficial effects:

[0099] The first aspect of the present disclosure provides a target dynamic threat evaluation method based on data optimization weight comprehensive weighting, which processes a plurality of evaluation indexes in the target threat, respectively establishes a subjective optimization model and an objective optimization model, and dynamically evaluates the target threat according to the obtained normalized decision matrix and comprehensive weight, realizes dynamic ordering of the target threat, effectively solves the problems of weak generalization ability and insufficient research in the implementation process of the existing target threat evaluation, and provides strong support for command and decision-making. The second aspect of the present disclosure provides a target dynamic threat evaluation system based on data optimization weight comprehensive weighting, which is designed by connecting the evaluation index processing module, the comprehensive weight module, the dynamic threat evaluation module, and the display and storage module in sequence, and fully excavates the support value of the evaluation index to the weight. BRIEF DESCRIPTION OF DRAWINGS

[0100] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and serve to explain the principles of the present disclosure. It is apparent that the accompanying drawings in the following description are only some embodiments of the present disclosure, and other drawings can be obtained according to the drawings without creative labor.

[0101] Figure 1 A step schematic diagram of the target dynamic threat evaluation method based on data optimization weight comprehensive weighting in the exemplary embodiment of the present disclosure is shown;

[0102] Figure 2 A flow schematic diagram of the target dynamic threat evaluation method based on data optimization weight comprehensive weighting in the exemplary embodiment of the present disclosure is shown;

[0103] Figure 3 A structure schematic diagram of the target dynamic threat evaluation system based on data optimization weight comprehensive weighting in the exemplary embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0104] Example implementations will now be described more fully with reference to the accompanying drawings. Example implementations can be implemented in any number of ways, and are not limited to the examples described herein. Rather, examples will be provided for illustration purposes, and one skilled in the art will recognize that modifications and other implementations are possible. Features described with respect to one implementation can be combined with features described with respect to another implementation.

[0105] Furthermore, the accompanying drawings are included to provide a further understanding of the present disclosure and are incorporated in and constitute a part of this specification. The drawings are not necessarily to scale, the same or similar reference numerals designate the same or similar parts throughout the several views, and the detailed description is presented in connection with the accompanying drawings.

[0106] The first aspect of the example embodiment provides a dynamic threat assessment method based on data optimization weight comprehensive weighting target, which can include the following steps: Figure 1 The first aspect of the example embodiment provides a dynamic threat assessment method based on data optimization weight comprehensive weighting target, which can include the following steps:

[0107] Step S101: For a specific target threat, a plurality of evaluation indexes in the target threat are obtained, and the plurality of evaluation indexes are processed to obtain a normalized decision matrix;

[0108] Step S102: A subjective weight optimization model and an objective weight optimization model are constructed in combination with the obtained normalized decision matrix, the subjective weight optimization model and the objective weight optimization model are solved respectively to obtain a subjective weight and an objective weight, and the subjective weight and the objective weight are processed to obtain a comprehensive weight;

[0109] Step S103: The target threat is dynamically evaluated by using the comprehensive weight and the normalized decision matrix to obtain a dynamic evaluation result of the target threat.

[0110] In the example embodiment of the present disclosure,

[0111] The above method of the example embodiment will be described in more detail below.

[0112] As shown in Figure 2 Step S101 includes the following sub-steps:

[0113] Sub-step S1011: For a specific target threat, a plurality of evaluation indexes are obtained through a sensor network or a command system;

[0114] Sub-step S1012: The plurality of evaluation indexes are quantitatively processed from the aspects of the type of the target threat, the interference ability of the target threat, and the killing ability of the target threat to obtain a plurality of quantitative evaluation indexes;

[0115] Sub-step S1013: The plurality of quantitative evaluation indexes are respectively optimized by using the least squares to obtain a plurality of optimized evaluation indexes;

[0116] In sub-step S1014, the multiple evaluation indexes are normalized by using fuzzy membership functions to obtain a normalized decision matrix.

[0117] In the evaluation of the air defense target threat, both sides keep secret from each other, and the intelligent side judges according to the target threat information obtained by the detection device and tracking device, and the local information mastered in peacetime. According to the combat principle and the characteristic information provided by the sensor, the main evaluation indexes affecting the target threat can be determined, including the type of target threat, the speed of target threat, the acceleration of target threat, the height of target threat, the distance of target threat, the heading angle of target threat, the jamming ability of target threat, and the killing ability of target threat.

[0118] These main evaluation indexes can be classified into qualitative indexes and quantitative indexes according to the attribute characteristics of multi-attribute decision; the qualitative indexes include the type of target threat, the jamming ability of target threat, and the killing ability of target threat; the quantitative indexes include the type of target threat, the speed of target threat, the acceleration of target threat, the height of target threat, the distance of target threat, and the heading angle of target threat.

[0119] These main evaluation indexes can be classified into benefit-type indexes and cost-type indexes according to the meaning of the attribute; the benefit-type indexes include the type of target threat, the speed of target threat, the acceleration of target threat, the jamming ability of target threat, and the killing ability of target threat; the cost-type indexes include the height of target threat, the distance of target threat, and the heading angle of target threat.

[0120] For a specific target threat, first, multiple evaluation indexes are obtained from the sensor network or the command system through step S1011.

[0121] In step S1012, for the qualitative indexes, i.e., the type of target threat, the jamming ability of target threat, and the killing ability of target threat, quantitative processing is performed to obtain multiple quantitative evaluation indexes. Here, the quantitative evaluation indexes can be represented by index numbers, as shown in Tables 1-3 below,

[0122] Table 1: Quantitative table of the type of target threat

[0123] Type of target threat Armed helicopter Large target threat Small target threat Indicator number 1 2 3

[0124] Table 2: Quantitative table of the jamming ability of target threat

[0125] Jamming capability of target threat None Weak Medium Strong Indicator number 1 2 3 4

[0126] Table 3: Quantitative table of the killing ability of target threat

[0127]

[0128] As can be seen from Tables 1-3, the multiple evaluation indicators mentioned in step S1011 are mainly specific indicators. In this embodiment, among the types of target threats, the indicator number for armed helicopters is 1, for large target threats it is 2, and for small target threats it is 3. The interference capability of target threats is divided into four levels: no interference capability is 1, some interference capability is 2, medium interference capability is 3, and strong interference capability is 4. Among the lethality of target threats, other weapons are 1, conventional weapons are 2, and nuclear weapons are 3.

[0129] In step S1013, if there are n target threats: s1, s2, ..., s n Choose m evaluation indicators from them, and use x ij Let x represent the observed value of the j-th evaluation metric for the threat of the i-th target. Here, if x represents the threat of n targets... ij They are basically the same, although x ij It is very important, but it is not very useful for comparing the feature levels of target threats. For gradual variation, x can be removed. ij Then we can obtain the expression for the optimized evaluation index:

[0130]

[0131] Among them, s j Let j represent the optimization evaluation index. Let represent the mean of the j-th assessment metric for all target threats. x ij represents the value of the j-th evaluation index for the i-th target threat; n represents the number of target threats.

[0132] It should be noted that if s k ≈0 (≤α), 1≤k≤m, α needs to be chosen based on the actual situation, then x k Can be removed; x k This represents the value of the k-th evaluation indicator; for qualitative indicators, when all indicator values ​​are the same, the evaluation indicator can be removed.

[0133] In step S1014, the expression for the fuzzy membership function is:

[0134]

[0135] Where, x s f represents a fuzzy membership function; s Indicates the evaluation index, m s ≤f s ≤M s, m s represents the minimum value, M s represents the maximum value, f represents an ideal moderate value; when x s =(f s -m s ) / (M s -m s ), it is a benefit type index; when x s =1-|f s -f| / (M s -m s ), it is a central type index; when x s =(M s -f s ) / (M s -m s ), it is a cost type index.

[0136] After the plurality of optimization evaluation indexes are normalized, a membership matrix X is obtained

[0137] X=(x ij ) n×m (3)

[0138] Wherein, n*m represents an n row m column normalized decision matrix.

[0139] The membership matrix is the normalized decision matrix finally obtained.

[0140] Step S102 includes the following sub-steps:

[0141] Sub-step S1021, a subjective weight optimization model is constructed, the subjective optimization model is solved, and a subjective weight is obtained;

[0142] Sub-step S1022, an objective weight optimization model is constructed according to the total deviation of the target threat and the positive and negative ideal targets, and the maximum entropy principle, and the objective weight optimization model is solved, and an objective weight is obtained;

[0143] Sub-step S1023, the subjective weight and the objective weight are processed by using a linear weighting method, and a comprehensive weight is obtained.

[0144] In sub-step S1021,

[0145] First, an expert group is selected from an expert library, the importance of each expert in the expert group is calculated, and the calculation formula is:

[0146]

[0147] Wherein, b i represents the importance of the i th expert; z irepresents the number of times that the ith expert meets the actual threat assessment test; Z represents the total number of threat assessment tests; and p represents the number of experts. When the number of experts is too large, the importance coefficient of the experts can be coupled, and the average method can be used to realize decoupling operation, that is,

[0148]

[0149] wherein k i represents the importance coefficient of the experts.

[0150] Secondly, the experts in the expert group subjectively evaluate each evaluation index, and the analytic network process (ANP) is used to calculate the initial index weight corresponding to each expert

[0151] Thirdly, according to the importance of each expert and the initial index weight corresponding to each expert, the preference degree of the expert group to multiple evaluation indexes is calculated, and a subjective weight optimization model is established by using the Hamming distance.

[0152] The expression for calculating the preference degree of the expert group to the evaluation indexes comprises:

[0153]

[0154] wherein d i represents the preference degree of the expert group to the ith evaluation index; represents the initial index weight of the jth evaluation index corresponding to the qth expert; k q represents the importance of the decoupled expert; x ij represents the observation value of the jth evaluation index of the ith target threat; and q represents the qth expert.

[0155] The expression of the subjective weight optimization model is:

[0156]

[0157] wherein d i represents the preference degree of the expert group to the ith evaluation index; s.t. represents “constrained to”; w j represents the initial index weight of the jth evaluation index; d(X, Y) represents the Hamming distance; F1 represents the subjective weight optimization model; m represents the number of evaluation indexes; n represents the number of target threats; and minF1 is obtained by minimizing the subjective weight optimization model F1.

[0158] The Hamming distance mentioned herein is:

[0159]

[0160] Finally, the Lagrange function is constructed, and the subjective weight optimization model is solved by using the Lagrange conditional extremum method to obtain the subjective weight.

[0161] Here, the expression of the Lagrange function is:

[0162]

[0163] wherein F1 represents the subjective weight optimization model; λ represents the Lagrange multiplier operator; w j represents the initial index weight of the jth evaluation index;

[0164] The expression of the subjective weight is:

[0165]

[0166] wherein w represents the subjective weight; -1 represents the reciprocal; and d represents the preference degree.

[0167] In the sub-step S1022,

[0168] First, the expression of the positive and negative ideal targets of the target threat is defined as:

[0169]

[0170] For a certain evaluation index, the total deviation of the target threat from the positive ideal target is represented by D+; the total deviation of the target threat from the negative ideal target is represented by D-; and the expression of the total deviation of the target threat from the positive and negative ideal targets is:

[0171]

[0172]

[0173] The selection of the index weight should make the total deviation of all target threats from the positive ideal target minimum and the total deviation of all target threats from the negative ideal target maximum, i.e.:

[0174]

[0175]

[0176] Combining all the evaluation indexes, we have:

[0177]

[0178] Secondly, the objective weight optimization model is constructed by using the maximum entropy principle, and the expression of the objective weight optimization model is:

[0179] ​​​

[0180] wherein D(w) represents the total deviation, represents the maximum entropy, μ represents a balance coefficient between the plurality of evaluation indexes, 0≤μ≤1, μ is determined according to specific conditions, and represents the balanced competition between conditions in calculation, and is usually taken as

[0181] Finally, a Lagrange function is constructed, and the objective weight optimization model is solved by using the Lagrange condition extremum method to obtain the objective weight.

[0182] The constructed Lagrange function is as follows:

[0183]

[0184] According to the necessary condition of extremum existence:

[0185]

[0186]

[0187] The expression of the objective weight can be obtained as follows:

[0188]

[0189] In step S1023, the expression of the comprehensive weight is as follows:

[0190]

[0191] wherein α represents a balance factor between the subjective weights, β represents a balance factor between the objective weights, 0≤α≤1, 0≤β≤1; represents the subjective weight; represents the objective weight, w j represents the comprehensive weight.

[0192] Step S103 includes the following sub-steps:

[0193] Sub-step S1031, the normalized decision matrix is weighted by using the comprehensive weight to obtain a comprehensive weighted decision matrix;

[0194] Sub-step S1032, the comprehensive weighted decision matrix is processed by using a TOPSIS algorithm or a VIKPR algorithm, and the single-moment target threat degree is obtained by using a relative closeness C i ,i=1,…,n or a threat compromise value Q i ,i=1,…,n.

[0195] In sub-step S1033, the time series weight is calculated by using the inverse form of Poisson distribution method or normal cumulative distribution method.

[0196] In sub-step S1034, the single-time target threat degree is weighted by using the time series weight to obtain a comprehensive weighted dynamic threat degree.

[0197] In sub-step S1035, the target threat is sorted according to the relative size of the comprehensive weighted dynamic threat degree, and the sorting result is the dynamic evaluation result of the target threat.

[0198] In sub-step S1031, the expression of the comprehensive weighted decision matrix can be:

[0199] V=(v ij ) n×m ,v ij =w j ×x ij (14)

[0200] wherein v ij represents the value of the weighted index.

[0201] In sub-step S1033,

[0202] If the Poisson distribution method is used to calculate the time series weight, the expression is:

[0203]

[0204] wherein η k ≥ 0, is usually taken as represents the Poisson distribution parameter.

[0205] If the normal cumulative distribution method is used to calculate the time series weight, the expression is:

[0206]

[0207] wherein F(t k ; μ p , σ p ) represents the normal cumulative distribution function, t represents the duration; t k represents the time weight at the kth moment; μ p represents the expectation of the normal distribution, σ p represents the standard deviation of the normal distribution,

[0208] Here, the cumulative distribution function of the normal distribution can be represented by a special function of an error function:

[0209]

[0210] In sub-step S1034, if there are p time points within the time sequence, for n target threats, according to the single time point decision method and the time sequence weight calculation method discussed above, the following cases exist:

[0211] If the TOPSIS algorithm and the Poisson distribution method are used, the comprehensive weighted dynamic threat degree expression includes:

[0212]

[0213] If the TOPSIS algorithm and the normal cumulative distribution method are used, the comprehensive weighted dynamic threat degree expression includes:

[0214]

[0215] If the VIKOR algorithm and the Poisson distribution method are used, the comprehensive weighted dynamic threat degree expression includes:

[0216]

[0217] If the VIKOR algorithm and the normal cumulative distribution method are used, the comprehensive weighted dynamic threat degree expression includes:

[0218]

[0219] wherein C ik represents the relative closeness at time t k ; Q ik represents the threat compromise value at time t k ; η k represents the time sequence weight using the Poisson distribution method; t k represents the time weight at the kth time point; λ represents the Lagrange multiplier method operator; and p represents the number of experts.

[0220] In sub-step S1035, the target threats are sorted according to the relative sizes of the comprehensive weighted dynamic threat degrees, and the sorting result is the dynamic assessment result of the target threats.

[0221] Here, if the target threat degree is represented by the relative closeness C i , i = 1, …, n, the greater the relative closeness, the greater the target threat degree; and if the target threat degree is represented by the threat compromise value Q i , the greater the threat compromise value, the greater the target threat degree.Let i = 1,..., n represent the target threat degree, then the smaller the threat compromise value is, the greater the target threat degree is.

[0222] The second aspect of the example embodiment provides a dynamic threat evaluation system of a target based on data optimization weight comprehensive weighting. As shown in the figure, the system comprises an evaluation index processing module, an evaluation index optimization weight module, a target threat dynamic evaluation module, a display module and a storage module, wherein, Figure 3

[0223] The evaluation index processing module is respectively connected with the command system or the sensor network and the evaluation index optimization weight module; the evaluation index processing module is used for obtaining a plurality of evaluation indexes in a target threat, and processing the plurality of evaluation indexes to obtain a normalized decision matrix; the evaluation index optimization weight module is respectively connected with the evaluation index processing module and the target threat dynamic evaluation module; the evaluation index optimization weight module comprises a subjective weight submodule, an objective weight submodule and a comprehensive weight submodule; the subjective weight submodule is used for generating a subjective weight; the objective weight submodule is used for generating an objective weight; and the comprehensive weight submodule is used for linearly weighting and processing the subjective weight and the objective weight to obtain a comprehensive weight.

[0224] The target threat dynamic evaluation module dynamically evaluates the target threat by using the normalized decision matrix and the comprehensive weight.

[0225] The display module is used for displaying information of the evaluation index processing module, the evaluation index optimization weight module and the target threat dynamic evaluation module.

[0226] The storage module is used for storing and counting information of the evaluation index processing module, the evaluation index optimization weight module and the target threat dynamic evaluation module.

[0227] It should be noted that the subjective weight submodule comprises an expert importance unit, an initial weight unit and a subjective weight solving unit; the expert importance unit and the initial weight unit are used for constructing a subjective weight model, and the subjective weight solving unit is used for solving the subjective weight model to obtain a subjective weight. The objective weight submodule comprises an objective weight construction unit, which is used for constructing an objective weight model; and an objective weight solving unit, which is used for solving the objective weight model to obtain an objective weight.

[0228] The evaluation index processing module directly interacts with the command system or the sensor network, and at least comprises an evaluation index database, data reading of the evaluation index, data calculation of the evaluation index and the like, to support obtaining an evaluation target, reading data of the evaluation index, quantifying a qualitative index and normalizing the evaluation index.

[0229] ​The target threat dynamic evaluation module obtains the target threat degree at a single moment by weighting and normalizing the decision matrix with comprehensive weights and using sorting algorithms such as the TOPSIS algorithm and the VIKOR algorithm, and obtains the comprehensive weighted dynamic threat degree in combination with a time series algorithm.

[0230] It should be noted that although several units of the system for action execution are mentioned in the foregoing detailed description, such a division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided into units embodied by multiple units. Part or all of the units can be selected to achieve the purpose of the present disclosure according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0231] Other embodiments of the present disclosure will be readily apparent to those skilled in the art upon considering the description herein, with the disclosure intended to be representative of the generic principles of the present disclosure and including any and all variations, uses, or adaptations of the present disclosure following in general the principles of the present disclosure and including such art-recognized equivalents and substitutes for the specific embodiments disclosed. The scope of the present disclosure is accordingly intended to be indicated by the appended claims rather than the description and examples herein, which are to be considered exemplary only.

Claims

1. A dynamic threat assessment method for targets based on data-optimized weighted comprehensive weighting, characterized in that, include: For a specific aerial target threat, multiple assessment indicators of the target threat are obtained, and the multiple assessment indicators are processed to obtain a normalized decision matrix; the multiple assessment indicators include type, speed, acceleration, altitude, range, heading angle, jamming capability and lethality; The steps of obtaining multiple evaluation indicators from a specific target threat and processing these indicators to obtain a normalized decision matrix include: For specific target threats, multiple evaluation metrics are acquired through sensor networks or command systems; The evaluation indicators are quantified in terms of the type, interference capability, and lethality of the target threat to obtain multiple quantitative evaluation indicators. The minimum variance method is used to optimize the multiple quantitative evaluation indicators to obtain multiple optimized evaluation indicators. The optimization evaluation indicators are normalized using fuzzy membership functions to obtain a normalized decision matrix; In the step of optimizing multiple quantitative evaluation indicators using the minimum variance method to obtain multiple optimized evaluation indicators, the expressions for the optimized evaluation indicators include: （1） in, Indicates the first j Optimize evaluation indicators ; The first, representing the threat to all targets j The average of the evaluation indicators, ; Indicates the first i The first target threat j The value of each evaluation indicator; n Indicates the number of target threats; if ,but Can be removed; express k The value of each evaluation indicator; In the step of normalizing multiple optimization evaluation indicators using fuzzy membership functions to obtain a normalized decision matrix... The expression for the fuzzy membership function includes: （2） in, Represents fuzzy membership functions; Indicates the evaluation indicators, , This represents the minimum value. Indicates the maximum value. This represents the ideal moderate value; The expression for the normalized decision matrix includes: (3) in, express n OK m Column normalized decision matrix; Based on the normalized decision matrix, a subjective weight optimization model and an objective weight optimization model are constructed. The subjective weight optimization model and the objective weight optimization model are solved respectively to obtain the subjective weight and the objective weight. The subjective weight and the objective weight are then processed to obtain the comprehensive weight. The steps of constructing a subjective weight optimization model and an objective weight optimization model by combining the normalized decision matrix, solving the subjective weight optimization model and the objective weight optimization model respectively to obtain the subjective weight and the objective weight, and processing the subjective weight and the objective weight to obtain the comprehensive weight include: Construct the subjective weight optimization model and solve the subjective weight optimization model to obtain the subjective weights; Based on the total deviation between the target threat and the positive and negative ideal targets, and the principle of maximum entropy, the objective weight optimization model is constructed, and the objective weight optimization model is solved to obtain the objective weights; The subjective weights and objective weights are processed using a linear weighting method to obtain a comprehensive weight; The target threat is dynamically evaluated using the comprehensive weights and the normalized decision matrix to obtain the dynamic evaluation result of the target threat.

2. The dynamic threat assessment method for targets based on data optimization and weighted comprehensive assessment as described in claim 1, characterized in that, The steps of constructing the subjective weight optimization model and solving the subjective weight optimization model to obtain the subjective weights include: An expert group is determined from the expert database, and the importance of each expert in the expert group is calculated; the expression for calculating the importance of each expert in the expert group includes: （4） in, Indicates the first i The importance of each expert; Indicates the first i The number of times an expert conducts a realistic threat assessment test; Z This indicates the total number of threat assessment tests. p Indicates the number of experts; Each expert subjectively evaluates the assessment indicators, and the initial indicator weights for each expert are calculated using the Analytic Hierarchy Process (AHP). ; Based on the importance of each expert and the initial index weight corresponding to each expert, the expert group's preference for the evaluation index is calculated, and a subjective weight optimization model is established using Hamming distance. The expression for calculating the expert group's preference for the evaluation index includes: （5） in, The expert panel indicated that it had no opinion on the matter. i Preference for each evaluation indicator; Indicates the first q The expert on the first j Initial indicator weights for each evaluation indicator; This indicates the importance of the experts after decoupling; Indicates the first The first target threat Observed values ​​of the evaluation indicators; q Indicates the first q One expert; The expression for the subjective weight optimization model includes: （6） in, The expert panel indicated that it had no opinion on the matter. i Preference for each evaluation indicator; st It means "bound to" or "subject to". Indicates the first j Initial indicator weights for each evaluation indicator; Indicates Hamming distance; This represents a subjective weight optimization model; m Indicates the number of evaluation indicators; n Represents the number of target threats; for the subjective weight optimization model Minimize the result to obtain min. ; Construct a Lagrange function and use the Lagrange conditional extremum method to solve the subjective weight optimization model to obtain the subjective weight; The expression for the Lagrange function includes: （7） in, This represents a subjective weight optimization model; Represents the Lagrange multiplier operator; Indicates the first j Initial indicator weights for each evaluation indicator; The expressions utilizing the subjective weights include: （8） in, Indicates subjective weight; -1 indicates the reciprocal; d Indicates the degree of preference.

3. The data-optimized weighted target dynamic threat assessment method according to claim 2, characterized in that, The steps of constructing the objective weight optimization model based on the total deviation between the target threat and the positive and negative ideal targets, and the principle of maximum entropy, and solving the objective weight optimization model to obtain the objective weights include: Construct a scheme to determine the total deviation between the target threat and the positive and negative ideal targets; The expressions for the positive and negative ideal targets include: （9） The expression for the total deviation between the target threat and the positive and negative ideal targets includes: （10） in, Indicates the ideal goal. ; Indicates a negative ideal goal. ; The expression of the objective weight optimization model includes: ， （11） in, Indicates the total deviation. ; Represents maximum entropy. , ; This represents the balance coefficient among multiple evaluation indicators. ; Construct a Lagrange function and use the Lagrange conditional extremum method to solve the objective weight optimization model to obtain the objective weight; The expression for the objective weight includes: （12）。 4. The dynamic threat assessment method for targets based on data optimization and weighted comprehensive assessment as described in claim 3, characterized in that, In the step of processing the subjective weights and objective weights using a linear weighting method to obtain the comprehensive weights... The expression for the comprehensive weight includes: （13） in, This represents a balancing factor between subjective weights. This represents a balancing factor among objective weights. ; Indicates subjective weighting; Indicates objective weighting. This indicates the overall weight.

5. The dynamic threat assessment method for targets based on data optimization and weighted comprehensive assessment as described in claim 4, characterized in that, The step of dynamically assessing the target threat using the comprehensive weights and the normalized decision matrix to obtain the dynamic assessment result of the target threat includes: The normalized decision matrix is ​​weighted using the comprehensive weights to obtain a comprehensive weighted decision matrix; The comprehensive weighted decision matrix is ​​processed using the TOPSIS algorithm or the VIKOR algorithm, and the result is determined by relative proximity. Or threaten a compromise value, To characterize the threat level of a target at a single moment; The weights of the time series can be obtained using the inverse form of the Poisson distribution or the normal cumulative distribution method. The threat level of a target at a single moment is weighted using the time series weights to obtain a comprehensive weighted dynamic threat level; The target threats are ranked according to the relative magnitude of the comprehensive weighted dynamic threat level, and the ranking result is the dynamic assessment result of the target threats.

6. The dynamic threat assessment method for targets based on data optimization and weighted comprehensive assessment as described in claim 5, characterized in that, In the step of weighting the normalized decision matrix using the comprehensive weights to obtain the comprehensive weighted decision matrix... The expression for the comprehensive weighted decision matrix includes: （14） in, This represents the value of the weighted index; In the step of obtaining the time series weights using the Poisson distribution method or the normal cumulative distribution method... The expression for obtaining the time series weights using the Poisson distribution method includes: （15） in, , , ; Represents the parameters of the Poisson distribution; The expression for obtaining the time series weights using the normal cumulative distribution method includes: （16） in, Represents the normal cumulative distribution function. ; t Indicates duration; express k The time weight of the i-th time step; This represents the expectation of a normal distribution. ; The standard deviation represents the normal distribution. ; ; In the step of weighting the target threat level at a single moment using the time series weights to obtain the comprehensive weighted dynamic threat level, If the TOPSIS algorithm and the Poisson distribution method are used, the comprehensive weighted dynamic threat expression includes: （17） If the TOPSIS algorithm and the normal cumulative distribution method are used, the comprehensive weighted dynamic threat expression includes: （18） If the VIKOR algorithm and the Poisson distribution method are used, the comprehensive weighted dynamic threat expression includes: （19） If the VIKOR algorithm and the normal cumulative distribution method are used, the comprehensive weighted dynamic threat expression includes: （20） in, express The relative proximity of moments; express The compromise value of threats at any given moment; This represents the weights of the time series using the Poisson distribution method; Indicates the first k Time weight of each moment; Represents the Lagrange multiplier operator; p This indicates the number of experts.

7. A dynamic threat assessment system based on data optimization weighted comprehensive weighting, applied to the data optimization weighted comprehensive weighted target dynamic threat assessment method as described in claim 1, characterized in that, include: The evaluation index processing module, evaluation index optimization weight module, target threat dynamic evaluation module, display module, and storage module are included. The evaluation index processing module is interactively connected to the command system or sensor network, as well as the evaluation index optimization weight module; the evaluation index processing module is used to obtain multiple evaluation indicators in the target threat, and process the multiple evaluation indicators to obtain a normalized decision matrix. The evaluation index optimization weight module is interactively connected to the evaluation index processing module and the target threat dynamic evaluation module, respectively. The evaluation index optimization weight module includes a subjective weight submodule, an objective weight submodule and a comprehensive weight submodule. The subjective weight submodule is used to generate subjective weights, the objective weight submodule is used to generate objective weights, and the comprehensive weight submodule is used to linearly weight the subjective weights and the objective weights to obtain the comprehensive weight. The target threat dynamic assessment module uses the normalized decision matrix and the comprehensive weight to dynamically assess the target threat; The display module is used to display information from the evaluation index processing module, the evaluation index optimization weight module, and the target threat dynamic evaluation module. The storage module is used to store and statistically analyze information from the evaluation index processing module, the evaluation index optimization weight module, and the target threat dynamic evaluation module.

Images