CN104899450A - Method for judging tuple linguistic model through comprehensive evaluation of target threat level - Google Patents

Abstract

The invention relates to a method for judging a tuple linguistic model through comprehensive evaluation of a target threat level. The method is realized according to the following methods: through uj<[1,h]aT[uj]=(ek, ajk)<Ex[-0.5, 0.5), obtaining a tuple linguistic model (ek, ajk) of a threat level judging value vj of a target Tj in a target set T; according to the tuple linguistic model corresponding to every target, determining the threat level of every target and determining a target threat sequence of all centralized targets according to the order of its tuple linguistic model from big to small. The target thread grade proposed by the invention comprehensively evaluates the grade judging value definition and the tuple linguistic concept, so as to further set up the tuple linguistic judging model and method of the corresponding target threat level comprehensive evaluation, so as to provide theoretical basis for the signifying operation of the target thread level comprehensive evaluation result, facilitate the implementation by computer, and provide the theoretical basis and intelligent support to guide the decision making; moreover, the method can be expanded and applied to multiattribute or rule group classification decision-making issues in economical management, information system management, information security, and other domains.

Description

A kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method

Technical field

The present invention relates to Object Threat Evaluation, particularly a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method.

Background technology

Object Threat Evaluation (Threat Assessment), sometimes also referred to as threat estimating or Threat verdict, is divided into target threat sequencing to assess the large class with target danger level comprehensive assessment two.Object Threat Evaluation and situation of battlefield are assessed closely related, its main task is the size or the threat level that go out unfriendly target threaten degree according to various inferred from input data, it is the important component part of Combat Command System, to directly have influence on Target Assignment and application of fire, be related to the performance of troop operation effect and fighting capacity ^[1-8].Since own mankind's war, Object Threat Evaluation is exactly one of problem of commander and military specialist's special concern always.May due to the war in past comparatively speaking, its complexity, scale, pace of change, technology content etc. are lower, rely on the knowledge of commander and military specialist, experience just can make target threat judgement, and then command troops.But along with the extensive utilization of a large amount of high and new technology in modern war, particularly in air defense operation, the modern times air strike environment faced becomes increasingly complex, the utilization of the air strike such as full spatial domain, multi rack secondary, multiple batches of, multi-direction, multi-level, continuous saturation attack mode pattern, have higher requirement to the correctness of air defense operational control decision-making and real-time, only the knowledge of Drawing upon commander and military specialist, experience carry out the requirement that judges to be difficult to even cannot meet modern operation with decision-making.Therefore, in recent years, one of very important active research content in military field was just become as the Object Threat Evaluation problem of one of basic decision in fight activity, particularly air defense operation.

A lot of research is all carried out to Object Threat Evaluation both at home and abroad, but owing to relating to military secret, be difficult in disclosed documents and materials, see the analysis of real Object Threat Evaluation problem case, great majority still concentrate in the research of theoretical model and method.We study discovery, the factor affecting targets'threat is a lot, comprise target type, apart from by the distance of defendance object, highly, speed, lateral range, interference performance, quantity etc., and the impact between these factors is all often mutual conflict, incommensurable.Therefore, Object Threat Evaluation can be summarized as the many attributes of a class or Multifactor Comprehensive Evaluation problem in essence ^[1].At present, target threat sequencing evaluation problem is studied often, and to target danger level comprehensive assessment problem, from existing documents and materials, most research is all assessed itself and target threat sequencing and is used as same class problem and solves, namely, after acquisition target threat integrated value, according to the threat level threshold value of setting, judge whether target threat integrated value meets threshold condition to determine the threat level of target.But, in actual operational commanding decision-making, for reaching the requirement of high-speed decision and shooting, a lot of ground to air missile weapon, particularly as short range ground to air missile weapon such as " rattle snakes ", what more pay close attention to is target danger level, and discusses angle from the concept and methodology of target danger level comprehensive assessment, and target danger level comprehensive assessment is also different from target threat sequencing assessment.For this reason, the present invention will from target danger level comprehensive assessment concept, set up the general model of target danger level comprehensive assessment, and its rationality, validity is analyzed in mathematics aspect, be intended to disclose in target danger level comprehensive assessment and occur " outwardly like science; and be actually not science or even pseudoscience " method, avoid having occurred commanding and decision-making error and " fail to see what Lushan Mountain really looks like ", the function and position of reduction Object Threat Evaluation in Combat Command System, promote quality and the confidence level of Object Threat Evaluation.

Summary of the invention

The object of the present invention is to provide a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method, set up the general model of target danger level comprehensive assessment problem, and provide the Fuzzy Linear method of weighting of its grade comprehensive assessment.

For achieving the above object, technical scheme of the present invention is: a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method, is characterized in that, realizes in accordance with the following steps:

Step S1: utilize invertible mapping τ: [1, h] → E × [-0.5,0.5), pass through υ _j∈ [1, h] a τ (υ _j)=(e _k, α _jk) ∈ E × [-0.5,0.5) obtain target T in object set T _jthreat level decision content υ _jtwo-tuple Linguistic Information Processing (e _k, α _jk), and according to described Two-tuple Linguistic Information Processing (e _k, α _jk) determine target T _jthreat level e _k, wherein, target danger level k=Round (υ _j), deviate α _jk=υ _j-k, E are threat level collection, and Round is round function, T={T ₁, T ₂..., T _nbe object set, 1≤j≤n, 1≤k≤h, n and h is the positive integer being more than or equal to 1;

Step S2: the Two-tuple Linguistic Information Processing corresponding according to each target, and the target threat sequencing determining all targets of target tightening by its Two-tuple Linguistic Information Processing order from big to small.

In an embodiment of the present invention, described threat level collection E is the ordered set comprising h threat level, i.e. an E={e ₁, e ₂..., e _h; Wherein, e _kfor a kth threat level, and e ₁>e ₂> ... >e _h, i.e. a kth threat level e _kthan kth+1 threat level e _k+1threat level high.

In an embodiment of the present invention, described threat level collection E is a language phrase collection comprising h language phrase, all corresponding language phrase of each threat level, and threat level e _kfor language phrase " threat of k level ".

In an embodiment of the present invention, target T _jthreat level decision content υ _jobtain in the following way:

${\mathrm{\&upsi;}}_j=(1,2,...,h){\left(U_j\right)}^T=\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{ku}}_{jk}$

Wherein, U _jfor target T _jfor all threat level e _kcomprehensis pertaining vector, i.e. U _j=(u _j1, u _j2..., u _jh), wherein, u _jk∈ [0,1] and due to obtain 1≤υ _j≤ h, i.e. target T _jthreat level decision content υ _jit is the real number between threat level 1 and h.

In an embodiment of the present invention, in described step S1, if target T _jthreat level decision content υ _jsatisfy condition k-0.5≤υ _j<k+0.5, then evaluating target T _jfor k level threatens target, i.e. target T _jthreat level be e _k, complete described Two-tuple Linguistic Information Processing (e _k, α _jk) middle e _kacquisition; Described Two-tuple Linguistic Information Processing (e _k, α _jk) in α _jkrepresent threat level decision content υ _jwith the deviate being obtained target danger level k by this threat level decision content, and represent overgauge and minus deviation with positive sign, negative sign respectively, and α _jk∈ [-0.5,0.5), complete described Two-tuple Linguistic Information Processing (e _k, α _jk) middle α _jkacquisition.

In an embodiment of the present invention, described Comprehensis pertaining vector U _j=φ (ω, μ _j), wherein, φ is a target danger level comprehensive assessment Aggregation Operators, and φ: [0,1] ^2m+h→ [0,1] ^h, φ (ω, μ _j)=(φ (ω, μ _j1), φ (ω, μ _j2), L, φ (ω, μ _jh)), u _jk=φ (ω, μ _jk) ∈ [0,1], μ _jfor target T _jstandardization threatening factors level characteristics value matrix, ω=(ω ₁, ω ₂..., ω _m) ^tthe weight vectors that threatening factors collection F is corresponding, described threatening factors collection F={f ₁, f ₂..., f _m, m be more than or equal to 1 positive integer.

In an embodiment of the present invention, following linear weighted function Integrated Evaluation Model is adopted to represent described Comprehensis pertaining vector U _j:

$U_j={\left(\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}\right)}_{1\&times;h}=(\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{j1}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}},\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{j2}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}},\mathrm{...},\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jh}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}),$ Wherein, $u_{jk}=\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}.$

In an embodiment of the present invention, described standardization threatening factors level characteristics value matrix μ _j=(μ _ijk) _{m × h}, and μ _jk=(μ _1jk, μ _2jk..., μ _mjk) ^tfor target T _jabout threat level e _kthe standardization feature value vector of all threatening factors; Described standardization threatening factors level characteristics value matrix μ _jobtained by following normalized transformation: y _ijk∈ Ra μ _ijk=φ ₁(y _ijk) ∈ [0,1], wherein, φ ₁a threatening factors rank feature values normalized transformation operator, and φ ₁: R → [0,1], R is set of real numbers, y _ijkfor target danger level eigenwert, i.e. this y _ijkfor target T _jabout threatening factors f _ithreat level e _keigenwert, y _ijk=f _ik(T _j).

In an embodiment of the present invention, in described step S2, as the threat level e of target _ktime different, the target that threat level is high, i.e. e _kthe target large or k is little, before its target threat sequencing comes; When threat level is identical, deviate α _jklittle target, before its target threat sequencing comes.

Compared to prior art, the present invention has following beneficial effect: a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method proposed by the invention, describe on the basis of target danger level Integrated Evaluation Model in system, propose rank decision content and the Two-tuple Linguistic Information Processing concept of target danger level comprehensive assessment, and establish corresponding target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method.The method not only reasonably can determine target danger level, and namely the threat size can also distinguishing identical threat level target sorts.By the representation of Two-tuple Linguistic Information Processing, the symbolism computing that can be realize target threat level comprehensive assessment result is provided fundamental basis, and is convenient to computing machine and realizes, to raising operational commanding level of decision-making and ageingly have important references and be worth.

In addition, because target danger level comprehensive assessment is the special many attributes of a class (or targets, criterion, factor, index) comprehensive evaluation (or decision-making) problem, the target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method that the present invention sets up can expansive approach to politics, multiple criteria group classification decision problem in the fields such as economy and society, such as, the predicting corporate failures in Financial Management field, credit rating of enterprise comprehensive evaluation and country risk rating comprehensive assessment, environment, the national energy policy analysis in the energy and ecomanagement field, natural gas line venture analysis and water resources management, the comprehensive assessment of matter input protective development grade, information management system, information security, human resource management field, marketing field and service outsourcing field etc.

Accompanying drawing explanation

Fig. 1 is the process flow diagram of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method in the present invention.

The rank decision content list of enemy's target of air attack that Fig. 2 adopts maximum membership grade principle to obtain for tradition.

Fig. 3 is the threat level comprehensive assessment result of enemy's target of air attack that adopts target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method in the present invention and obtain and sorted lists.

Embodiment

Below in conjunction with accompanying drawing, technical scheme of the present invention is specifically described.

The invention provides a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method, as shown in Figure 1, realize in accordance with the following steps:

Step S1: utilize invertible mapping τ: [1, h] → E × [-0.5,0.5), pass through υ _j∈ [1, h] a τ (υ _j)=(e _k, α _jk) ∈ E × [-0.5,0.5) obtain target T in object set T _jthreat level decision content υ _jtwo-tuple Linguistic Information Processing (e _k, α _jk), and according to described Two-tuple Linguistic Information Processing (e _k, α _jk) determine target T _jthreat level e _k, wherein, target danger level k=Round (υ _j), deviate α _jk=υ _j-k, E are threat level collection, and Round is round function, T={T ₁, T ₂..., T _nbe object set, 1≤j≤n, 1≤k≤h, n and h is the positive integer being more than or equal to 1;

Obviously, the inverse mapping τ of τ ^-1: E × [-0.5,0.5) → [1, h] make

(e _k,α _jk)∈E×[-0.5,0.5)aτ ^-1(e _k,α _jk)＝υ _j∈[1,h]

Step S2: the Two-tuple Linguistic Information Processing corresponding according to each target, determines the target threat sequencing of all targets of target tightening by its Two-tuple Linguistic Information Processing order from big to small; In the present embodiment, as the threat level e of target _ktime different, the target that threat level is high, i.e. e _kthe target large or k is little, before its target threat sequencing comes; When threat level is identical, deviate α _jklittle target, before its target threat sequencing comes.Such as, if (e _k, α _jk), (e _l, α _il) be target T respectively _jand T _irank decision content υ _j, υ _itwo-tuple Linguistic Information Processing, specify that its size is as follows:

(A) if e _k>e _l(or k<l), then (e _k, α _jk) > (e _l, α _il);

(B) if e _k=e _l(or k=l), then as 1. α _jk=α _iltime, (e _k, α _jk)=(e _l, α _il); 2. α _jk> α _iltime, (e _k, α _jk) < (e _l, α _il); 3. α _jk< α _iltime, (e _k, α _jk) > (e _l, α _il).

Situation (A) represents, if target danger level is high, no matter then its deviate, the Two-tuple Linguistic Information Processing of its respective objects threat level decision content is large.And situation (B) represents, under the condition that target danger level is identical, the size of the Two-tuple Linguistic Information Processing of target threat rank decision content depends on deviate, and deviate is less, and its corresponding Two-tuple Linguistic Information Processing is larger.So, can according to Two-tuple Linguistic Information Processing (e _k, α _jk) to target T _j(j=1,2 ..., the grade that n) impends Comprehensive Assessment and the sequence of threat size.

Further, in the present embodiment, described threat level collection E is the ordered set comprising h threat level, i.e. an E={e ₁, e ₂..., e _h; Wherein, e _kfor a kth threat level, and e ₁>e ₂> ... >e _h, i.e. a kth threat level e _kthan kth+1 threat level e _k+1threat level high.And in the present embodiment, described threat level collection E is a language phrase collection comprising h language phrase, all corresponding language phrase of each threat level, and threat level e _kfor language phrase " threat of k level ".Such as, the threat level collection E={e of 13 grade ₁, e ₂, e ₃may be defined as: language phrase e ₁=" 1 grade of threat " namely " threatens very large ", e ₂=" 2 grades of threats " namely " threatens general ", e ₃=" 3 grades of threats " namely " threatens very little ".Usually, threat level collection E={e ₁, e ₂..., e _horder and maximization, minimization computing etc. should be met.

The concept of Two-tuple Linguistic Information Processing appears in uncertainty decision problem the earliest, is now developed, is applied to and carries out target danger level assessment.In the present embodiment, in described step S1, if target T _jrank decision content be υ _j∈ [1, h], then its Two-tuple Linguistic Information Processing can be expressed as binary ordered group (e _k, α _jk), wherein, e _k∈ E and α _jk∈ [-0.5,0.5), concrete meaning can be described as: (1) e _kthe E={e of prior defined good order ₁, e ₂..., e _hin " threat of k level ", represent by rank decision content υ _jobtain, with the threat level pressed close to most in E; (2) α _jkis-symbol branch value, represents υ _jwith target T therefrom _jthe deviate of threat level comprehensive assessment result and k.

In the present embodiment, if target T _j(j=1,2 ..., rank decision content υ n) _jsatisfy condition

k-0.5≤υ _j<k+0.5

Then evaluating target T _jfor the target that k level threatens, i.e. target T _jthreat level be e _k.Belonging to the extent of deviation of k level threat in order to portray target, being expressed as α by deviate _jk=υ _j-k.Obviously ,-0.5≤α _jk<0.5.α _jkimplication be very clearly.If α _jk∈ [-0.5,0) larger (namely | α _jk| less), then υ _j∈ (k-1, k) is larger, and more close to k, its respective objects T _jalso the threat of k level is reduced to gradually by belonging to the threat of k-1 level.Otherwise, if α _jk∈ (0,0.5) is less, then υ _j∈ (k, k+1) is less, and more close to k, corresponding target T _jthe degree belonging to the threat of k level is also larger.If α _jk=0, then υ _j=k, its respective objects T _jjust in time belong to k level to threaten.

Further, in the present embodiment, target T _jthreat level decision content υ _jobtain in the following way:

${\mathrm{\&upsi;}}_j=(1,2,...,h){\left(U_j\right)}^T=\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{ku}}_{jk}$

Wherein, U _jfor target T _jfor all threat level e _k(k=1,2 ..., Comprehensis pertaining vector h), U _j=(u _j1, u _j2..., u _jh), wherein, u _jk∈ [0,1] and due to 1≤υ can be obtained _j≤ h, i.e. υ _j1 nondimensional quantitative index, between 1 grade of (i.e. e ₁) threaten and h level (i.e. e _h) between threat.

And in traditional target danger level comprehensive assessment, usually utilize maximum membership grade principle, namely by

$u_{j\stackrel{\&OverBar;}{k}}=\underset{1\&le;k\&le;h}{max}\left\{u_{jk}\right\}$

Evaluating target T _jbelong to individual threat level namely in actual computation process, although target T _jrank decision content υ _jbe the real number between 1 and h, but be not in most cases integer, and under maximum membership grade principle background, a kind of nature of employing, simple method are: if

[υ _j]＝k

Then by target T _jbe assessed as the threat of k level and e _k, wherein [υ _j] represent be not more than υ _jmaximum integer.

In the present embodiment, target danger level comprehensive assessment problem is below considered: suppose there are 5 crowdes of enemy's target of air attack T _j(j=1,2 ..., 5), by formula calculate its rank decision content, as shown in Figure 2 the rank decision content of enemy's target of air attack.

By Fig. 1 and Shi [υ _j]=k can obtain:

[υ ₁]＝1，[υ ₂]＝[υ ₃]＝[υ ₄]＝2，[υ ₅]＝3

Thus enemy's target of air attack T can be made _j(j=1,2 ..., 5) threat level Comprehensive Assessment result be: T ₁belong to 1 grade and threaten target, T ₂, T ₃and T ₄belong to 2 grades and threaten target, T ₅belong to 3 grades and threaten target.But such conclusion seems inconsistent with intuitive judgment.In fact, carefully analyze discovery, υ ₁=1.98 and υ ₂=2.03 are all comparatively close to 2, and differ very little between the two, υ ₁with υ ₂corresponding target T ₁and T ₂be assessed as 1 grade of threat respectively, 2 grades of threats are irrational.Intuitively, T ₁and T ₂identical threat level that is the 2nd grade should be belonged to.Same analysis can be seen, υ ₂=2.03, υ ₃=2.34 and υ ₄=2.98 be all be less than 3 number, but their difference is obvious: υ ₂=2.03 and 2 closely, and its with 2 difference degree compare υ ₃the difference degree of=2.34 and 2 is little, compares υ ₄the difference degree of=2.98 and 2 is less.In addition, υ ₄=2.98 are comparatively close to 3.Therefore, υ ₂, υ ₃and υ ₄corresponding target T ₂, T ₃and T ₄it is irrational for being all assessed as 2 grades of threats.Intuitively, T ₂and T ₃2 grades should be belonged to and threaten target, and T ₄3 grades should be belonged to and threaten target.Further, T ₁, T ₂and T ₃all belong to 2 grades and threaten target, but their subjection degree is not exclusively the same.Target T ₄and T ₅situation also similar.Therefore, be necessary, on the basis of evaluation each batch of target danger level, to distinguish the difference degree between each batch of target in identical threat level.

And by adopting the general model of target danger level comprehensive assessment problem proposed by the invention, then consider 5 batches of enemy's Air attack threat grade comprehensive assessment problems in Fig. 2, obtain target T _j(j=1,2,3,4,5) threat level decision content υ _jtwo-tuple Linguistic Information Processing; Utilize concept and the size comparison of Two-tuple Linguistic Information Processing, the threat level of all 5 batches of targets can be determined and threaten size sequence, the threat level comprehensive assessment result of enemy's target of air attack and sequence listed by Fig. 3.As seen from Figure 2, according to the target danger level of Two-tuple Linguistic Information Processing evaluation and by formula [υ _jit is different that]=k evaluates, but intuitively, the former result is relatively more rational, and matches with analysis above.Therefore, these 5 crowdes of enemy's target of air attack T _jthe threat level comprehensive assessment result of (j=1,2,3,4,5) should be: T ₁, T ₂and T ₃belong to 2 grades and threaten target, T ₄and T ₅belong to 3 grades and threaten target, and target threat sequencing is T ₁> T ₂> T ₃> T ₄> T ₅.

Can find out significantly from said process, utilize maximum membership grade principle evaluating target threat level, sometimes may there are some irrational phenomenons, trace it to its cause and be that maximum membership grade principle only considers the relative size of degree of membership, do not consider the position difference of each threat level.And the general model of target danger level comprehensive assessment problem proposed by the invention and the Fuzzy Linear method of weighting of grade comprehensive assessment thereof, threat level decision content υ _jreflect the information of target danger level Comprehensis pertaining and threat level (i.e. level position) 2 aspects, use υ _jevaluating target T _j(j=1,2 ..., threat level n) want Billy with maximum membership grade principle more comprehensively, more objective.Usually, threat level judges υ _jless, target T _j(j=1,2 ..., threat level n) is higher.

Further, in the present embodiment, described Comprehensis pertaining vector U _j=φ (ω, μ _j), wherein, φ is a kind of target danger level comprehensive assessment Aggregation Operators, and φ: [0,1] ^2m+h→ [0,1] ^h, φ (ω, μ _j)=(φ (ω, μ _j1), φ (ω, μ _j2), L, φ (ω, μ _jh)), u _jk=φ (ω, μ _jk) ∈ [0,1], μ _jfor target T _jstandardization threatening factors level characteristics value matrix, ω=(ω ₁, ω ₂..., ω _m) ^tthe weight vectors that threatening factors collection F is corresponding, described threatening factors collection F={f ₁, f ₂..., f _m, m be more than or equal to 1 positive integer.In the present embodiment, U _j=(u _j1, u _j2..., u _jh) can membership vector be seen as, and the threat level Comprehensis pertaining vector matrix form that the n in object set T criticizes target can be expressed as U=(u intuitively _jk) _{n × h}, be called target danger level Comprehensis pertaining matrix.

Further, in the present embodiment, following linear weighted function Integrated Evaluation Model is adopted to represent described Comprehensis pertaining vector U _j:

$U_j={\left(\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}\right)}_{1\&times;h}=(\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{j1}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}},\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{j2}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}},\mathrm{...},\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jh}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}),$ Wherein, $u_{jk}=\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}.$

Further, in the present embodiment, described standardization threatening factors level characteristics value matrix μ _j=(μ _ijk) _{m × h}, and μ _jk=(μ _1jk, μ _2jk..., μ _mjk) ^tfor target T _jabout threat level e _kthe standardization feature value vector of all threatening factors, the division of threat level is a fuzzy concept comprising larger subjectivity.Therefore, μ _j1 fuzzy relation between F and E can be regarded as, thus μ _ijkbe exactly T _jabout f _ito threat level e _kdegree of membership, μ _jkthen T _jabout e _kthe membership vector of all threatening factors.Also μ _jkbe called T _jthreatening factors level assessment vector, μ _jbe called T _jthreatening factors level assessment matrix.

Further, in the present embodiment, described standardization threatening factors level characteristics value matrix μ _jobtained by following normalized transformation: y _ijk∈ Ra μ _ijk=φ ₁(y _ijk) ∈ [0,1], wherein, φ ₁a kind of threatening factors rank feature values normalized transformation operator, and φ ₁: R → [0,1], R is set of real numbers, y _ijkfor target danger level eigenwert, i.e. this y _ijkfor target T _jabout threatening factors f _ithreat level e _keigenwert, y _ijk=f _ik(T _j), Y can be expressed as intuitively with matrix form _j=(y _ijk) _{m × h}, be called target T _jthreat level eigenvalue matrix.At matrix Y _jin, the i-th row represents target T _jabout threatening factors f _ithe eigenwert of all threat levels, and target T is shown in kth list _jabout threat level e _kthe eigenwert of all threatening factors, be designated as vectorial y _jk=(y _1jk, y _2jk..., y _mjk) ^t(k=1,2 ..., h).But, due to target danger level eigenwert y _ijktype normally different, therefore type unification and nondimensionalization process need be done to it, namely adopt y _ijk∈ Ra μ _ijk=φ ₁(y _ijk) ∈ [0,1] carries out normalized transformation to it.

Be more than preferred embodiment of the present invention, all changes done according to technical solution of the present invention, when the function produced does not exceed the scope of technical solution of the present invention, all belong to protection scope of the present invention.

Claims (9)

1. a target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method, is characterized in that, realizes in accordance with the following steps:

Step S1: utilize invertible mapping τ: [1, h] → E × [-0.5,0.5), pass through υ _j∈ [1, h] a τ (υ _j)=(e _k, α _jk) ∈ E × [-0.5,0.5) obtain target T in object set T _jthreat level decision content υ _jtwo-tuple Linguistic Information Processing (e _k, α _jk), and according to described Two-tuple Linguistic Information Processing (e _k, α _jk) determine target T _jthreat level e _k, wherein, target danger level k=Round (υ _j), deviate α _jk=υ _j-k, E are threat level collection, and Round is round function, T={T ₁, T ₂..., T _nbe object set, 1≤j≤n, 1≤k≤h, n and h is the positive integer being more than or equal to 1;

Step S2: the Two-tuple Linguistic Information Processing corresponding according to each target, and the target threat sequencing determining all targets of target tightening by its Two-tuple Linguistic Information Processing order from big to small.

2. a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method according to claim 1, it is characterized in that, described threat level collection E is the ordered set comprising h threat level, i.e. an E={e ₁, e ₂..., e _h; Wherein, e _kfor a kth threat level, and e ₁>e ₂> ... >e _h, i.e. a kth threat level e _kthan kth+1 threat level e _k+1threat level high.

3. a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method according to claim 2, it is characterized in that, described threat level collection E is a language phrase collection comprising h language phrase, all corresponding language phrase of each threat level, and threat level e _kfor language phrase " threat of k level ".

4. a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method according to claim 1, is characterized in that, target T _jthreat level decision content υ _jobtain in the following way:

${\mathrm{\&upsi;}}_j=(1,2,\mathrm{...},h){\left(U_j\right)}^T=\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{ku}}_{jk}$

Wherein, U _jfor target T _jfor all threat level e _kcomprehensis pertaining vector, i.e. U _j=(u _j1, u _j2..., u _jh), wherein, u _jk∈ [0,1] and due to obtain 1≤υ _j≤ h, i.e. target T _jthreat level decision content υ _jit is the real number between threat level 1 and h.

5. a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method according to claim 2 or 4, is characterized in that, in described step S1, if target T _jthreat level decision content υ _jsatisfy condition k-0.5≤υ _j<k+0.5, then evaluating target T _jfor k level threatens target, i.e. target T _jthreat level be e _k, complete described Two-tuple Linguistic Information Processing (e _k, α _jk) middle e _kacquisition; Described Two-tuple Linguistic Information Processing (e _k, α _jk) in α _jkrepresent threat level decision content υ _jwith the deviate being obtained target danger level k by this threat level decision content, and represent overgauge and minus deviation with positive sign, negative sign respectively, and α _jk∈ [-0.5,0.5), complete described Two-tuple Linguistic Information Processing (e _k, α _jk) middle α _jkacquisition.

6. a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method according to claim 4, is characterized in that, described Comprehensis pertaining vector U _j=φ (ω, μ _j), wherein, φ is a target danger level comprehensive assessment Aggregation Operators, and φ: [0,1] ^2m+h→ [0,1] ^h, φ (ω, μ _j)=(φ (ω, μ _j1), φ (ω, μ _j2), L, φ (ω, μ _jh)), u _jk=φ (ω, μ _jk) ∈ [0,1], μ _jfor target T _jstandardization threatening factors level characteristics value matrix, ω=(ω ₁, ω ₂..., ω _m) ^tthe weight vectors that threatening factors collection F is corresponding, described threatening factors collection F={f ₁, f ₂..., f _m, m be more than or equal to 1 positive integer.

7. a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method according to claim 6, is characterized in that, adopts following linear weighted function Integrated Evaluation Model to represent described Comprehensis pertaining vector U _j:

$U_j={\left(\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}\right)}_{1\&times;h}=(\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{j1}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}},\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{j2}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}},\mathrm{...},\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jh}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}),$ Wherein, $u_{jk}=\frac{{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}{\underset{k=1}{\overset{h}{\&Sigma;}}{\mathrm{\&omega;}}^T{\mathrm{\&mu;}}_{jk}}.$

8. a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method according to claim 6, is characterized in that, described standardization threatening factors level characteristics value matrix μ _j=(μ _ijk) _{m × h}, and μ _jk=(μ _1jk, μ _2jk..., μ _mjk) ^tfor target T _jabout threat level e _kthe standardization feature value vector of all threatening factors; Described standardization threatening factors level characteristics value matrix μ _jobtained by following normalized transformation: y _ijk∈ R a μ _ijk=φ ₁(y _ijk) ∈ [0,1], wherein, φ ₁a threatening factors rank feature values normalized transformation operator, and φ ₁: R → [0,1], R is set of real numbers, y _ijkfor target danger level eigenwert, i.e. this y _ijkfor target T _jabout threatening factors f _ithreat level e _keigenwert, y _ijk=f _ik(T _j).

9. a kind of target danger level comprehensive assessment Two-tuple Linguistic Information Processing decision method according to claim 1, is characterized in that, in described step S2, as the threat level e of target _ktime different, the target that threat level is high, i.e. e _kthe target large or k is little, before its target threat sequencing comes; When threat level is identical, deviate α _jklittle target, before its target threat sequencing comes.