CN104809235A - Program evaluation system and program evaluation method - Google Patents

Abstract

The invention provides a program evaluation system and a program evaluation method. The program evaluation system comprises an input unit, a program index attribute space construction unit, a weighted index space construction unit, a program evaluation model construction unit, an optimal solution determination unit and an obtainment unit; the input unit is used for inputting weighting methods, programs to be evaluated and evaluation indexes; the program index attribute space construction unit obtains a decision making matrix according to the programs to be evaluated and an evaluation index calculation formula; the weighted index space construction unit calculates the weight of the evaluation index of each program according to the weighting methods and multiplies the weight by a linear expression coefficient of each weighting method, so that a compound weight vector is obtained; the program evaluation model construction unit utilizes the decision making matrix and the compound weight vectors to construct target optimization models according to the evaluation method; the optimal solution determination unit adopts the Lagrangian multiplier method to solve optimal solutions of the program evaluation models; the obtainment unit obtains an optimal linear expression coefficient vector and an optimal compound weight vector of each evaluation index of each program according to the audience information and optimal solutions of the programs to be evaluated, and thereby a comprehensive evaluation value of each program is obtained. The invention reflects objective and subjective decision making at the same time, realizes the scientific weighting of evaluation indexes, and increases the accuracy of evaluation.

Description

A kind of program evaluation system and method

Technical field

The present invention relates to field of broadcast televisions, more specifically, relate to a kind of program evaluation system and method.

Background technology

In field of broadcast televisions, usually adopting multiple metrics evaluation user to the preference of program, is an important step of multiobjectives decision to the assignment of evaluation criterion weight.The weight of index refers to the reflection being marked on different significance level in evaluation procedure, is a kind of subjective assessment of index relative importance in decision-making (or assessment) problem and the comprehensive measurement that objectively responds.Whether reasonable the assignment of weight is, plays vital effect to the scientific rationality of evaluation result; If the weight of a certain factor changes, whole evaluation result will be affected.Therefore, the assignment of weight must accomplish science and objective, and this just requires to seek suitable Weight Determination.

The domestic and international defining method about evaluation index weight coefficient has a variety of at present, and according to raw data source and the difference of computation process when calculating weight coefficient, these methods roughly can divide subjective weighting method and the large class of objective weighted model two.Subjective weights Evaluation Method takes method qualitatively, rule of thumb carry out subjective judgement by expert and obtain flexible strategy, and then comprehensive assessment is carried out to index, as analytical hierarchy process, expert survey, Fuzzy Analysis Method, binomial coefficient method etc., wherein, analytical hierarchy process is the maximum method used in practical application, and it is by challenge stratification, by qualitative question quantification.Objective Weight assessment rule carries out comprehensive assessment according to the relation of the correlationship between historical data research index or index and assessment result, mainly contain the methods such as Information Entropy, principal component analysis (PCA), average variance method, VC Method, wherein, Information Entropy more, the data that this enabling legislation uses are decision matrixs, and determined attribute weight reflects the dispersion degree of property value.Subjective weighting method can embody the micro-judgment of decision maker, and the relative importance of attribute generally can not violate the general knowledge of people.But its randomness is comparatively large, accuracy of determination and less reliable.There is the objective standard of the power of tax in objective weighted model, can utilize certain mathematical model, by calculating the weight coefficient of attribute.Its shortcoming is the subjective preference information such as Subjective Knowledge and experience ignoring decision maker, occurs the irrational phenomenon of weight coefficient sometimes.

Summary of the invention

In view of the above problems, the object of this invention is to provide a kind of reflection subjective decision and objective making decision, realize program evaluation system and method that TV programme evaluation index science composes power.

According to an aspect of the present invention, provide a kind of program evaluation system, comprising: input block, for inputting tax power method, program to be evaluated and evaluation index, wherein, described tax power method comprises subjective weights method and Objective Weighting; Program Criterion Attribute space construction unit, obtains the decision matrix of each evaluation index composition of each program, i.e. program Criterion Attribute space according to the program to be evaluated of input block input and the computing formula of evaluation index; Compose power index space construction unit, according to the tax power method of input block input calculate each evaluation index of each program weight and with the Linearly Representation coefficient of each tax power method, be multiplied and obtain right vector, namely power index space is composed, wherein, relevant to described decision matrix according to the weight vectors that Objective Weighting obtains, right vector F

F＝WΘ＝[F ₁F ₂… F _n] ^T

$W=\left[\begin{array}{cccc}{w}_{\mathrm{1,1}}& w_{\mathrm{1,2}}& ...& w_{1,l}\\ w_{\mathrm{2,1}}& w_{\mathrm{2,2}}& ...& w_{2,l}\\ .& .& & .\\ .& .& ...& .\\ .& .& & .\\ w_{n,1}& w_{n,2}& ...& w_{n,l}\end{array}\right]=\left[\begin{array}{cccc}{w}_1& w_2& ...& w_1\end{array}\right]$

$\mathrm{\Θ}=\left[\begin{array}{c}{\mathrm{\θ}}_1\\ {\mathrm{\θ}}_2\\ .\\ .\\ .\\ {\mathrm{\θ}}_l\end{array}\right]$

F _n＝w _n,1θ ₁+w _n,2θ ₂+…+w _n,lθ _l

Wherein, matrix W is the weight vectors matrix that every evaluation index of each program obtains according to various tax power method, and Θ is the Linearly Representation coefficient vector of each tax power method, w _n,lfor composing according to l kind the weight that power method obtains the n-th evaluation index, weight is not less than the number that zero, n is evaluation index, and l is the number of tax power method, w _lobtain the weight vectors of the weight composition of each evaluation index for composing power method according to l kind, and in each weight vectors, weight sum equals 1, θ _lbe the Linearly Representation coefficient that l kind composes power method, θ _k>=0, k=1,2 ..., l, f _nit is the combining weights of the n-th evaluation index; Program evaluation model construction unit, utilizes decision matrix and right vector establishing target Optimized model, i.e. program evaluation model according at least two kinds of evaluation methods, and wherein, described objective optimization model is double-goal optimal model or Model for Multi-Objective Optimization; Optimum solution determining unit, adopts method of Lagrange multipliers to solve the optimum solution of above-mentioned program evaluation model; Obtain unit, the optimum linearity table obtaining each evaluation index of each program according to the audience information of program to be evaluated and described optimum solution goes out coefficient vector and optimum combination weight vector, thus obtains the comprehensive evaluation value of each program.

According to another aspect of the present invention, a kind of program evaluation method is provided, comprises: select tax power method, program to be evaluated and evaluation index, wherein, described tax power method comprises subjective weights method and Objective Weighting; Build program Criterion Attribute space, namely obtain the decision matrix of each evaluation index composition of each program according to the program to be evaluated of input block input and the computing formula of evaluation index, i.e. program Criterion Attribute space; Build and compose power index space, that is, according to described tax power method calculate each evaluation index of each program weight vectors and with the Linearly Representation coefficient of each tax power method, be multiplied and obtain right vector, namely compose power index space; Build program evaluation model, utilize decision matrix and right vector to build Bi-objective or Model for Multi-Objective Optimization, i.e. program evaluation model according at least two kinds of evaluation methods; Method of Lagrange multipliers is adopted to solve the optimum solution of above-mentioned program evaluation model; The optimum linearity table obtaining each evaluation index of each program according to the audience information of program to be evaluated and described optimum solution goes out coefficient vector and optimal weights vector, obtains the comprehensive evaluation value of each program.

Program evaluation system of the present invention and method achieve the combination weighting considering multiple subjective weights method and Objective Weighting, and utilize at least two kinds of evaluation methods to build Bi-objective or Model for Multi-Objective Optimization, subjective decision and objective making decision can be reflected simultaneously, the science realizing TV programme evaluation index composes power, improves accuracy and the reliability of program evaluation.

Accompanying drawing explanation

By reference to the content below in conjunction with the description of the drawings and claims, and understand more comprehensively along with to of the present invention, other object of the present invention and result will be understood and easy to understand more.In the accompanying drawings:

Fig. 1 is the formation block diagram of program evaluation system of the present invention;

Fig. 2 is the process flow diagram of program evaluation method of the present invention;

Fig. 3 is the process flow diagram of program evaluation model construction method of the present invention;

Fig. 4 is the process flow diagram of the construction method of the double-goal optimal model that the present invention is based on discrete maximization and unitization constraint condition;

Fig. 5 is the process flow diagram of the method for solving of program evaluation model optimum solution of the present invention.

Label identical in all of the figs indicates similar or corresponding feature or function.

Embodiment

In the following description, for purposes of illustration, in order to provide the complete understanding to one or more embodiment, many details have been set forth.But, clearly, also these embodiments can be realized when there is no these details.Below with reference to accompanying drawing, specific embodiments of the invention are described in detail.

Below with reference to accompanying drawing, specific embodiments of the invention are described in detail.

Fig. 1 is the formation block diagram of program evaluation system of the present invention, and as shown in Figure 1, program evaluation system of the present invention comprises:

Input block 110, for selecting tax power method, program to be evaluated and evaluation index, wherein, described tax power method comprises subjective weights method and Objective Weighting, such as, input block 110 can be touch-screen, computer, keyboard, mouse etc., it shows tax power method, program to be evaluated and evaluation index and selects for user, user can select m program to be evaluated, is designated as S={S ₁, S ₂..., S _m, n evaluation index, is designated as P={P ₁, P ₂..., P _n, compose power method and comprise the subjective weights method of such as analytical hierarchy process, expert survey, Fuzzy Analysis Method, binomial coefficient method and the Objective Weighting of such as Information Entropy, principal component analysis (PCA), average variance method, VC Method.

Program Criterion Attribute space construction unit 120, the program to be evaluated inputted according to input block 110 and the computing formula of evaluation index obtain the decision matrix of each evaluation index composition of each program, i.e. program Criterion Attribute space, preferably, described program Criterion Attribute space also comprises the specified decision matrix after by described decision matrix standardization processing, such as, i-th program S _ito a jth evaluation index P _jproperty value be designated as a _ij, A=(a _ij) _{m × n}be called decision matrix, wherein, i=1,2 ..., m, j=1,2 ..., n; Through the matrix B=(b of standardization processing _ij) _{m × n}be called normalized decision matrix, b _ijrepresent i-th program S _ito a jth evaluation index P _jstandardization property value, then the i-th row of matrix B represents i-th program S _ito the normal value of the property value of n evaluation index, and for example, evaluation index P _jfor audience ratings, then

Compose power index space construction unit 130, the tax power method inputted according to input block 110 calculate each evaluation index of each program weight and with the Linearly Representation coefficient of each tax power method, be multiplied and obtain right vector, namely power index space is composed, wherein, the weight of each evaluation index obtained according to Objective Weighting is relevant to described decision matrix, right vector F

F＝WΘ＝[F ₁F ₂… F _n] ^T

$W=\left[\begin{array}{cccc}{w}_{\mathrm{1,1}}& w_{\mathrm{1,2}}& ...& w_{1,l}\\ w_{\mathrm{2,1}}& w_{\mathrm{2,2}}& ...& w_{2,l}\\ .& .& & .\\ .& .& ...& .\\ .& .& & .\\ w_{n,1}& w_{n,2}& ...& w_{n,l}\end{array}\right]=\left[\begin{array}{cccc}{w}_1& w_2& ...& w_1\end{array}\right]$

$\mathrm{\Θ}=\left[\begin{array}{c}{\mathrm{\θ}}_1\\ {\mathrm{\θ}}_2\\ .\\ .\\ .\\ {\mathrm{\θ}}_l\end{array}\right]$

F _n＝w _n,1θ ₁+w _n,2θ ₂+…+w _n,lθ _l

Wherein, matrix W is the weight vectors matrix that every evaluation index of each program obtains according to various tax power method, and Θ is the Linearly Representation coefficient vector of each tax power method, w _n,lfor composing according to l kind the weight that power method obtains the n-th evaluation index, weight is not less than the number that zero, n is evaluation index, and l is the number of tax power method, w _lobtain the weight vectors of the weight composition of each evaluation index for composing power method according to l kind, and in each weight vectors, weight sum equals 1, θ _lbe the Linearly Representation coefficient that l kind composes power method, θ _k>=0, k=1,2 ..., l, f _nit is the combining weights of the n-th evaluation index.

Program evaluation model construction unit 140, decision matrix and right vector is utilized to build Bi-objective or Model for Multi-Objective Optimization according at least two kinds of evaluation methods, i.e. program evaluation model, such as, according to the double-goal optimal model that deviation maximization and unitization constraint condition two kinds of evaluation methods utilize decision matrix and right vector to build.

Optimum solution determining unit 150, adopts method of Lagrange multipliers to solve the optimum solution of above-mentioned program evaluation model.

Obtain unit 160, the optimum linearity table obtaining each evaluation index of each program according to the audience information of program to be evaluated and described optimum solution goes out coefficient vector and optimum optimum combination weight vector, thus obtains the comprehensive evaluation value of each program.

Preferably, program evaluation model construction unit 140 comprises: the first single object optimization model construction unit, utilizes decision matrix and right vector to build a kind of single object optimization model according to a kind of evaluation method; Second single object optimization model construction unit, utilizes decision matrix and right vector to build another kind of single object optimization model according to another kind of evaluation method; Double-goal optimal model construction unit, utilizes linear weight sum method that above-mentioned two kinds of single object optimization model linear combinations are built double-goal optimal model.

Foregoing description shows the program evaluation model construction unit 140 building double-goal optimal model, but the present invention is not limited to this, program evaluation model construction unit 140 can comprise multiple single object optimization model construction unit, utilizes linear weighting method to obtain Model for Multi-Objective Optimization.

The program evaluation model of above-mentioned program evaluation system is the comprehensive program evaluation model to program evaluation index subjective weights and Objective Weight, can comprehensively several evaluation methods evaluate above-mentioned model, both the micro-judgment of decision maker had been embodied, reflect again the dispersion degree of property value, reach as broadcast TV program index system carries out the object that science composes power.

Fig. 2 is the process flow diagram of program evaluation method of the present invention, and as shown in Figure 2, program evaluation method of the present invention comprises:

First, in step S210, select tax power method, program to be evaluated and evaluation index, wherein, described tax power method comprises subjective weights method and Objective Weighting;

In step S220, build program Criterion Attribute space, that is, namely obtain the decision matrix of each evaluation index composition of each program according to the program to be evaluated of input block input and the computing formula of evaluation index, i.e. program Criterion Attribute space;

In step S230, build and compose power index space, that is, according to described tax power method calculate each evaluation index of each program weight and with the Linearly Representation coefficient of each tax power method, be multiplied and obtain right vector, namely compose power index space;

In step S240, build program evaluation model, utilize decision matrix and right vector establishing target Optimized model, i.e. program evaluation model according to evaluation method, concrete building process will describe in figures 3 and 4;

In step s 250, adopt method of Lagrange multipliers to solve the optimum solution of above-mentioned program evaluation model, solution procedure will be described in Figure 5 particularly;

In step S260, the optimum linearity table obtaining each evaluation index of each program according to the audience information of program to be evaluated and described optimum solution goes out coefficient vector and optimum combination weight vector, thus obtains each program comprehensive evaluation value.

Preferably, in step S220, also comprise and standardization processing is carried out to described decision matrix.

Fig. 3 is the process flow diagram of program evaluation model construction method of the present invention, and as shown in Figure 3, described program evaluation model construction method comprises:

First, in step S310, utilize decision matrix and right vector to build a kind of single object optimization model according to a kind of evaluation method;

In step s 320, decision matrix and right vector is utilized to build another kind of single object optimization model according to another kind of evaluation method;

In step S330, utilize linear weight sum method that above-mentioned two kinds of single object optimization model linear combinations are built double-goal optimal model.

Above-mentioned process flow diagram only gives the process flow diagram of the building method of double-goal optimal model, but the present invention is not limited to this, when structure Model for Multi-Objective Optimization makes, only needs to build multiple single object optimization model according to the method described above again by its linear combination.

Fig. 4 is the process flow diagram of the construction method of the double-goal optimal model that the present invention is based on discrete maximization and unitization constraint condition, and as shown in Figure 4, the construction method based on the double-goal optimal model of discrete maximization and unitization constraint condition comprises:

First, in step S410, evaluation method based on deviation maximization utilizes decision matrix or specified decision matrix to build single object optimization model, particularly, utilizes each program in specified decision matrix to build mean dispersion error matrix J to the difference of the standardization property value of each index ₁, $J_1=\left[\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1,k\≠i}{\overset{m}{\mathrm{\Σ}}}\right|b_{i1}-b_{k1}|,\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1,k\≠i}{\overset{m}{\mathrm{\Σ}}}|b_{i2}-b_{k2}|,...,\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1,k\≠i}{\overset{m}{\mathrm{\Σ}}}|b_{\mathrm{in}}-b_{\mathrm{kn}}\left|\right];$ Then, in step s 320, mean dispersion error matrix and right vector is utilized to build single object optimization model based on deviation maximization, that is,

$M_1\left(F\right)=J_{1}F=J_1\mathrm{W\Θ},\begin{array}{c}\max M_1\left(\mathrm{\Θ}\right)=J_1\mathrm{W\Θ}\\ s.t.\left\{\begin{array}{c}{\mathrm{\Θ}}^T\mathrm{\Θ}=1\\ \mathrm{\Θ}\≥0\end{array}\right.\end{array};$

In the step s 420, evaluation method based on unitization constraint condition utilizes decision matrix or specified decision matrix to build single object optimization model, particularly, each program in specified decision matrix is utilized to build decision-making synthetical matrix J to the standardization property value sum of each index ₂, then, decision-making synthetical matrix and right vector is utilized to build single object optimization model based on unitization constraint condition, that is, $M_2\left(F\right)=J_{2}F=J_2\mathrm{W\Θ},\begin{array}{c}\max M_2\left(\mathrm{\Θ}\right)=\max \left(J_2\mathrm{W\Θ}\right)\\ s.t.\left\{\begin{array}{c}{\mathrm{\Θ}}^T\mathrm{\Θ}=1\\ \mathrm{\Θ}\≥0\end{array}\right.\end{array};$

In step S430, utilize weigthed sums approach that above-mentioned two kinds of single object optimization model linear combinations are built the double-goal optimal model based on deviation maximization and unitization constraint condition, described double-goal optimal model is $M_3\left(F\right)=J_{3}F=J_3\mathrm{W\Θ},\begin{array}{c}\max M_3\left(\mathrm{\Θ}\right)=J_3\mathrm{W\Θ}\\ s.t.\left\{\begin{array}{c}{\mathrm{\Θ}}^T\mathrm{\Θ}=1\\ \mathrm{\Θ}\≥0\end{array}\right.\end{array}$ Wherein, J ₃for biobjective scheduling matrix of coefficients, aJ ₁+ bJ ₂=J ₃, a and b is linear weighted function coefficient, a+b=1.Preferably, a=b=0.5.

Program evaluation method of the present invention builds the combination weights method based on deviation maximization and unitization constraint condition, the comprehensive evaluation value of each program is not only made to widen the difference between different grades as far as possible, even if the comprehensive evaluation value of each program disperses as far as possible, the comprehensive evaluation value of each program is also made to maximize as far as possible, the science that achieves composes power, ensure that the accuracy that program is evaluated and reliability.

Fig. 5 is the process flow diagram of the method for solving of program evaluation model optimum solution of the present invention, and as shown in Figure 5, the method for solving of program evaluation model optimum solution comprises:

First, in step S510, Lagrange conversion is carried out to program evaluation model, such as, construct the Lagrangian function of the double-goal optimal model based on deviation maximization and unitization constraint condition $L({\mathrm{\θ}}_1,{\mathrm{\θ}}_2,...,{\mathrm{\θ}}_l)=J_3\mathrm{W\Θ}+\mathrm{\λ}({\mathrm{\Θ}}^T\mathrm{\Θ}-1)=J_3\underset{k=1}{\overset{l}{\mathrm{\Σ}}}{W}_k{\mathrm{\θ}}_k+\mathrm{\λ}(\underset{k=1}{\overset{l}{\mathrm{\Σ}}}{\mathrm{\θ}}_k^2-1);$

In step S520, to through Lagrange conversion the differentiate of program evaluation model, find first order derivative be zero Linearly Representation coefficient, carry it into constraint condition and obtain optimum solution, wherein, constraint condition is Θ ^tΘ=1, and simultaneously meet Θ>=0, such as, solve the first order derivative of routine function, find first order derivative be zero Linearly Representation coefficient, that is, make try to achieve θ _k=-J ₃w _k/ 2 λ, k=1,2 ..., l, carries it into Θ ^tΘ=1, and meet Θ>=0 simultaneously, try to achieve so the optimum solution of this double-goal optimal model is ${\mathrm{\θ}}_k^{*}=J_3{W}_k/\sqrt{\underset{k=1}{\overset{l}{\mathrm{\Σ}}}{\left(J_3{W}_k\right)}^2},k=\mathrm{1,2},...,l.$

Preferably, in step S520, the optimum solution of program evaluation model is normalized, such as, right be normalized, that is, ${\mathrm{\θ}}_k^{**}={\mathrm{\θ}}_k^{*}/\underset{k=1}{\overset{l}{\mathrm{\Σ}}}{\mathrm{\θ}}_k^{*}=J_3{W}_k/\underset{k=1}{\overset{l}{\mathrm{\Σ}}}\left(J_3{W}_k\right),k=\mathrm{1,2},...,l,$ The weighing vector of program index is made to meet normalization constraint condition.

In addition, preferably, according in upper example, based on the optimum solution of the double-goal optimal model of deviation maximization and unitization constraint condition, the optimum combination weight vector of this double-goal optimal model is obtained the comprehensive evaluation value utilizing deviation maximization and unitization constraint condition two kinds of evaluation methods to obtain each program is

In sum, the program evaluation system and method that propose according to the present invention is described in an illustrative manner with reference to accompanying drawing.But, it will be appreciated by those skilled in the art that the system and method that the invention described above is proposed, various improvement can also be made on the basis not departing from content of the present invention.Therefore, protection scope of the present invention should be determined by the content of appending claims.

Claims (10)

1. a program evaluation system, comprising:

Input block, for inputting tax power method, program to be evaluated and evaluation index, wherein, described tax power method comprises subjective weights method and Objective Weighting;

Program Criterion Attribute space construction unit, obtains the decision matrix of each evaluation index composition of each program, i.e. program Criterion Attribute space according to the program to be evaluated of input block input and the computing formula of evaluation index;

Compose power index space construction unit, according to the tax power method of input block input calculate each evaluation index of each program weight and with the Linearly Representation coefficient of each tax power method, be multiplied and obtain right vector, namely power index space is composed, wherein, relevant to described decision matrix according to the weight vectors that Objective Weighting obtains, right vector F

F＝WΘ＝[F ₁F ₂… F _n] ^T

$W=\left[\begin{array}{cccc}{w}_{1,1}& w_{\mathrm{1,2}}& ...& w_{1,l}\\ w_{\mathrm{2,1}}& w_{\mathrm{2,2}}& ...& w_{2,l}\\ .& .& & .\\ .& .& ...& .\\ .& .& & .\\ w_{n,1}& w_{n,2}& ...& w_{n,l}\end{array}\right]=\left[\begin{array}{cccc}{w}_1& w_2& ...& w_l\end{array}\right]$

$\mathrm{\Θ}=\left[\begin{array}{c}{\mathrm{\θ}}_1\\ {\mathrm{\θ}}_2\\ .\\ .\\ .\\ {\mathrm{\θ}}_l\end{array}\right]$

F _n＝w _n,1θ ₁+w _n,2θ ₂+…+w _n,lθ _l

Wherein, matrix W is the weight vectors matrix that every evaluation index of each program obtains according to various tax power method, and Θ is the Linearly Representation coefficient vector of each tax power method, w _n,lfor composing according to l kind the weight that power method obtains the n-th evaluation index, weight is not less than the number that zero, n is evaluation index, and l is the number of tax power method, w _lobtain the weight vectors of the weight composition of each evaluation index for composing power method according to l kind, and in each weight vectors, weight sum equals 1, θ _lbe the Linearly Representation coefficient that l kind composes power method, f _nit is the combining weights of the n-th evaluation index;

Program evaluation model construction unit, utilizes decision matrix and right vector establishing target Optimized model, i.e. program evaluation model according at least two kinds of evaluation methods, and wherein, described objective optimization model is double-goal optimal model or Model for Multi-Objective Optimization;

Optimum solution determining unit, adopts method of Lagrange multipliers to solve the optimum solution of above-mentioned program evaluation model;

Obtain unit, the optimum linearity table obtaining each evaluation index of each program according to the audience information of program to be evaluated and described optimum solution goes out coefficient vector and optimum combination weight vector, thus obtains the comprehensive evaluation value of each program.

2. program evaluation system according to claim 1, wherein, described program Criterion Attribute space also comprises the specified decision matrix after by described decision matrix standardization processing.

3. program evaluation system according to claim 1, wherein, described program evaluation model construction unit comprises:

First single object optimization model construction unit, utilizes decision matrix and right vector to build a kind of single object optimization model according to a kind of evaluation method;

Second single object optimization model construction unit, utilizes decision matrix and right vector to build another kind of single object optimization model according to another kind of evaluation method;

Double-goal optimal model construction unit, utilizes linear weight sum method that above-mentioned two kinds of single object optimization model linear combinations are built double-goal optimal model.

4. program evaluation system according to claim 3, wherein, described program evaluation model is the double-goal optimal model utilizing decision matrix and right vector to build according to the evaluation method of deviation maximization and unitization constraint condition.

5. a program evaluation method, comprising:

Select tax power method, program to be evaluated and evaluation index, wherein, described tax power method comprises subjective weights method and Objective Weighting;

Build program Criterion Attribute space, namely obtain the decision matrix of each evaluation index composition of each program according to the program to be evaluated of input block input and the computing formula of evaluation index, i.e. program Criterion Attribute space;

Build and compose power index space, that is, according to described tax power method calculate each evaluation index of each program weight vectors and with the Linearly Representation coefficient of each tax power method, be multiplied and obtain right vector, namely compose power index space;

Build program evaluation model, utilize decision matrix and right vector to build Bi-objective or Model for Multi-Objective Optimization, i.e. program evaluation model according at least two kinds of evaluation methods;

Method of Lagrange multipliers is adopted to solve the optimum solution of above-mentioned program evaluation model;

The optimum linearity table obtaining each evaluation index of each program according to the audience information of program to be evaluated and described optimum solution goes out coefficient vector and optimal weights vector, obtains the comprehensive evaluation value of each program.

6. program evaluation method according to claim 5, wherein, described structure program Criterion Attribute space comprises carries out standardization processing to described decision matrix.

7. program evaluation method according to claim 5, wherein, described structure program evaluation model comprises:

Decision matrix and right vector is utilized to build a kind of single object optimization model according to a kind of evaluation method;

Decision matrix and right vector is utilized to build another kind of single object optimization model according to another kind of evaluation method;

Utilize weigthed sums approach that above-mentioned two kinds of single object optimization model linear combinations are built double-goal optimal model.

8. program evaluation method according to claim 7, wherein, described structure program evaluation model comprises:

Evaluation method based on deviation maximization utilizes decision matrix and right vector to build single object optimization model based on deviation maximization, and wherein, the single object optimization model based on deviation maximization is M ₁(F)=J ₁f=J ₁w Θ, $\begin{array}{c}\max M_1\left(\mathrm{\Θ}\right)=J_1\mathrm{W\Θ}\\ s.t.\left\{\begin{array}{c}{\mathrm{\Θ}}^T\mathrm{\Θ}=1\\ \mathrm{\Θ}\≥0\end{array}\right.\end{array}$ Wherein, J ₁for mean dispersion error matrix, $J_1=\left[\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1,k\≠i}{\overset{m}{\mathrm{\Σ}}}\right|b_{i1}-b_{k1}|,\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1,k\≠i}{\overset{m}{\mathrm{\Σ}}}|b_{i2}-b_{k2}|,...,\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1,k\≠i}{\overset{m}{\mathrm{\Σ}}}|b_{\mathrm{in}}-b_{\mathrm{kn}}\left|\right],$ Wherein, b _ijrepresent i-th program S _ito a jth evaluation index P _jstandardization property value, i=1,2 ..., m, m are the number of program to be evaluated, j=1,2 ..., n;

Evaluation method based on unitization constraint condition utilizes decision matrix and right vector to build the single object optimization model of unitization constraint condition, and the single object optimization model based on unitization constraint condition is M ₂(F)=J ₂f=J ₂w Θ, $\begin{array}{c}\max M_2\left(\mathrm{\Θ}\right)=J_2\mathrm{W\Θ}\\ s.t.\left\{\begin{array}{c}{\mathrm{\Θ}}^T\mathrm{\Θ}=1\\ \mathrm{\Θ}\≥0\end{array}\right.\end{array}$ Wherein, J ₂for decision-making synthetical matrix, $J_2=[\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{b}_{i1},\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{b}_{i2},...,\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{b}_{\mathrm{in}}];$

Utilize linear weight sum method that above-mentioned two single object optimization model linear combinations are built the double-goal optimal model based on deviation maximization and unitization constraint condition, described double-goal optimal model is M ₃(F)=J ₃f=J ₃w Θ, $\begin{array}{c}\max M_3\left(\mathrm{\Θ}\right)=J_3\mathrm{W\Θ}\\ s.t.\left\{\begin{array}{c}{\mathrm{\Θ}}^T\mathrm{\Θ}=1\\ \mathrm{\Θ}\≥0\end{array}\right.\end{array}$ Wherein, J ₃for biobjective scheduling matrix of coefficients, aJ ₁+ bJ ₂=J ₃, a and b is linear weighted function coefficient, a+b=1.

9. program evaluation method according to claim 8, wherein, the method for solving of the optimum solution of the described double-goal optimal model based on deviation maximization and unitization constraint condition comprises:

Biobjective scheduling program evaluation model based on deviation maximization and unitization constraint condition is carried out Lagrange conversion $L({\mathrm{\θ}}_1,{\mathrm{\θ}}_2,...,{\mathrm{\θ}}_l)=J_3\mathrm{W\Θ}+\mathrm{\λ}({\mathrm{\Θ}}^T\mathrm{\Θ}-1)=J_3\underset{k=1}{\overset{l}{\mathrm{\Σ}}}{W}_k{\mathrm{\θ}}_k+\mathrm{\λ}(\underset{k=1}{\overset{l}{\mathrm{\Σ}}}{\mathrm{\θ}}_k^2-1);$

To through Lagrange conversion program evaluation model carry out differentiate, find first order derivative be zero Linearly Representation coefficient, carry it into constraint condition and obtain optimum solution, wherein, constraint condition is Θ ^tΘ=1, and meet Θ>=0 simultaneously, the optimum solution of this double-goal optimal model is

To optimum solution be normalized, obtain normalization optimum solution, that is, ${\mathrm{\θ}}_k^{**}={\mathrm{\θ}}_k^{*}/\underset{k=1}{\overset{l}{\mathrm{\Σ}}}{\mathrm{\θ}}_k^{*}=J_3{W}_k/\underset{k=1}{\overset{l}{\mathrm{\Σ}}}\left(J_3{W}_k\right),k=\mathrm{1,2},...,l.$

10. program evaluation method according to claim 9, wherein, the optimum combination weight vector of the described double-goal optimal model based on deviation maximization and unitization constraint condition ${F_j}^{**}=w_{j,1}{{\mathrm{\θ}}_1}^{**}+w_{j,2}{{\mathrm{\θ}}_2}^{**}+\·\·\·+w_{j,l}{{\mathrm{\θ}}_l}^{**},$ The comprehensive evaluation value of each program is $Z_i=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{b}_{\mathrm{ij}}{F_j}^{**}.$