CN104812027A - Network selection method based on intuitionistic fuzzy set multi-attribute decision-making - Google Patents

Abstract

The invention discloses a network selection method based on intuitionistic fuzzy set multi-attribute decision-making. The method comprises the following steps: determining an intuitionistic fuzzy set decision-making matrix; converting each attribute value in the intuitionistic fuzzy set decision-making matrix into a trapezoidal fuzzy number to obtain a trapezoidal fuzzy number decision-making matrix, and simultaneously obtaining a membership function of each trapezoidal fuzzy number; by use of a distance formula of an intuitionistic fuzzy set, converting the distance relation of the intuitionistic fuzzy set into an area relation of the membership function of a corresponding trapezoidal fuzzy number set, and according to the trapezoidal fuzzy number decision-making matrix, obtaining a total deviation between each network to be selected and all other networks to be selected under each attribute; based on a deviation maximum idea, establishing a weight model of attributes, and according to a total deviation value, obtaining a weight value of each attribute; and by use of an intuitionistic fuzzy set weighed averaging operator (IFWA), calculating integrated attribute values of the networks to be selected, and selecting an optimum network from the networks to be selected. According to the invention, the problem of network selection in a heterogeneous network environment can be effectively solved.

Description

Based on the network selecting method of intuitionistic Fuzzy Sets multiple attribute decision making (MADM)

Technical field

The present invention relates to a kind of network selecting method based on intuitionistic Fuzzy Sets multiple attribute decision making (MADM), belong to technical field of the computer network.

Background technology

Along with the fast development of wireless communication technology, various wireless communication system is the network environment that user provides multiple isomery, comprising Wireless Personal Network (as Bluetooth), WLAN (wireless local area network) (as Wi-Fi), wireless MAN (as Wimax), public mobile network (as GSM, GPRS, UMTS) etc., making user be faced with multiple choices when selecting access network.Because above-mentioned various network also exists larger difference in coverage, message transmission rate, average packet time delay, service price, mobility support ability etc., therefore, when access or handover network, need first to solve network select permeability, make user reduce service cost while the service of acquisition high network quality, realize reasonable disposition and the utilization of Internet resources simultaneously.

Multiple attribute decision making (MADM) is the decision problem that social economy and field of engineering technology extensively exist, and due to the ambiguity of the complexity of objective things, uncertainty and human thinking, decision information is often expressed with fuzzy information.1986, Bulgaria scholar Atanassov proposes the concept of intuitionistic Fuzzy Sets (Intuitionistic Fuzzy Sets), this theory, to mode that the description of transaction attribute provides more choices, has stronger expressive ability when processing uncertain information.

Network selection procedures under heterogeneous network environment, can be regarded as intuitionistic Fuzzy Sets multiple attribute decision making (MADM) process, its target is compared by the synthesized attribute value of the network to accessible or switching to be selected, determines the quality of various network, therefrom select optimum network; But, due to knowledge or the limited experience of policymaker, or the complexity of policy setting, the weight information of each attribute of network is always unknown, therefore determines that the weight of attribute is very important.

Intuitionistic Fuzzy Sets multiple attributive decision making method mainly contains: first, subjective weighting method, mainly fuzzy synthetic evaluation model, is repeatedly with the method determination attribute weight such as method, paried comparison method by expert's point-score, root method, statistics of extremes, then solves synthesized attribute value with decision matrix computing; Second, based on the multiple attributive decision making method of intuitionistic Fuzzy Sets distance, as the sort method (TOPSIS), multiple criteria compromise solution ranking method (VIKOR), gray relative analysis method (GRA), multidimensional preference linear Programming Analysis method (LINMAP) etc. of similarity to ideal solution; 3rd, based on the multiple attributive decision making method of intuitionistic Fuzzy Sets entropy, according to Intuitionistic Fuzzy Entropy formula, based on the weight of entropy assessment determination attribute.

But, above-mentioned intuitionistic Fuzzy Sets multiple attributive decision making method network trade-off decision under not being suitable for heterogeneous network environment, this is because: for first method, expert's point-score is difficult to be applied in practice decision process due to the dependence to subjective priori, root method, paried comparison rule suppose that the pass between each factor is linear relationship, adopt simple weighted mean method to determine weight, but in the decision-making that network is selected, because the unit of each attribute of network is different, dimension is different, the order of magnitude is different, this hypothesis is not always set up, for second method, TOPSIS, VIKOR, GRA method is calculating the constant existing in weight process and need artificially to determine, and network environment is in change real-time dynamicly, policymaker is difficult to these constants of real-time determination, and the range formula that this method adopts just brings the parameter of intuitionistic Fuzzy Sets into calculating as single numerical value, and due to the complexity of network environment, network to be selected is for the degree of membership of relevant attribute, the fuzzy information that non-affiliated degree etc. can not lean on single numerical value just can be represented represents, therefore above-mentioned range formula well can not distinguish the uncertain information amount of different intuitionistic Fuzzy Sets in some cases, for the third method, entropy according to each attribute determines weight, do not consider the different otherness of network to be selected under same attribute, if the entropy of a certain network attribute is very large, but the difference of all-network under this attribute is very little, the difference of the synthesized attribute value of each network so finally obtained is also just little, is unfavorable for last network trade-off decision.

Therefore, existing intuitionistic Fuzzy Sets multiple attributive decision making method is owing to accurately can not weigh the distance of intuitionistic Fuzzy Sets, weight cannot be determined dynamically according to real-time property value, and the weighted value drawn sometimes and be unfavorable for sequence and the decision-making of network to be selected, so the problem that network selects effectively cannot be solved.

Summary of the invention

In view of the foregoing, the object of the present invention is to provide a kind of network selecting method based on intuitionistic Fuzzy Sets multiple attribute decision making (MADM), the method can be intuitionistic Fuzzy Sets at the property value of network to be selected and in the completely unknown heterogeneous network environment of weight, according to the weight of real-time property value determination attribute, and then the network selected best access or switch from network to be selected, effective solution network select permeability.

For achieving the above object, the present invention is by the following technical solutions:

Based on the network selecting method of intuitionistic Fuzzy Sets multiple attribute decision making (MADM), comprising:

S1: determine intuitionistic Fuzzy Sets decision matrix;

S2: each property value in this intuitionistic Fuzzy Sets decision matrix is converted to Trapezoid Fuzzy Number, obtains Trapezoid Fuzzy Number decision matrix, obtain the membership function of each Trapezoid Fuzzy Number simultaneously;

S3: the range formula utilizing intuitionistic Fuzzy Sets, the distance relation of intuitionistic Fuzzy Sets is converted to the area relationship of the membership function of corresponding Trapezoid Fuzzy Number collection, according to this Trapezoid Fuzzy Number decision matrix, obtain each network to be selected and the total deviation of other networks to be selected all under each attribute;

S4: based on deviation maximization thought, sets up the weight model of attribute, according to the total deviation value obtained in step S3, obtains the weighted value of each attribute;

S5: in conjunction with the weighted value of each attribute, utilizes intuitionistic Fuzzy Sets algebraic mean operator IFWA to calculate the synthesized attribute value of network to be selected, selects optimum network according to result of calculation from network to be selected.

Further,

Intuitionistic Fuzzy Sets decision matrix in described step S1 is:

R＝(r _ij) _m×n

Wherein: r _ij=(μ _ij, v _ij), i ∈ M, j ∈ N, M={1,2 ... m}, N={1,2 ... n},

μ _ijrepresent network A _ithere is attribute x _jdegree, v _ijrepresent network A _inot there is attribute x _jdegree, A={A ₁, A ₂..., A _mrepresent the set of m network to be selected, X={X ₁, X ₂..., X _nrepresent the set of n attribute of critic network performance.

In described step S2, intuitionistic Fuzzy Sets decision matrix is converted to Trapezoid Fuzzy Number decision matrix, by following Mapping implementation:

Wherein,

For arbitrary Intuitionistic Fuzzy Numbers α=(μ _α, v _α, π _α), corresponding Trapezoid Fuzzy Number is:

Then intuitionistic fuzzy counts to being mapped as of Trapezoid Fuzzy Number: have:

$\left\{\begin{array}{c}a={\mathrm{\μ}}_{\mathrm{\α}}\\ b={\mathrm{\μ}}_{\mathrm{\α}}+\frac{{\mathrm{\π}}_{\mathrm{\α}}}{4}\\ c={\mathrm{\μ}}_{\mathrm{\α}}+\frac{{3\mathrm{\π}}_{\mathrm{\α}}}{4}\\ d={\mathrm{\μ}}_{\mathrm{\α}}+{\mathrm{\π}}_{\mathrm{\α}}\end{array}\right.$

In described step S2, the membership function of described Trapezoid Fuzzy Number is:

In described step S3, the range formula of intuitionistic Fuzzy Sets is:

$d(A_i,A_k)=d({\tilde{A}}_i,{\tilde{A}}_k)=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ω}}_{j}d({\tilde{r}}_{\mathrm{ij}},{\tilde{r}}_{\mathrm{kj}})$

Wherein, ω _jrepresent a jth attribute x of network _jweight properties, weight sets ω={ ω ₁, ω ₂..., ω _n, $(\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ω}}_j^2=1,{\mathrm{\ω}}_j\∈[\mathrm{0,1}],j\∈N).$

In described step S3, the formula calculating each network to be selected and the total deviation of other networks to be selected all under each attribute is:

$D\left(\mathrm{\ω}\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{D}_{\mathrm{ij}}\left(\mathrm{\ω}\right)=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\ω}}_{j}d(\widetilde{r_{\mathrm{ij}}},\widetilde{r_{\mathrm{kj}}})$

In described step S4, the weight model of attribute is:

$\max D\left(\mathrm{\ω}\right)=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{D}_j\left(\mathrm{\ω}\right)=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\ω}}_{j}d({\tilde{r}}_{\mathrm{ij}},{\tilde{r}}_{\mathrm{kj}})$

$s.t.\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ω}}_j^2=1{,\mathrm{\ω}}_j\∈\left[\mathrm{0,1}\right],i=\mathrm{1,2},...,n$

In described step S4, the weighted value computing formula of attribute is:

${\mathrm{\ω}}_j=\frac{\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{m}{\mathrm{\Σ}}}d({\tilde{r}}_{\mathrm{ij}},{\tilde{r}}_{\mathrm{kj}})}{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{m}{\mathrm{\Σ}}}d({\tilde{r}}_{\mathrm{ij}},{\tilde{r}}_{\mathrm{kj}})}$

In described step S5, the synthesized attribute value formula calculating network to be selected is:

$r_i=\mathrm{IFWA}(r_{i1},r_{i2},...r_{\mathrm{in}})=(1-\underset{i=1}{\overset{n}{\mathrm{\Π}}}{(1-{\mathrm{\μ}}_i)}^{{\mathrm{\ω}}_i},\underset{i=1}{\overset{n}{\mathrm{\Π}}}{v}_i^{{\mathrm{\ω}}_i})$

The invention has the advantages that:

1, utilize the property value based on intuitionistic Fuzzy Sets to represent and network to be selected being subordinate to and non-affiliated information for each attribute effectively can show decision information;

2, intuitionistic Fuzzy Sets is used Trapezoid Fuzzy Number approximate representation, the distance of intuitionistic Fuzzy Sets is converted to the area relationship being included in the membership function that points countless in degree of membership interval forms by the distance of limited point, improves the computational accuracy of distance;

3, based on the thought determination weight model of deviation maximization, pole is beneficial to the otherness embodied between network, makes the sequence of network more accurate.

Accompanying drawing explanation

Fig. 1 is method flow schematic diagram of the present invention.

Fig. 2 to Fig. 7 is the possible position relationship schematic diagrames of six kinds of two Trapezoid Fuzzy Number membership functions respectively.

Embodiment

Below in conjunction with drawings and Examples, the present invention is described in further detail.

Fig. 1 is method flow schematic diagram of the present invention, and as shown in the figure, the network selecting method based on intuitionistic Fuzzy Sets multiple attribute decision making (MADM) disclosed by the invention, comprises step:

S1: determine intuitionistic Fuzzy Sets decision matrix;

To the multiple attribute of the network to be selected in heterogeneous network, with the property value of each attribute of intuitionistic fuzzy set representations, obtain decision matrix.

Intuitionistic Fuzzy Sets Multiple Attribute Decision Model is:

If the set of m to be selected network is A={A ₁, A ₂..., A _m, the set of n attribute of evaluating network performance is X={X ₁, X ₂..., X _n, separately remember M={1,2 ... m}, N={1,2 ... n};

Attribute weight ω={ ω ₁, ω ₂..., ω _n, $(\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ω}}_j^2=1,{\mathrm{\ω}}_j\∈[\mathrm{0,1}],j\∈N),$ ω _jrepresent a jth attribute x of network _jweight, then intuitionistic Fuzzy Sets decision matrix is:

R＝(r _ij) _m×n(1)

Wherein, r _ij=(μ _ij, v _ij), i ∈ M, j ∈ N, r _ijrepresent i-th network A _ifor a jth attribute X _jproperty value, μ _ijrepresent network A _ithere is attribute x _jdegree, v _ijrepresent network A _inot there is attribute x _jdegree.

S2: each property value in intuitionistic Fuzzy Sets decision matrix is converted to Trapezoid Fuzzy Number, obtains the membership function of each Trapezoid Fuzzy Number, determine Trapezoid Fuzzy Number decision matrix;

S21: Intuitionistic Fuzzy Numbers (each property value in intuitionistic Fuzzy Sets decision matrix) is converted to Trapezoid Fuzzy Number:

For arbitrary Intuitionistic Fuzzy Numbers α=(μ _α, v _α, π _α), corresponding Trapezoid Fuzzy Number is:

Then intuitionistic fuzzy counts to being mapped as of Trapezoid Fuzzy Number: wherein:

$\left\{\begin{array}{c}a={\mathrm{\μ}}_{\mathrm{\α}}\\ b={\mathrm{\μ}}_{\mathrm{\α}}+\frac{{\mathrm{\π}}_{\mathrm{\α}}}{4}\\ c={\mathrm{\μ}}_{\mathrm{\α}}+\frac{{3\mathrm{\π}}_{\mathrm{\α}}}{4}\\ d={\mathrm{\μ}}_{\mathrm{\α}}+{\mathrm{\π}}_{\mathrm{\α}}\end{array}\right.---\left(2\right)$

S22: the membership function determining Trapezoid Fuzzy Number;

Trapezoid Fuzzy Number membership function its definition is as follows:

S23: intuitionistic Fuzzy Sets decision matrix is converted to Trapezoid Fuzzy Number decision matrix;

Based on the mapping in step S21 by intuitionistic Fuzzy Sets decision matrix R=(r _ij) _{m × n}be converted to Trapezoid Fuzzy Number decision matrix $\tilde{R}={\left({\tilde{r}}_{\mathrm{ij}}\right)}_{m\×n}={(a_{\mathrm{ij}},b_{\mathrm{ij}},c_{\mathrm{ij}},d_{\mathrm{ij}})}_{m\×4n},$ That is:

Wherein, each property value r _ijcorresponding Trapezoid Fuzzy Number is that is,

Meanwhile, according to formula (3), the membership function that each Trapezoid Fuzzy Number is corresponding is obtained.

S3: the range formula utilizing intuitionistic Fuzzy Sets, the distance relation of intuitionistic Fuzzy Sets is converted to the area relationship of the membership function of corresponding Trapezoid Fuzzy Number collection, according to Trapezoid Fuzzy Number decision matrix, obtain each network to be selected and the total deviation of other networks to be selected all under each attribute;

S31: the range formula determining Intuitionistic Fuzzy Numbers;

The range formula of Intuitionistic Fuzzy Numbers is defined as follows:

If two arbitrary Intuitionistic Fuzzy Numbers α ₁=(μ ₁, v ₁, π ₁) and α ₂=(μ ₂, v ₂, π ₂), corresponding Trapezoid Fuzzy Number is respectively with have ${\widetilde{\mathrm{\α}}}_1=(a_1,b_1,c_1,d_1),{\widetilde{\mathrm{\α}}}_2=(a_2,b_2,c_2,d_2),$ The two corresponding Trapezoid Fuzzy Number of distance then between two Intuitionistic Fuzzy Numbers represents, that is:

$d({\mathrm{\α}}_1,{\mathrm{\α}}_2)=d({\widetilde{\mathrm{\α}}}_1,{\widetilde{\mathrm{\α}}}_2)=S_C-S_I---\left(6\right)$

Wherein, S _crepresent by Trapezoid Fuzzy Number with the outmost boundary line of membership function and horizontal cross shaft and degree of membership be the area of the rectilinear(-al) of 1, S _irepresent Trapezoid Fuzzy Number with the common factor of area of membership function;

Then,

For two intuitionistic Fuzzy Sets, as network A to be selected _iand A _k, the property set of this two intuitionistic Fuzzy Sets is designated as: A _i={ r _i1, r _i2... r _in, A _k={ r _k1, r _k2... r _kn,

Each property value being converted to the set that corresponding Trapezoid Fuzzy Number obtains Trapezoid Fuzzy Number is ${\tilde{A}}_i=\{{\tilde{r}}_{i1},{\tilde{r}}_{i2},...{\tilde{r}}_{\mathrm{in}}\},{\tilde{A}}_k=\{{\tilde{r}}_{k1},{\tilde{r}}_{k2},...{\tilde{r}}_{\mathrm{kn}}\},$ So, network A _iand A _kdifference be expressed as by the distance of the intuitionistic Fuzzy Sets of the two:

$d(A_i,A_k)=d({\tilde{A}}_i,{\tilde{A}}_k)=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ω}}_{j}d({\tilde{r}}_{\mathrm{ij}},{\tilde{r}}_{\mathrm{kj}})---\left(7\right)$

S32: based on the range formula of intuitionistic Fuzzy Sets, solves each network to be selected and the total deviation of other networks to be selected all under each attribute;

Based on formula (7), and formula (4),

For attribute X _jif, network A to be selected _iwith other network A to be selected all _kdeviation be expressed as:

$D_{\mathrm{ij}}\left(\mathrm{\ω}\right)=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}d({\tilde{r}}_{\mathrm{ik}},{\tilde{r}}_{\mathrm{jk}})\·{\mathrm{\ω}}_j---\left(8\right)$

Wherein, i ∈ M, j ∈ N,

For attribute X _j, all networks to be selected and other deviations of network to be selected all are expressed as:

$D_j\left(\mathrm{\ω}\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{D}_{\mathrm{ij}}\left(\mathrm{\ω}\right)---\left(9\right)$

Wherein, j ∈ N,

For all properties, all networks to be selected and other the total deviation of network to be selected under whole attribute all are:

$D\left(\mathrm{\ω}\right)=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{D}_j\left(\mathrm{\ω}\right)=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\ω}}_{j}d({\tilde{r}}_{\mathrm{ij}},{\tilde{r}}_{\mathrm{kj}})---\left(10\right)$

S4: based on deviation maximization thought, sets up the weight model of attribute, according to the total deviation value obtained in step S3, obtains the weighted value of each attribute;

S41: based on the range formula of deviation maximization thought and intuitionistic Fuzzy Sets, the Weight of attribute is converted into nonlinear programming problem, and the weight model obtaining attribute is as follows:

$\begin{array}{c}\max D\left(\mathrm{\ω}\right)=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{D}_j\left(\mathrm{\ω}\right)=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\ω}}_{j}d\left(\widetilde{r_{\mathrm{ij}},}\widetilde{r_{\mathrm{kj}}}\right)\\ s.t.\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ω}}_j^2=1,{\mathrm{\ω}}_j\∈\left[\mathrm{0,1}\right],j\∈N\end{array}---\left(11\right)$

S42: structure Lagrange function solves formula (11):

$L(\mathrm{\ω},\mathrm{\λ})=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{\ω}}_{j}d({\tilde{r}}_{\mathrm{ij}},{\tilde{r}}_{\mathrm{kj}})+\frac{\mathrm{\λ}}{2}(\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ω}}_j^2-1)---\left(12\right)$

Respectively partial derivative is asked for ω and λ in formula (12):

$\frac{\∂L}{{\∂\mathrm{\ω}}_j}=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{m}{\mathrm{\Σ}}}d({\tilde{r}}_{\mathrm{ij}},{\tilde{r}}_{\mathrm{kj}})+\mathrm{\λ}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ω}}_j=0---\left(13\right)$

$\frac{\∂L}{\∂\mathrm{\λ}}=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ω}}_j^2-1=0---\left(14\right)$

Obtain unitization attribute weight computing formula:

${\mathrm{\ω}}_j^{*}=\frac{\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{m}{\mathrm{\Σ}}}d({\tilde{r}}_{\mathrm{ij}},{\tilde{r}}_{\mathrm{kj}})}{\sqrt{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{\left[\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{m}{\mathrm{\Σ}}}d({\tilde{r}}_{\mathrm{ij}},{\tilde{r}}_{\mathrm{kj}})\right]}^2}},j\∈N---\left(15\right)$

Unitization weight shown in formula (15) is normalized and obtains normalized weight computing formula:

${\mathrm{\ω}}_j=\frac{\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{m}{\mathrm{\Σ}}}d({\tilde{r}}_{\mathrm{ij}},{\tilde{r}}_{\mathrm{kj}})}{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{m}{\mathrm{\Σ}}}d({\tilde{r}}_{\mathrm{ij}},{\tilde{r}}_{\mathrm{kj}})},j\∈N---\left(16\right)$

S5: in conjunction with the weighted value of attribute, utilizes intuitionistic Fuzzy Sets algebraic mean operator IFWA, calculates the synthesized attribute value of network to be selected, select optimum access network according to result of calculation from network to be selected.

Network A to be selected _icommunity set be A _i={ r _i1, r _i2... r _in, utilize intuitionistic Fuzzy Sets algebraic mean operator IFWA to calculate the synthesized attribute value of network to be selected:

$r_i=\mathrm{IFWA}(r_{i1},r_{i2},...r_{\mathrm{in}})=(1-\underset{i=1}{\overset{n}{\mathrm{\Π}}}{(1-{\mathrm{\μ}}_i)}^{{\mathrm{\ω}}_i},\underset{i=1}{\overset{n}{\mathrm{\Π}}}{v}_i^{{\mathrm{\ω}}_i})---\left(17\right)$

According to the synthesized attribute value r that formula (17) calculates _i, treat network selection network rank, and therefrom select optimum network.

Fig. 2 to Fig. 7 is the possible position relationship schematic diagrames of six kinds of two Trapezoid Fuzzy Number membership functions respectively.As shown in the figure, in above-mentioned steps S31, the position relationship of the membership function of two concrete Trapezoid Fuzzy Numbers has six kinds, is respectively:

1), in two Trapezoid Fuzzy Number membership functions shown in Fig. 2, condition is: a ₁≤ d ₁≤ a ₂,

Two fuzzy number distances: $d({\widetilde{\mathrm{\α}}}_1,{\widetilde{\mathrm{\α}}}_2)=S_C-S_I=S_{\mathrm{ABCD}}=\frac{c_2-b_1+d_2-a_1}{2}---\left(61\right)$

2), in two Trapezoid Fuzzy Number membership functions shown in Fig. 3, condition is: a ₁≤ a ₂≤ d ₁≤ d ₂and c ₁≤ b ₂,

Two fuzzy number distances: $d({\widetilde{\mathrm{\α}}}_1,{\widetilde{\mathrm{\α}}}_2)=S_C-S_I=\frac{c_2-b_1+d_2-a_1}{2}-\underset{a_2}{\overset{d_1}{\∫}}\min ({\mathrm{\μ}}_{{\widetilde{\mathrm{\α}}}_1},{\mathrm{\μ}}_{{\widetilde{\mathrm{\α}}}_2})\mathrm{dt}\left(62\right)$

3), in two Trapezoid Fuzzy Number membership functions shown in Fig. 4, condition is: a ₁≤ a ₂≤ d ₁≤ d ₂and c ₁>=b ₂,

Two fuzzy number distances: $d({\widetilde{\mathrm{\α}}}_1,{\widetilde{\mathrm{\α}}}_2)=S_C-S_I=\frac{c_2-b_1+d_2-a_1}{2}-\frac{d_1-a_2+c_1-b_2}{2}---\left(63\right)$

4), in two Trapezoid Fuzzy Number membership functions shown in Fig. 5, condition is: a ₁≤ a ₂≤ d ₂≤ d ₁, b ₁≤ b ₂, c ₁>=c ₂,

Two fuzzy number distances: $d({\widetilde{\mathrm{\α}}}_1,{\widetilde{\mathrm{\α}}}_2)=S_C-S_I=\frac{d_1-a_1+c_1-b_1}{2}-\frac{d_2-a_2+c_2-b_2}{2}---\left(64\right)$

5), in two Trapezoid Fuzzy Number membership functions shown in Fig. 6, condition is: a ₁≤ a ₂≤ d ₂≤ d ₁, b ₁≤ b ₂, c ₁≤ c ₂,

Two fuzzy number distances: $d({\widetilde{\mathrm{\α}}}_1,{\widetilde{\mathrm{\α}}}_2)=S_C-S_I=S_{\mathrm{ABEK}}+S_{\mathrm{CFP}}+S_{\mathrm{PQD}}---\left(65\right)$

6), in two Trapezoid Fuzzy Number membership functions shown in Fig. 7, condition is: a ₁≤ a ₂≤ d ₂≤ d ₁, b ₁>=b ₂, c ₁>=c ₂,

Two fuzzy number distances: $d({\widetilde{\mathrm{\α}}}_1,{\widetilde{\mathrm{\α}}}_2)=S_C-S_I=S_{\mathrm{GCDH}}+S_{\mathrm{PFB}}+S_{\mathrm{PAE}}---\left(66\right)$

In formula (7) can according to 1)-6) in condition choice for use formula (61)-(66) in one calculate.

The present invention is the network selecting method based on intuitionistic Fuzzy Sets multiple attribute decision making (MADM), to the property value of the multiple attribute of the network to be selected in heterogeneous network, represent with intuitionistic Fuzzy Sets decision matrix, utilize mapping relations that intuitionistic Fuzzy Sets decision matrix is converted to Trapezoid Fuzzy Number decision matrix, obtain the membership function of each Trapezoid Fuzzy Number simultaneously; Then, set up the range formula of intuitionistic Fuzzy Sets, the distance relation of intuitionistic Fuzzy Sets is converted to the area relationship of the membership function of corresponding Trapezoid Fuzzy Number collection, based on intuitionistic Fuzzy Sets range formula, obtains each network to be selected and the total deviation of other networks to be selected all under each attribute; Then, based on deviation maximization thought, set up the weight model of attribute, obtain the weighted value of each attribute; Finally, in conjunction with the Attribute Weight weight values obtained, utilize intuitionistic Fuzzy Sets algebraic mean operator IFWA to try to achieve the synthesized attribute value of network to be selected, each network to be selected is sorted, therefrom select optimum access or handover network.The present invention utilizes the property value based on intuitionistic Fuzzy Sets to represent being subordinate to and non-affiliated information of each attribute of network to be selected, effectively can show decision information, intuitionistic Fuzzy Sets is used Trapezoid Fuzzy Number approximate representation, improve the precision calculating intuitionistic Fuzzy Sets distance, achieve Attribute Weight weight values and carry out self-adaptative adjustment with the dynamic change of property value, can effectively solve network select permeability.

The above know-why being preferred embodiment of the present invention and using; for a person skilled in the art; when not deviating from the spirit and scope of the present invention; any based on apparent changes such as the equivalent transformation on technical solution of the present invention basis, simple replacements, all belong within scope.

Claims (9)

1. based on the network selecting method of intuitionistic Fuzzy Sets multiple attribute decision making (MADM), it is characterized in that, comprising:

S1: determine intuitionistic Fuzzy Sets decision matrix;

S2: each property value in this intuitionistic Fuzzy Sets decision matrix is converted to Trapezoid Fuzzy Number, obtains Trapezoid Fuzzy Number decision matrix, obtain the membership function of each Trapezoid Fuzzy Number simultaneously;

S3: the range formula utilizing intuitionistic Fuzzy Sets, the distance relation of intuitionistic Fuzzy Sets is converted to the area relationship of the membership function of corresponding Trapezoid Fuzzy Number collection, according to this Trapezoid Fuzzy Number decision matrix, obtain each network to be selected and the total deviation of other networks to be selected all under each attribute;

S4: based on deviation maximization thought, sets up the weight model of attribute, according to the total deviation value obtained in step S3, obtains the weighted value of each attribute;

S5: in conjunction with the weighted value of each attribute, utilizes intuitionistic Fuzzy Sets algebraic mean operator IFWA to calculate the synthesized attribute value of network to be selected, selects optimum network according to result of calculation from network to be selected.

2., as claimed in claim 1 based on the network selecting method of intuitionistic Fuzzy Sets multiple attribute decision making (MADM), it is characterized in that, the intuitionistic Fuzzy Sets decision matrix in described step S1 is:

R＝(r _ij) _m×n

Wherein: r _ij=(μ _ij, v _ij), i ∈ M, j ∈ N, M={1,2 ... m}, N={1,2 ... n},

μ _ijrepresent network A _ithere is attribute x _jdegree, v _ijrepresent network A _inot there is attribute x _jdegree, A={A ₁, A ₂..., A _mrepresent the set of m network to be selected, X={X ₁, X ₂..., X _nrepresent the set of n attribute of critic network performance.

3., as claimed in claim 2 based on the network selecting method of intuitionistic Fuzzy Sets multiple attribute decision making (MADM), it is characterized in that, in described step S2, intuitionistic Fuzzy Sets decision matrix is converted to Trapezoid Fuzzy Number decision matrix, by following Mapping implementation:

Wherein,

For arbitrary Intuitionistic Fuzzy Numbers α=(μ _α, v _α, π _α), corresponding Trapezoid Fuzzy Number is:

Then intuitionistic fuzzy counts to being mapped as of Trapezoid Fuzzy Number: have:

。

4., as claimed in claim 3 based on the network selecting method of intuitionistic Fuzzy Sets multiple attribute decision making (MADM), it is characterized in that, in described step S2, the membership function of described Trapezoid Fuzzy Number is:

。

5., as claimed in claim 4 based on the network selecting method of intuitionistic Fuzzy Sets multiple attribute decision making (MADM), it is characterized in that, in described step S3, the range formula of intuitionistic Fuzzy Sets is:

Wherein, ω _jrepresent a jth attribute x of network _jweight properties, weight sets ω={ ω ₁, ω ₂..., ω _n,

6., as claimed in claim 5 based on the network selecting method of intuitionistic Fuzzy Sets multiple attribute decision making (MADM), it is characterized in that, in described step S3, the formula calculating each network to be selected and the total deviation of other networks to be selected all under each attribute is:

。

7., as claimed in claim 6 based on the network selecting method of intuitionistic Fuzzy Sets multiple attribute decision making (MADM), it is characterized in that, in described step S4, the weight model of attribute is:

。

8., as claimed in claim 7 based on the network selecting method of intuitionistic Fuzzy Sets multiple attribute decision making (MADM), it is characterized in that, in described step S4, the weighted value computing formula of attribute is:

。

9., as claimed in claim 8 based on the network selecting method of intuitionistic Fuzzy Sets multiple attribute decision making (MADM), it is characterized in that, in described step S5, the synthesized attribute value formula calculating network to be selected is:

。