CN105722149A - Topology construction excitation method based on reputation value - Google Patents

Abstract

The invention discloses a topology construction excitation method based on a reputation value. The method comprises steps of setting a wireless ubiquitous network communication scene; and assuming that there is a business demand node, a surround idle network region set K={k1, k2, ..., kn} and a relay node set gama={R1, R2, ..., Rn} corresponding to all idle network regions in a jam network region A in the scene. The method is applied to a wireless ubiquitous environment; by use of network performance indexes, an optimal idle network region is selected; based on history interaction information, by constructing an efficient mathematic model, a direct reputation value and an indirect reputation value of node evaluation are calculated, thereby obtaining a comprehensive reputation value of the node; and then via the periodic topology construction excitation algorithm, a topology structure connected to the node is adjusted, so cooperation nodes are allowed to be connected with each other while selfish nodes are repelled to network edges, thereby greatly accelerating enthusiasm of cooperation and forwarding of nodes with low reputation values.

Description

Topological structure motivational techniques based on credit value

Technical field

The present invention relates to a kind of wireless general in the environment based on the topological structure motivational techniques of credit value, belong to multimedia communication technology field.

Background technology

Development along with wireless communication technology, different requirements for coverage, number of users, type of service and service quality (QoS), research and develop various cellular communications network, WLAN (WLAN), Wireless Personal Network (WPAN) and satellite communication network at interior various wireless communication network at present, and more new wireless network is also continuing to bring out.These networks adopt different network technologies to supplement each other, jointly exist, and constitute the wireless ubiquitous environment of ubiquitous Multi net voting multi-service fusion.But, cluster characteristic wireless general user at ambient determines user traffic and concentrates on a certain network or specific cell (such as Cellular Networks community), causes that the play of this network or Zone flow is risen and the relative free situation of adjacent isomorphism or heterogeneous network resource, business.Connect by constituting the peer-to-peer network of Ad-hoc formula between terminal node and cooperate forwarding data, thus realizing being transferred to adjacent dormant network by the service traffics of business demand node from congested network, it is achieved wireless general each network traffics at ambient are balanced with load.But, Ad-hoc network has the features such as dynamic, opening and autonomy, the behavior making node faces serious selfishness and credit problems, is embodied in the uncontrolled use etc. of the uncooperative behavior of selfish node, " hitchhiking " and cooperative nodes resource.Based on the research direction that the trust management of prestige cooperates as excitation node, its Major Difficulties is in that how to utilize rationally effective mathematical model that the prestige of node is evaluated all-sidedly and accurately on the reputation information basis collected and utilize the purpose of certain mechanism realization excitation and punishment.Therefore, also increasingly receive publicity for the Study on Incentive Mechanism based on credit value.

Now, wireless general machine-processed for node activations and algorithm research at ambient is broadly divided into three directions: based on prestige, based on ideal money (including auction mechanism), based on game theoretic motivational techniques.But there is a lot of weak point in above-mentioned incentive mechanism, including: 1, only consider single-point-excitation, it does not have divide cooperative nodes and selfish node from the whole network angular area, cause that the search location to cooperative nodes and data forwarding performance decline.2, the mechanism having is more difficult acceptance in offered load, as increased extra hardware device or " bank " to Manage Currency based on ideal money mechanism.3, it is better than in cooperation policy and fairness based on credit mechanism based on game theory mechanism, but because the unstability of wireless network, the distributivity of wireless Ubiquitous Network interior joint and mobility, game theory result sometimes can not directly be applied.It is further noted that the performance of whole network is not optimum state under equilibrium state in betting model.And the present invention can solve the problems referred to above well.

Summary of the invention

Present invention aim at for above-mentioned the deficiencies in the prior art, inspiration problem especially for the load balance in the wireless general region of Multi net voting at ambient and multiple terminals node, a kind of topological structure motivational techniques based on credit value are proposed, the method be applied to wireless general in the environment, network performance index is utilized to select optimum dormant network region, based on history mutual information, the direct credit value to Node evaluation and indirect credit value is calculated by building effective mathematical model, thus obtaining the comprehensive credit value of egress, then pass through periodic topological structure excitation algorithm and adjust the topological structure that node connects, the network edge that cooperative nodes is connected with each other and selfish node is ostracised, and then it is greatly promoted the enthusiasm that the cooperation of low credit value node forwards.

This invention address that its technical problem is adopted the technical scheme that: a kind of topological structure motivational techniques based on credit value, the method setting wireless Ubiquitous Network communication scenes, it is assumed that there is business demand node s, ambient idle network area collection Κ={ k in this scene in the A of congested network region₁,k₂,…k_nAnd Κ in via node collection Γ={ R corresponding to each dormant network region₁,R₂,…R_n}。

Method flow:

Step 1: business demand node s selects performance-optimal network region k according to network capacity c and offered load l two indices from set Κ^*And correspondence via node collection R^*。

Step 2: business demand node s with broadcast mode to R^*In via node send service request and consider remaining power energy e, node load l' according to principle of optimality from R^*The optimum via node r of middle selection^*As service providing node.Business demand node s records via node interactive information, and 1 represents acceptance request and elected as service providing node by business demand node s, and 0 represents reject the service request.

Step 3: when the topological structure cycle arrives, according to history mutual information between node, the direct credit value of computing node.First, its recorded node history mutual information is divided in different time sections by business demand node s τ at timed intervals, then calculates the credit value in each time period, and computing formula is as follows:

${\mathrm{DT}}_{\mathrm{sj}}^t\left(n\right)=\left\{\begin{array}{ccc}\min (1,{\mathrm{DT}}_{\mathrm{sj}}^t(n-1)+a*k_g);& \mathrm{accept}& \mathrm{request}\\ \max (0,{\mathrm{DT}}_{\mathrm{sj}}^t(n-1)-m^2*k_m);& \mathrm{refuse}& \mathrm{request}\end{array}\right.$

WhereinRepresenting the t time period interior nodes s credit rating value to node j, n represents the interactive information article number in the t time period.A is that node j provides the number of times forwarding service, κ continuously for s_gIt is excitation factor, a* κ_gOn the basis of upper once credit rating, cooperative nodes is encouraged in a linear fashion.M is that node j refuses to provide the number of times forwarding service, κ for s continuously_mIt is penalty factor, m²*κ_mBy square in the way of on the basis of upper once credit rating, selfish node is punished.The seriality of a was refused to provide for s to forward service to calculate for initiateing with j the last time, and m is in like manner.The feature decayed to characterize the importance of node historical transactional information in time remote, utilizes the characteristic of quadratic function to describe the weight w (t) of directly credit value in different time sections, is shown below, it may be assumed that

$w\left(t\right)={(\frac{t}{STL}-1)}^2,(t=0,1,2,...,STL)$

Wherein w (t) is with the t function successively decreased, and STL (storetimelimit) is the reputation information storage time limit of default, and the reputation information record more than STL is no longer participate in calculating.It is weighted on average obtaining the node s direct credit value to j to credit value in different time sections with w (t):

${\mathrm{DT}}_{sj}=\frac{1}{STL}\underset{t=1}{\overset{STL}{\Σ}}({\mathrm{DT}}_{sj}^t*w(t\left)\right).$

Step 4: on direct credit value basis, uses each recommended node recommendation reputation value of behavior similarity Weight and utilizes the indirect credit value of normalization factor computing node.In single recommended node situation, the direct credit value of recommended node i is μ by node s, it is recommended that the direct credit value of objective appraisal node j is β by node i, then the indirect credit value of objective appraisal node j is μ β by node s；In many recommended nodes situation, in order to ensure the verity of recommendation behavior, by behavior similarity Weight recommendation reputation value.(i, s) is shown below the credit rating behavior similarity Sim of s and i, namely

$Sim\left(i,s\right)=\frac{{\mathrm{\Σ}}_{v\∈CZ\left(s,i\right)}{\mathrm{\μ}}_{sv}{\mathrm{\μ}}_{iv}}{\sqrt{{\mathrm{\Σ}}_{v\∈CZ\left(s,i\right)}{\mathrm{\μ}}_{sv}^2}\sqrt{{\mathrm{\Σ}}_{v\∈CZ\left(s,i\right)}{\mathrm{\μ}}_{sv}^2}}$

Wherein (s i) represents there was mutual node set with s and i to C Ζ.As weight, behavior similarity is described the node s indirect credit value to objective appraisal node j is:

${\mathrm{IDT}}_{sj}=\frac{1}{{\mathrm{\μ}}_{s1}+{\mathrm{\μ}}_{s2}+...+{\mathrm{\μ}}_{sn}}\underset{m=1}{\overset{n}{\Σ}}{\mathrm{\μ}}_{sm}^2*Sim(m,s)*{\mathrm{\β}}_{mj}$

Wherein μ_smRepresent node s to recommended node m ∈ 1,2 ..., the direct credit value of n}, β_mjRepresent the recommended node m direct credit value to objective appraisal node j.Calculate the direct credit value of gained and the indirect credit value of this step gained according to step 3, calculate comprehensive credit value T_sjFor:

T_sj=α * DT_sj+(1-α)*IDT_sj

Wherein α ∈ [0,1].

Step 5: after the comprehensive credit value of node updates, business demand node s calls topological structure excitation algorithm and completes self topology adjustment.Topological structure excitation algorithm is divided into node link control algolithm (i.e. LCA, Linkcontrolalgorithm) and node connection request verification algorithm (i.e. CVA, Connectionrequestverificationalgorithm)；LCA is responsible for deleting and the linking and to high credit value node initiation connection request, and set node self maintained three set of low credit value node, including:

(1) it is familiar with node set FS_s.Represent there was mutual node set with node s.When the topological structure cycle each time arrives, node s calculates the direct credit value of all nodes and indirect credit value in this set, then obtains comprehensive credit value and updates FS_s。FS_s=i | whohascooperatedwiths}.

(2) high credit value node set HTS_s.Represent credit value T_siMore than minimum prestige threshold value T_minNode set.HTS_s=i | T_si> T_min, HTS_s∈FS_s。

(3) neighbor node set NS_s.Expression and node s have the existing node set of annexation.About NS_sThree parameter respectively τ_min: NS_sIn minimum nodes.τ_max: the NS set according to node self computing capability_sIn maximum nodes, τ_max> τ_min。τ_base: the credit value that topological structure process interior joint makes great efforts to safeguard is higher than T_minNodes, τ_min< τ_base< τ_max.Node s is from HTS_sIn choose the node updates NS of high credit value_s, reach at NS_sInterior joint number is not less than τ_baseBasis on optimization NS_sThe purpose of interior joint credit value.NS_s=i | whoconnectswiths}, NS_s∈FS_s。

CVA is responsible for the credit value of the node of assessment initiation connection request and chooses whether to be attached with it according to the interstitial content in self neighbor node set.

Beneficial effect:

1, the general Multi net voting multiple terminals scene at ambient of the present invention first setting wireless, interactive history information based on business demand node, the direct credit value of utilization time attenuation function weighted calculation interaction node, and Behavior-based control similarity weight calculates the indirect credit value of interaction node with normalization factor, thus obtaining the comprehensive credit value of interaction node.

2, the present invention periodically adjusts network topology structure by adjusting three set of node maintenance, thus efficiently differentiating cooperative nodes and selfish node, and encourages selfish node to pass through cooperation forwarding self credit value of raising.

3, the present invention is simple, it is easy to accomplish, there is good application prospect.

Accompanying drawing explanation

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

Detailed description of the invention

Below in conjunction with Figure of description, the invention is described in further detail.

The present invention is at the wireless general heterogeneous network that there is multiple access in the environment, as: such as the network such as cellular basestation, WLAN, WIMAX, set business demand node s in this traffic model to be within the scope of cellular basestation A and connect base station network, business demand node s rises due to its place Zone flow play, need to carry out service traffics transfer, and near business demand node s, there is n isomery of relative free, i.e. (or isomorphism) network area Κ={ k₁,k₂,…k_nAnd Κ in via node collection Γ={ R corresponding to each dormant network region₁,R₂,…R_n}.Then node s selects the network area k of performance the best according to network capacity c and offered load l two indices from set Κ^*As business migration object, corresponding optimum via node integrates as R^*.For network area k_i, its network capacity c is more big, the more low then network performance of offered load lMore high.Definition is as shown in Equation 1, it may be assumed that

${\mathrm{\&eta;}}_{k_i}=l_{k_i}^{-1}+c_{k_i},i\&Element;\{1,2,...,n\}$ Formula 1

To the all-network region in Κ, shown in formula 2, optimizing model processes availability energy optimum network region k^*, optimizing model is:

$k^{*}=\arg \max {\mathrm{\&eta;}}_{k_i}$

$\begin{array}{cc}s.t& {\mathrm{\&eta;}}_{k_i}=l_{k_i}^{-1}+c_{k_i}\end{array}$ Formula 2

k_i∈ Κ={ k₁,k₂,…,k_n}

Owing to there is following two hindering factor: 1) community A and network area k^*Isomery, therefore both communication standards are different, for instance under 3G environment, and China Mobile adopts TD-SCDMA, and CHINAUNICOM adopts WCDMA；2) community A and network area k^*Communication coverage area is not overlapping, and node s must complete the migration of grouped data by the via node closed on.Node s with broadcast mode to R^*In via node send service request and record via node interactive information, 0 represent reject the service request, 1 represent accept request and elected as service providing node, service providing node r by business demand node s^*Need to meet:

$r^{*}=\mathrm{argmax}\frac{e_{k_i}}{l_{k_i}^{\&prime;}}$ Formula 3

Wherein k_i∈ Κ={ k₁,k₂,…,k_n,Expression remaining power energy,Represent node load.

When the topological structure cycle arrives, the first direct credit value of computing node.Direct credit value be requesting node without the recommendation of intermediary node direct credit rating value to service node, between node, history mutual information is to calculate the basic document of direct credit value, we first at timed intervals τ its recorded node history mutual information is divided in different time sections, then calculating the credit value in each time period, computing formula is as follows:

${\mathrm{DT}}_{\mathrm{sj}}^t\left(n\right)=\left\{\begin{array}{ccc}\min (1,{\mathrm{DT}}_{\mathrm{sj}}^t(n-1)+a*k_g);& \mathrm{accept}& \mathrm{request}\\ \max (0,{\mathrm{DT}}_{\mathrm{sj}}^t(n-1)-m^2*k_m);& \mathrm{refuse}& \mathrm{request}\end{array}\right.$ Formula 4

WhereinRepresenting the t time period interior nodes s credit rating value to node j, n represents the interactive information article number in the t time period.A is that node j provides the number of times forwarding service, κ continuously for s_gIt is excitation factor, a* κ_gOn the basis of upper once credit rating, cooperative nodes is encouraged in a linear fashion.M is that node j refuses to provide the number of times forwarding service, κ for s continuously_mIt is penalty factor, m²*κ_mBy square in the way of on the basis of upper once credit rating, selfish node is punished.The seriality of a was refused to provide for s to forward service to calculate for initiateing with j the last time, and m is in like manner.

Decaying due to importance in time remote of node historical transactional information, we adopt the quadratic function that Open Side Down to describe the weight w (t) of directly credit value in different time sections, are shown below, it may be assumed that

$w\left(t\right)={(\frac{t}{STL}-1)}^2,(t=0,1,2,...,STL)$ Formula 5

Wherein w (t) is with the t function successively decreased, and STL (storetimelimit) is the reputation information storage time limit of default, and the reputation information record more than STL is no longer participate in calculating.Therefore the direct credit value expression formula of j can be obtained by node s:

${\mathrm{DT}}_{sj}=\frac{1}{STL}\underset{t=1}{\overset{STL}{\&Sigma;}}({\mathrm{DT}}_{sj}^t*w(t\left)\right)$ Formula 6

After the direct credit value of node obtains, use each recommended node recommendation reputation value of behavior similarity Weight and utilize the indirect credit value of normalization factor computing node.In single recommended node situation, the indirect credit value expression formula of objective appraisal node j is by node s:

IDT_sj=μ β formula 7

Wherein μ is the node s direct credit value to recommended node i, and β is the recommended node i direct credit value to objective appraisal node j.

In many recommended nodes situation, when recommending for binode, node h and i has the direct credit value β to node j respectively_hj、β_ijAnd recommend to node s, μ_shAnd μ_siThe respectively node s direct credit value to h, i.Then node s can calculate the indirect credit value IDT of node j_sjFor:

${\mathrm{IDT}}_{sj}=\left(\frac{{\mathrm{\&mu;}}_{sh}}{{\mathrm{\&mu;}}_{sh}+{\mathrm{\&mu;}}_{si}}\right){\mathrm{\&mu;}}_{sh}{\mathrm{\&beta;}}_{hj}+\left(\frac{{\mathrm{\&mu;}}_{si}}{{\mathrm{\&mu;}}_{sh}+{\mathrm{\&mu;}}_{si}}\right){\mathrm{\&mu;}}_{si}{\mathrm{\&beta;}}_{ij}$ Formula 8

Usually, when recommended node number is more than 2, formula 8 is extended to:

$\begin{array}{c}{\mathrm{IDT}}_{sj}=\left(\frac{{\mathrm{\&mu;}}_{s1}}{{\mathrm{\&mu;}}_{s1}+{\mathrm{\&mu;}}_{s2}+\mathrm{...}+{\mathrm{\&mu;}}_{sn}}\right){\mathrm{\&mu;}}_{s1}{\mathrm{\&beta;}}_{1j}+\left(\frac{{\mathrm{\&mu;}}_{s2}}{{\mathrm{\&mu;}}_{s1}+{\mathrm{\&mu;}}_{s2}+\mathrm{...}+{\mathrm{\&mu;}}_{sn}}\right){\mathrm{\&mu;}}_{s2}{\mathrm{\&beta;}}_{2j}+\\ \mathrm{...}+\left(\frac{{\mathrm{\&mu;}}_{sn}}{{\mathrm{\&mu;}}_{s1}+{\mathrm{\&mu;}}_{s2}+\mathrm{...}+{\mathrm{\&mu;}}_{sn}}\right){\mathrm{\&mu;}}_{sn}{\mathrm{\&beta;}}_{nj}\\ =\frac{1}{{\mathrm{\&mu;}}_{s1}+{\mathrm{\&mu;}}_{s2}+\mathrm{...}+{\mathrm{\&mu;}}_{sn}}\underset{m=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&mu;}}_{sm}^2{\mathrm{\&beta;}}_{mj}\end{array}$ Formula 9

WhereinM ∈ 1,2 ..., n} is normalization factor, it is ensured that IDT_sj∈ (0,1).Prove as follows:

By μ_sm∈ (0,1) and β_mj∈ (0,1), can obtain μ_smβ_mj∈ (0,1)；Therefore:

$\begin{array}{c}{\mathrm{IDT}}_{sj}<\frac{{\mathrm{\&mu;}}_{s1}}{{\mathrm{\&mu;}}_{s1}+{\mathrm{\&mu;}}_{s2}+...+{\mathrm{\&mu;}}_{sn}}+\frac{{\mathrm{\&mu;}}_{s2}}{{\mathrm{\&mu;}}_{s1}+{\mathrm{\&mu;}}_{s2}+...+{\mathrm{\&mu;}}_{sn}}+\\ \mathrm{...}+\frac{{\mathrm{\&mu;}}_{sn}}{{\mathrm{\&mu;}}_{s1}+{\mathrm{\&mu;}}_{s2}+...+{\mathrm{\&mu;}}_{sn}}=1\end{array}$ Formula 10

Meanwhile, be easy to get IDT_sj> 0, therefore IDT_sj∈(0,1)。

In order to ensure the verity of recommendation behavior, by behavior similarity Weight recommendation reputation value.The credit rating behavior similarity Sim of s and i (i, s) is shown below, it may be assumed that

$Sim(i,s)=\frac{{\mathrm{\&Sigma;}}_{v\&Element;CZ(s,i)}{\mathrm{\&mu;}}_{sv}{\mathrm{\&mu;}}_{iv}}{\sqrt{{\mathrm{\&Sigma;}}_{v\&Element;CZ(s,i)}^2{\mathrm{\&mu;}}_{sv}^2}\sqrt{{\mathrm{\&Sigma;}}_{v\&Element;CZ(s,i)}{\mathrm{\&mu;}}_{iv}^2}}$ Formula 11

Wherein, (s i) represents there was mutual node set with s and i to C Ζ.Obtain the node s indirect credit value to objective appraisal node j in conjunction with formula 9, (11) be:

${\mathrm{IDT}}_{sj}=\frac{1}{{\mathrm{\&mu;}}_{s1}+{\mathrm{\&mu;}}_{s2}+...+{\mathrm{\&mu;}}_{sn}}\underset{m=1}{\overset{n}{\&Sigma;}}{\mathrm{\&mu;}}_{sm}^2*Sim(m,s)*{\mathrm{\&beta;}}_{mj}$ Formula 12

According to formula 6 and formula 12, the comprehensive credit value of node s computing node j as shown in Equation 13:

T_sj=α * DT_sj+(1-α)*IDT_sjFormula 13

Wherein α ∈ [0,1], it is possible to adjust the size of α, for direct credit value DT_ilWith indirect credit value IDT_ilDistribute different weight preferences.

Node s calls topological structure excitation algorithm and completes self topology adjustment.Node topology structure excitation algorithm is divided into node link control algolithm (LCA-Linkcontrolalgorithm) and node connection request verification algorithm (CVA-Connectionrequestverificationalgorithm).LCA is responsible for deleting and the linking and to high credit value node initiation connection request, and set node self maintained three set of low credit value node, including:

(1) it is familiar with node set FS_s.Represent there was mutual node set with node s.When the topological structure cycle each time arrives, node s calculates the direct credit value of all nodes and indirect credit value in this set, then obtains comprehensive credit value and updates FS_s。FS_s=i | whohascooperatedwiths}.

(2) high credit value node set HTS_s.Represent credit value T_siMore than minimum prestige threshold value T_minNode set.HTS_s=i | T_si> T_min, HTS_s∈FS_s。

(3) neighbor node set NS_s.Expression and node s have the existing node set of annexation.About NS_sThree parameter respectively τ_min: NS_sIn minimum nodes.τ_max: the NS set according to node self computing capability_sIn maximum nodes, τ_max> τ_min。τ_base: the credit value that topological structure process interior joint makes great efforts to safeguard is higher than T_minNodes, τ_min< τ_base< τ_max.Node s is from HTS_sIn choose the node updates NS of high credit value_s, reach at NS_sInterior joint number is not less than τ_baseBasis on optimization NS_sThe purpose of interior joint credit value.NS_s=i | whoconnectswiths}, NS_s∈FS_s。

The LCA of the present invention is divided into three steps to carry out, including:

1, FS is updated according to formula 13_s、HTS_sAnd NS_sIn node credit value.

2、HTS_sSelect FS_sMiddle credit value is more than minimum prestige threshold value T_minAnd not at HTS_sIn node be added, namely ${\mathrm{HTS}}_s+\left\{i\right|i\&Element;{\mathrm{FS}}_s\&cap;i\&NotElement;{\mathrm{HTS}}_s\&cap;T_i>T_{min}\}.$

3, first by NS_sMiddle credit value is lower than T_minThe link of node delete because mutual through after a while between node, it may appear that some node credit values reduce and not in the situation of satisfied requirement；Then, if NS_sInterior joint number N um (NS_s) < τ_base, from HTS_sNode e=Max_T (the HTS that prioritizing selection credit value is big_s), if NS_sDo not comprise e and then send connection request to e, if request is by verifying, NS will be added to_s, i.e. NS_s+ { e} otherwise continues traversal Max_T (HTS_s-e}), until Num (NS_s)≥τ_base.If τ_base≤Num(NS_s)≤τ_max, then NS is compared_sMiddle credit value minimum node p=Min_T (NS_s) and q=Max_T (HTS_s) credit value size, if credit value T_p< T_qThen to q transmission connection request, if request is by verifying, deletes the connection with p the connection of foundation and q, repeat said process and NS_s-{Min_T(NS_s) and NS_s+{Max_T(HTS_s).The mark that LCA algorithm terminates is Num (NS_s) between τ_baseAnd τ_maxBetween and NS_sThe credit value of interior joint is all higher than HTS_sThe credit value of interior joint or HTS_sFor sky, being described as of formulation:

Formula 14

Owing to node is more prone to the node cooperation high with credit value and refuses the connection request of low credit value node, therefore some credit value relatively low but more than T_minIts Num of node (NS) long-term less than τ_base, it is impossible to improve its credit value by the effective cooperation with neighbor node, thus causing vicious cycle.For the problems referred to above, it is assumed that n topological structure periodic knot Num (NS) < τ continuously_base, then τ is reduced by decline factor delta ∈ (0,1)_baseValue be τ '_base, the definition of δ is as shown in Equation 11, it may be assumed that

$\mathrm{\&delta;}=\frac{{\mathrm{\&tau;}}_{base}-{\mathrm{\&tau;}}_{base}^{\&prime;}}{{\mathrm{\&tau;}}_{base}-{\mathrm{\&tau;}}_{\min }}$ Formula 15

τ ' can be solved by formula 11_base:

τ'_base=(1-δ) * τ_base+δ*τ_minFormula 16

CVA is responsible for the credit value of the node of assessment initiation connection request and chooses whether to be attached with it according to the interstitial content in self neighbor node set.Assuming that connection request node is s, requested node is r.Neighbor node quantity Num (NS according to r_r) and τ_maxMagnitude relationship be divided into two kinds of situations to judge, including:

1, the neighbor node quantity Num (NS of node r_s)≥τ_max.If not for the history credit rating information of s (namely be strange node to r, s) in node r, then refusal accepts the connection request of s；Otherwise, if the connection request accepting s depends on the credit value T of s_sWith Min_T (NS_r) credit valueMagnitude relationship between the two, ifThen replace s and Min_T (NS_r)；Otherwise refuse its connection request.

2, the neighbor node quantity Num (NS of node r_r) < τ_max.If node r has the credit value T of the history credit rating information to s and s_s> T_min, then its connection request is accepted.If not history credit rating information to s in node r, then r is with probability P_AccAccept the connection request of s, and give its initial credit value 0.5.P_AccDefinition as shown in Equation 13, its implication is: it is more big that more few its of the neighbor node of node r accepts the probability of s connection request, otherwise more little.

$P_{Acc}=1-\frac{Num\left({\mathrm{NS}}_r\right)}{{\mathrm{\&tau;}}_{max}}$ Formula 17

As it is shown in figure 1, the invention provides a kind of topological structure motivational techniques based on credit value, the process that implements of the method includes as follows:

Step 1: select the performance-optimal network region k near business demand node s according to formula 2^*And correspondence via node collection R^*。

Step 2: business demand node s with broadcast mode to R^*Middle via node sends service request and records via node interactive information, and 0 represents reject the service request, and 1 expression accepts request and elected as service providing node, service providing node r by business demand node s^*Determine by formula 3.

Step 3: when the topological structure cycle arrives, using formula 6 calculates business demand node s to FS_sThe direct credit value DT of interior joint_sj。

Step 4: on direct credit value basis, adopts behavior similarity mathematical model shown in formula 11 to calculate business demand node s to FS_sThe indirect credit value IDT of interior joint_sj, and the direct credit value DT of integrating step 3 gained_sjBring formula 13 into and calculate business demand node s to FS_sThe comprehensive credit value T of interior joint_sj。

Step 5: after the comprehensive credit value of node updates, business demand node s calls topological structure excitation algorithm and completes self topology adjustment.

A kind of topological structure motivational techniques based on the credit value above embodiment of the present invention provided are described in detail, for one of ordinary skill in the art, thought according to the embodiment of the present invention, all will change in specific embodiments and applications, in sum, this specification content should not be construed as limitation of the present invention.

Claims (6)

1. the topological structure motivational techniques based on credit value, it is characterised in that described method setting wireless Ubiquitous Network communication scenes, it is assumed that there is the business demand node s in the A of congested network region, ambient idle network area collection Κ={ k in this scene₁,k₂,…k_nAnd Κ in via node collection Γ={ R corresponding to each dormant network region₁,R₂,…R_n, comprise the steps:

Step 1: business demand node s selects performance-optimal network region k according to network capacity c and offered load l two indices from set Κ^*And correspondence via node collection R^*；

Step 2: business demand node s with broadcast mode to R^*In via node send service request and consider remaining power energy e, node load l' according to principle of optimality from R^*The optimum via node r of middle selection^*As service providing node, business demand node s records via node interactive information, and 1 represents acceptance request and elected as service providing node by business demand node s, and 0 represents reject the service request；

Step 3: when the topological structure cycle arrives, according to history mutual information between node, τ calculates credit value in each time period at timed intervalsUse weighting function w (t) weighting decayed in timeObtain the direct credit value of node；

Step 4: on direct credit value basis, each recommended node recommendation reputation value of utilization behavior similarity Weight also utilizes the indirect credit value of normalization factor computing node, calculate the direct credit value of gained and the indirect credit value of this step gained according to step 3, calculate comprehensive credit value T_sj；

Step 5: after the comprehensive credit value of node updates, business demand node s calls topological structure excitation algorithm and completes self topology adjustment.

2. a kind of topological structure motivational techniques based on credit value according to claim 1, it is characterised in that the described direct credit value computational methods of method step 3 interior joint include:

First, its recorded node history mutual information is divided in different time sections by business demand node s τ at timed intervals, then calculates the credit value in each time period, and computing formula is as follows:

${\mathrm{DT}}_{sj}^t\left(n\right)=\left\{\begin{array}{ccc}\min \left(\begin{array}{cc}1,& {\mathrm{DT}}_{sj}^t\left(n-1\right)+a*{\mathrm{\&kappa;}}_g\end{array}\right);& accept& forwarding\\ \max \left(\begin{array}{cc}0,& {\mathrm{DT}}_{sj}^t\left(n-1\right)-m^2*{\mathrm{\&kappa;}}_m\end{array}\right);& refuse& forwarding\end{array}\right.$

WhereinRepresenting the t time period interior nodes s credit rating value (its initial value is 0.5 within each time period) to node j, n represents the interactive information article number in the t time period, and a is that node j provides the number of times forwarding service, κ continuously for s_gIt is excitation factor, a* κ_gOn the basis of upper once credit rating, cooperative nodes being encouraged in a linear fashion, m is that node j refuses to provide the number of times forwarding service, κ for s continuously_mIt is penalty factor, m²*κ_mBy square in the way of on the basis of upper once credit rating, selfish node is punished, the seriality of a was refused to provide for s to forward service to calculate for initiateing with j the last time, m is in like manner, the feature decayed to characterize the importance of node historical transactional information in time remote, utilize the characteristic of quadratic function to describe the weight w (t) of directly credit value in different time sections, it is shown below, it may be assumed that

$w\left(t\right)={(\frac{t}{STL}-1)}^2,(t=0,1,2,...,STL)$

Wherein w (t) is with the t function successively decreased, STL (storetimelimit) is the reputation information storage time limit of default, reputation information record more than STL is no longer participate in calculating, and is weighted on average obtaining the node s direct credit value to j to credit value in different time sections with w (t):

${\mathrm{DT}}_{sj}=\frac{1}{STL}\underset{t=1}{\overset{STL}{\&Sigma;}}({\mathrm{DT}}_{sj}^t*w(t\left)\right).$

3. a kind of topological structure motivational techniques based on credit value according to claim 1, it is characterised in that the described indirect credit value computational methods of method step 4 interior joint include:

In single recommended node situation, the direct credit value of recommended node i is μ by node s, it is recommended that the direct credit value of objective appraisal node j is β by node i, then the indirect credit value of objective appraisal node j is μ β by node s；In many recommended nodes situation, in order to ensure the verity of recommendation behavior, by behavior similarity Weight recommendation reputation value, the credit rating behavior similarity Sim of s and i (i, s) is shown below, it may be assumed that

$Sim(i,s)=\frac{{\mathrm{\&Sigma;}}_{v\&Element;CZ(s,i)}{\mathrm{\&mu;}}_{sv}{\mathrm{\&mu;}}_{iv}}{\sqrt{{\mathrm{\&Sigma;}}_{v\&Element;CZ(s,i)}{\mathrm{\&mu;}}_{sv}^2}\sqrt{{\mathrm{\&Sigma;}}_{v\&Element;CZ(s,i)}{\mathrm{\&mu;}}_{iv}^2}}$

Wherein C Ζ (s, i) represent with s and i had mutual node set, as weight, behavior similarity is described the node s indirect credit value to objective appraisal node j is:

${\mathrm{IDT}}_{sj}=\frac{1}{{\mathrm{\&mu;}}_{s1}+{\mathrm{\&mu;}}_{s2}+...+{\mathrm{\&mu;}}_{sn}}\underset{m=1}{\overset{n}{\&Sigma;}}{\mathrm{\&mu;}}_{sm}^2*Sim(m,s)*{\mathrm{\&beta;}}_{mj}$

Wherein μ_smRepresent node s to recommended node m ∈ 1,2 ..., the direct credit value of n}, β_mjRepresent the recommended node m direct credit value to objective appraisal node j；For direct credit value DT_sjWith indirect credit value IDT_sjDistributing different weight preferences, obtain the node s comprehensive credit value to objective appraisal node j, expression formula is as follows, it may be assumed that

T_sj=α * DT_sj+(1-α)*IDT_sj

Wherein α ∈ [0,1].

4. a kind of topological structure motivational techniques based on credit value according to claim 1, it is characterized in that, in described method step 5, topological structure excitation algorithm is divided into node link control algolithm (i.e. LCA, and node connection request verification algorithm (i.e. CVA, Connectionrequestverificationalgorithm) Linkcontrolalgorithm)；LCA is responsible for deleting and the linking and to high credit value node initiation connection request, and set node self maintained three set of low credit value node, including:

(1) it is familiar with node set FS_s, represent there was mutual node set with node s, when the topological structure cycle each time arrives, node s calculates the direct credit value of all nodes and indirect credit value in this set, then obtains comprehensive credit value and updates FS_s, FS_s=i | whohascooperatedwiths}；

(2) high credit value node set HTS_s, represent credit value T_siMore than minimum prestige threshold value T_minNode set, HTS_s=i | T_si> T_min, HTS_s∈FS_s；

(3) neighbor node set NS_s, expression and node s have the existing node set of annexation；NS_sThree parameter respectively τ_min: NS_sIn minimum nodes, τ_max: the NS set according to node self computing capability_sIn maximum nodes, τ_max> τ_min, τ_base: the credit value that topological structure process interior joint makes great efforts to safeguard is higher than T_minNodes, τ_min< τ_base< τ_max, node s is from HTS_sIn choose the node updates NS of high credit value_s, reach at NS_sInterior joint number is not less than τ_baseBasis on optimization NS_sThe purpose of interior joint credit value, NS_s=i | whoconnectswiths}, NS_s∈FS_s；CVA is responsible for the credit value of the node of assessment initiation connection request and chooses whether to be attached with it according to the interstitial content in self neighbor node set.

5. a kind of topological structure motivational techniques based on credit value according to claim 1, it is characterized in that, described method is to utilize network performance index to select optimum dormant network region, based on history mutual information, the direct credit value to Node evaluation and indirect credit value is calculated by building effective mathematical model, thus obtaining the comprehensive credit value of egress, then pass through periodic topological structure excitation algorithm and adjust the topological structure that node connects so that cooperative nodes is connected with each other and selfish node is ostracised network edge.

6. a kind of topological structure motivational techniques based on credit value according to claim 1, it is characterised in that described method be applied to wireless general in the environment.