CN103647671A - Gur Game based crowd sensing network management method and system - Google Patents

Abstract

The invention discloses a Gur Game based crowd sensing network management method and system. The method comprises the steps of sending information quality requirements of sensing tasks to user intelligent equipment; performing information fusion on information contribution amount initial values and calculating the information contribution amount assigned to each user through a Gur Game iteration process; performing participant selection and feeding back selection results back to the user intelligent equipment; obtaining returns after the selected users complete the tasks. The system comprises a crowd sensing server, a network platform information center and a user intelligent terminal. The crowd sensing server and the user intelligent terminal perform information interaction through the network platform information center. By means of the method and the system, the automata theory and the iterative computation are applied in sensing task assignment calculation, so that the complexity in implementation is reduced; by raising a reverse auction reward mechanism, the reward expenditure of operators is reduced while user participation is promoted, and a win-win situation is achieved.

Description

A kind of gunz aware network management method and system thereof based on Gur Game

Technical field

The invention belongs to technology of wireless sensing network field, be specifically related to a kind of gunz aware network management method and system thereof based on Gur Game.

Background technology

Occurred in recent years cheap, wireless and be easy to the mobile device of programming, for example smart mobile phone and panel computer, with embedded type sensor as accelerometer, gyroscope, GPS, camera and microphone.The multimedia that these are integrated and position tracking function are to enable " property of participation remote sensing " model of multiple new application program and innovation.The smart machine that its task has been disposed, forms interactive and sensor network participatory, makes the public and professional user, collection, analysis and shared Indigenous knowledge.

Mass-rent refers to Yi Ge company or mechanism were carried out past task by employee, is contracted out to the way of unspecific (and normally large-scale) popular network with free voluntary form.The task of mass-rent is normally born by individual, if but relate to task of needing multiple person cooperational to complete, also likely to rely on the form of the individuality production of increasing income to occur.The smart mobile phone of today has fundamentally changed the understanding of traditional " mass-rent ", the property of participation of an emerging type, the application program of oriented mission.It is intended to support so-called " effort of citizen's science " for Knowledge Discovery, understand their suggestion of human behavior and measure/assess.

One of traditional method is CenceMe project, and it carries out perception by mobile phone transducer to people's animation, and releases news from trend social networks; Traditional other method is EmotionSense project, and it is by service condition and the psychological knowledge of mobile phone, analyst's social situation, the information of mood state; A traditional method is again Citizen noise pollution monitoring, and the method by mobile phone Mike and GPS locator, is added up city noise pollution situation everywhere automatically, and provides " Noise map " for user.But because above-mentioned conventional method has, lack the design of the reasonable dispatched mode of perception task and lack the limitation that user recruits measure, the large data analysis and processing, information quality assurance, secret protection, user management and the excitation user participation mechanism distributed and energy efficient that therefore how to solve in mass-rent process become a difficult problem in the urgent need to address.

Summary of the invention

In order to overcome the above-mentioned defect of prior art, one of object of the present invention be to propose a kind of simple, robustness is high, realize the gunz aware network management method based on Gur Game of perception task reasonable distribution.For this reason, the present invention is achieved through the following technical solutions:

A gunz aware network management method based on Gur Game, comprises the steps:

Gunz aware services device sends the information quality demand of perception task to each user's smart machine by network platform information centre;

Each user's smart machine is sent to network platform information centre by contribute information amount initial value respectively, described network platform information centre carries out after use processing contribute information amount initial value, consider user's smart machine status of energy consumption and user awareness capacity of water, by Gur Game iterative process, calculate the contribute information amount of distributing to each user;

Each user's smart machine is sent to network platform information centre according to the contribute information amount of distributing by required remuneration, by described network platform information centre, is carried out participant's selection and selection result is fed back to each user's smart machine;

When selected user completes after this subtask, described network platform information centre is to user's payt.

Further, the method that described network platform information centre carries out use processing to contribute information amount initial value is:

Contribute information amount is carried out to partition of the level, in each rank, include a stable state and at least one intermediateness;

Foundation is with the automatic Gur Game structure of chain, by Gur Game structure determine each state respectively with the relation of Reward-Penalty Functions, described Reward-Penalty Functions comprises the reward function of following formula

with penalty P ^j _i(k):

$R_i^j\left(k\right)=g(I(k-1),j)$

${P_i}^j\left(k\right)=1-R_i^j\left(k\right)$

In formula, I (k-1) is the information quality factor, $I(k-1)=\tanh \left(\mathrm{\η}\ln \frac{u^a(k-1)}{u^r}\right);$

U ^a(k-1) be all user profile contribution amount initial value sums,

Wherein, p _ifor information Effective Probability, u _ifor each user's contribute information amount initial value, N is total number of users, i ∈ N, j ∈ N, u ^rinformation quality demand for perception task.

Further, described foundation with the method for the Gur Game structure of automatic chain is:

Determine the stable state of contribute information amount;

Determine intermediateness number corresponding with stable state in each rank;

All states connect successively by automatic chain;

The Reward-Penalty Functions migratory direction of stable state and intermediateness is set:

If the reward function of stable state, moves back this stable state; If the penalty of stable state, migrates to respectively all intermediatenesses adjacent with this stable state;

If the reward function of intermediateness, jumps to the corresponding stable state of this intermediateness; If the penalty of intermediateness, towards with this intermediateness the rightabout of corresponding stable state migrate to next adjacent stable state or intermediateness;

Further, by Gur Game iterative process, the method that calculates the contribute information amount of distributing to each user is:

First, M=N is set, M is for finally selecting the participation value number of carrying out this subtask; Make all users' contribute information amount surpass perception task information quality demand,

use parameter

replace the parameters u in Gur Game structure ^r;

Then, by Gur Game algorithm, calculate alternative parameter u ^rafter distribute to each user's contribute information amount

Further, described Gur Game algorithm comprises the steps:

A1) initial value that current iteration number of times k is set is 0, and it is K that total iterations is set;

A2) initialization user smart machine;

A3) calculate all user profile contribution amount initial value sums ${\tilde{u}}^a(k-1),{\tilde{u}}^a(k-1)=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{p}_i{\tilde{u}}_i;$ If

or

perform step A4), otherwise jump to steps A 6); ζ ∈ [0,0.2] wherein;

A4) by following various computing information quality factor respectively

reward function

with penalty P ^j _i(k):

$\tilde{I}(k-1)=\tanh \left(\mathrm{\η}\ln \frac{{\tilde{u}}^a(k-1)}{{\tilde{u}}^r}\right);$

$R_i^j\left(k\right)=g(\stackrel{\‾}{I}(k-1),j);$

${P_i}^j\left(k\right)=1-R_i^j\left(k\right);$

A5) Gur Game structure generation random number seed ∈ [0,1], by judging the size of this random number and reward function, further determines state transition direction:

If

according to reward function, point to and migrate to NextState; Otherwise, by penalty, point to and migrate to NextState, jump to steps A 7);

A6) calculate the mean value that current all user's energy consume and further judge state transition direction; In formula, the energy consumption of each user i is

in formula,

for dump energy, E _ifor primary power, γ is the transforming factor being consumed to energy by perception information,

contribute information amount initial value for each user;

A7) k=k+1 is set, if k>K exits this iterative computation; Otherwise, enter next iteration, jump to steps A 3);

A8) select k ^*the result of step is as the contribute information amount of distributing to each user,

If exist in iterative process ${\tilde{u}}^a\left(k\right)\∈[{\tilde{u}}^r,(1+\mathrm{\ζ}){\tilde{u}}^r],$ ? $k^{*}=\arg {\min }_k({\tilde{u}}^a\left(k\right)-{\tilde{u}}^r),\∀{\tilde{u}}^a\left(k\right)\∈({\tilde{u}}^r,\∞);$

Otherwise $k^{*}=\arg {\min }_k{\tilde{u}}^a\left(k\right),{\tilde{u}}^a\left(k\right)\∈[{\tilde{u}}^r,(1+\mathrm{\ζ}){\tilde{u}}^r],$ Wherein, ζ ∈ [0,0.2].

Further, in described steps A 6, judge that the method for state transition direction is:

If $\frac{\mathrm{\&xi;}(k-1)-\stackrel{\&OverBar;}{\mathrm{\&xi;}}(k-1)}{\stackrel{\&OverBar;}{\mathrm{\&xi;}}(k-1)}>\&Element;,$ Migration left; If $\frac{\mathrm{\&xi;}(k-1)-\stackrel{\&OverBar;}{\mathrm{\&xi;}}(k-1)}{\stackrel{\&OverBar;}{\mathrm{\&xi;}}(k-1)}<-\&Element;,$ Migration to the right; Otherwise, keep current state constant.

Further, described network platform information centre carries out the method for participant's selection and is:

B1) the remuneration b sending to network platform information centre according to user _in user sorted, make b ₁≤ b ₂≤ ... ≤ b _n, another i=N;

B2) if ${\tilde{u}}^a-p_i{\tilde{u}}_i\&GreaterEqual;u^r,$ Order ${\tilde{u}}^a={\tilde{u}}^a-p_i{\tilde{u}}_i,M=M-\left\{i\right\},i=i-1,$ Re-execute this step B2; Otherwise, turn to step B3;

B3) finally selecting the participant who carries out this subtask is M user,

Another object of the present invention is to propose a kind of gunz aware network management system based on Gur Game, this system comprises gunz aware services device, network platform information centre and user's intelligent terminal, between described gunz aware services device and user's intelligent terminal, by network platform information centre, carry out information interaction, wherein:

Described gunz aware services device, for gathering and store user's perception information around, also for send the information quality demand of perception task and the final perception information that receives user's intelligent terminal feedback to user's intelligent terminal;

Described network platform information centre, carries out use processing for the contribute information amount initial value to receiving;

Described user's intelligent terminal, the contribute information amount of giving each user for dispensed, also for sending the required remuneration of user to network platform information centre.

Further, described network platform information centre comprises:

Perception task dispatch module, for being distributed to each user's intelligent terminal by wireless senser by the information quality demand of perception task;

Data fusion analysis platform, for carrying out use processing to the perception information receiving;

Described user's stimulating platform, for selecting the participant who carries out this subtask.

Further, described user's intelligent terminal comprises:

Electric weight monitoring module, for the dump energy of supervisory user intelligent terminal;

Gur Game sensed activation recommended engine, for sending contribute information amount initial value and by determining with the interaction of the network platform information centre contribute information amount of distributing to each user to network platform information centre;

The module of bidding, recommends to send amount of money award to network platform information centre for the preliminary perception action of sending according to Gur Game sensed activation recommended engine, also for completing after this subtask as user, and the amount of money award that reception network platform information is returned;

Sensed activation decision module, selects information for receiving the participant that network platform information sends, if selected, informs that user carries out this subtask, final perception information is returned to gunz aware services device by network platform information centre simultaneously.

Compared with prior art, the beneficial effect that the present invention reaches is:

1) method and system of the present invention is applied to automaton theory in perception task Distribution Calculation, traditional Gur Game theory is improved in automaton structure and the division of state iteration, when realizing distributed and high robust, reduced the complexity in implementing;

2) method and system of the present invention can, according to user's dump energy state and perception size, be realized the reasonable distribution of perception task;

3) method and system of the present invention completes the Distribution Calculation of perception task by limited number of time iteration, greatly reduces the complexity of enforcement;

4) in method and system of the present invention, pass through to propose the reward mechanism of reverse auction, when promoting that user participates in, reduced the award expenditure of operator, realize doulbe-sides' victory.

Accompanying drawing explanation

Fig. 1 is the Organization Chart that the present invention is based on the gunz aware network management system of Gur Game;

Fig. 2 is the Gur Game structural representation in the present invention, u in figure _max=2, β=2, have maximum 2 intermediatenesses under each stable state, in figure, arrow shows the migratory direction of each state under award or penalty;

Fig. 3 is the example with the reward function of different j and α;

Fig. 4 is the drawing of the award machine based on reverse auction in the present invention;

Fig. 5 rewards the changing trend diagram of probability with iterations;

Fig. 6 is that the information quality that obtains is with the changing trend diagram of iterations;

Fig. 7 is that difference is bidded under pattern and rewarded the changing trend diagram of total expenditure with simulation time scope.

Embodiment

Below in conjunction with accompanying drawing, the specific embodiment of the present invention is described in further detail.

In this example, proposed a kind of gunz aware network management system based on Gur Game, this system was comprised of two stages.First stage relates to the interaction between network platform information centre and the built-in Gur Game engine of user's smart machine.The input of Gur Game engine is user's smart machine dump energy level and the QoI demand of perception task.In the iterative step each time of Gur Game, smart machine is by the output of Gur Game iteration, and the information quality contribution level of preparation, returns described to network platform information centre.Then the latter, according to all participants' that receive provisioning information quality contribution level, calculates award-probability penalty, and returns it to each user's Gur Game engine.Based on this feedback, each user's Gur Game automaton changes its current state, and produces new perception action, i.e. new contribute information level.Gur Game engine produces best result by the method that iterates, can meet QoI requirement, energy consumption that again can balance different user.

Second step relates to mutual between module of bidding of network platform information centre and user's smart machine.After the first step, the network platform obtains each user's preliminary information.Then, user sends to platform their bid information, and this just represents their the expectation payment remuneration to the contribute information of Board Lot.User's stimulating platform is selected to meet perception mission bit stream quality requirement, can reduce again the user group of award total expenditure as final contribution user, these end users, according to the determined final information quality contribution level of Gur Game engine in the first step, are mail to data fusion analysis platform by perception data.User's stimulating platform according to user's bid information to its payt.

As shown in Figure 1, this system comprises gunz aware services device, network platform information centre and user's intelligent terminal, between gunz aware services device and user's intelligent terminal, by network platform information centre, carries out information interaction, wherein:

1) gunz aware services device, for gathering and store user's perception information around, also for send the information quality demand of perception task and the final perception information that receives user's intelligent terminal feedback to user's intelligent terminal;

2) network platform information centre, carries out use processing for the contribute information amount initial value to receiving; The detailed structure at this center is as follows:

Perception task dispatch module, for being distributed to each user's intelligent terminal by wireless senser by the information quality demand of perception task;

Data fusion analysis platform, for carrying out use processing to the perception information receiving;

Described user's stimulating platform, for selecting the participant who carries out this subtask.

3) user's intelligent terminal, the contribute information amount of giving each user for dispensed, also, for send the required remuneration of user to network platform information centre, the detailed structure of this terminal is as follows:

Electric weight monitoring module, for the dump energy of supervisory user intelligent terminal;

Gur Game sensed activation recommended engine, for sending contribute information amount initial value and by determining with the interaction of the network platform information centre contribute information amount of distributing to each user to network platform information centre;

The module of bidding, recommends to send amount of money award to network platform information centre for the preliminary perception action of sending according to Gur Game sensed activation recommended engine, also for completing after this subtask as user, and the amount of money award that reception network platform information is returned;

Sensed activation decision module, selects information for receiving the participant that network platform information sends, if selected, informs that user carries out this subtask, final perception information is returned to gunz aware services device by network platform information centre simultaneously.

In this example, also introduce a kind of gunz aware network management method based on Gur Game, the detailed process that realizes the method is as follows:

Steps A, gunz aware services device send the information quality demand of perception task to each user's smart machine by network platform information centre;

Information quality demand, refers to the ageing etc. of information accuracy, information integrity, information, is abstract to application information needed speciality, and its definition is independent of mode and the process of application use information.The information quality demand of gunz perception task, often needs participant that the multimedia messages based on context environmental is provided, and the ageing and integrality of information has higher requirements.

Step B, each user's smart machine are sent to network platform information centre by contribute information amount initial value respectively, described network platform information centre carries out after use processing contribute information amount initial value, consider user's smart machine status of energy consumption and user awareness capacity of water, by Gur Game iterative process, calculate the contribute information amount of distributing to each user;

In this step B, the detailed method that network platform information centre carries out use processing and Gur Game iterative process to contribute information amount is:

B1, contribute information amount is carried out to partition of the level, set up with the Gur Game structure of chain automatically.

Gur Game is a kind of distributed algorithm based on finite-state automata (Finite State Automaton), the state that each individual automaton comprises similar number, and wherein half is " participation " state, half is " free time " state.The initial condition of each automaton is random appointment, according to " rewarding-punishment " probability function of definition, changes independently between adjacent states.Iteration all can be upgraded Gur Game " award-punishment " probability each time, thereby affects the state transition of next iteration.By each state transition like this, final all automatons reach the optimum state of convergence.This algorithm is completely distributed, it is the energy consumption state that user both need not predict other users, also need not carry out any type of information interaction with other users, pass through independently iteration adjustment state of automata completely, by the method for repetition test, reach the global optimum of system.Inventor greatly expands traditional Gur Game, makes it to have the ability that multiclass sensed activation is recommended of making, and each recommends (being sensed activation) corresponding to contribute information to a certain degree.Traditional Gur Game algorithm determines for each automaton only provides two kinds of candidates, i.e. " participation " and " free time ".Inventor expands it, and the information quality of the required contribution of user is divided into some ranks, as word, voice, image, video and corresponding characteristic (as resolution etc.).Each rank is all bound one by one with the state of automaton, and then the automaton of a plurality of traditional Gur Game is merged into one with the automaton of a plurality of stable states.Like this, defined new Gur Game automaton model will have the ability that multiple propelling movement determines of making.

State in Gur Game automaton structure is divided into two kinds, and a kind of is stable state, and a kind of is intermediateness.In example shown in Fig. 2, there is 0,1,2 three stable state, represent successively to complete the partition of the level to contribute information amount by three information quality contribution levels from low to high.Each stable state has again left and right two attached intermediatenesses (stable state of two sections only has an attached intermediateness) simultaneously.The number of intermediateness is more, and it is slower that Gur Game iterative process reaches the speed of convergence; If intermediateness number is very few, also may cannot reach iteration convergence because state transition crosses violent.Therefore, the number of intermediateness should be adjusted according to actual conditions.

B2, definite " rewarding-punishment " probability function

Between state and state, by " rewarding-punishment " probability function, move, it distributes as shown in Figure 2.During migration, if the reward function of stable state moves back this stable state; If the penalty of stable state, migrates to respectively all intermediatenesses adjacent with this stable state; If the reward function of intermediateness, jumps to the corresponding stable state of this intermediateness; If the penalty of intermediateness, towards with this intermediateness the rightabout of corresponding stable state migrate to next adjacent stable state or intermediateness.Reward-Penalty Functions comprises the reward function of following formula

with penalty P ^j _i(k):

$R_i^j\left(k\right)=g(I(k-1),j)$

${P_i}^j\left(k\right)=1-R_i^j\left(k\right)$

In formula, I (k-1) is the information quality factor, u ^a(k-1) be all user profile contribution amount initial value sums,

wherein, p _ifor information Effective Probability, u _ifor each user's contribute information amount initial value, N is total number of users, i ∈ N, j ∈ N, u ^rinformation quality demand for perception task.By this design, the effect reaching as shown in Figure 3, transverse axis is the information quality factor, the longitudinal axis is for rewarding probability, its value not only depends on the current information quality that can reach, also relevant with current residing state of automata: if current total information quality contribution lower than information quality demand, the user that contribution margin is less has the higher probability moving to high contribution level state; If current total information quality contribution is higher than information quality demand, the user that contribution margin is larger has the higher probability moving to low contribution level state.Therefore, state transition is able to correct guidance, has accelerated system convergence rate.

B3, Gur Game iterative process

First, each user sends a random value to network platform information centre, represents the information quality contribution level of its preparation, and the initial condition of its Gur Game automaton is exactly certain stable state that this random value is corresponding.Network platform information centre is after collecting all users' information quality contribution level, and by information fusion, computing information quality factor, then according to the residing state position of each user, calculates " rewarding-punishment " probability, and feed back to all users.User receives after " rewarding-punishment " probability, carries out state transition, and then produces new provisioning information quality contribution level, and this value is mail to network platform information centre again, thereby enter next round iterative process.If all users' information quality contribution summation differs larger with the information quality demand of perception task, can one straight through " reward-punish " probability, carry out state transition; If the gap of the information quality demand of all users' information quality contribution summation and perception task reaches in certain limit, by the mode of " moving to left " and " moving to right ", information quality contribution level to each user is finely tuned, when its summation is slowly changed, the energy progressively reducing between each user consumes level of difference, reaches certain user profile contribution fairness.After iteration surpasses default maximum times, choose in historical iterative information and can meet information quality demand, can reduce again one group of contribute information level that user's energy consumes level of difference and distribute, as final result feedback to each user.

Step C, each user's smart machine are sent to network platform information centre according to the contribute information amount of distributing by required remuneration, by described network platform information centre, carried out participant's selection and selection result is fed back to each user's smart machine.When selected user completes after this subtask, described network platform information centre is to user's payt, and process as shown in Figure 4.

The method that described network platform information centre carries out participant's selection is in fact a kind of reward mechanism based on reverse auction.

C1) the remuneration b sending to network platform information centre according to user _in user sorted, make b ₁≤ b ₂≤ ... ≤ b _n, another i=N;

C2) if ${\tilde{u}}^a-p_i{\tilde{u}}_i\&GreaterEqual;u^r,$ Order ${\tilde{u}}^a={\tilde{u}}^a-p_i{\tilde{u}}_i,M=M-\left\{i\right\},i=i-1,$ Re-execute this step B2; Otherwise, turn to step B3;

C3) finally selecting the participant who carries out this subtask is M user,

Due to before Gur Game iterative process, make all users' contribute information amount surpass perception task information quality demand,

therefore above-mentioned user choosing method can't cause perception task to be not being met.Through proof, the method can be removed the higher user of quotation, can make again total information quality meet the demand of perception task.

Inventor has further verified that by following test data the present invention has significant effect:

Fig. 5 rewards the changing trend diagram of probability with iterations.Can find out, along with the increase of iterations, reward probability and move closer to maximum 1.In conjunction with Fig. 2, can find out, this Gur Game automaton that represents this each user all will stably remain on certain stable state, thereby the iterative process of whole system reaches convergence.

Fig. 6 is that the information quality that obtains is with the changing trend diagram of iterations.Can find out, along with the increase of iterations, all users' information quality contribution sum, moves closer to the information quality demand of perception task, and the most always reaches convergence.In conjunction with Fig. 5 and Fig. 6, can observe out, reward trough and the information quality of probability curve and contribute the crest of sum to there is relation one to one.

Fig. 7 is that difference is bidded under pattern and rewarded the changing trend diagram of total expenditure with simulation time scope.In contrasted pedestal method, all users' contribute information amount sum equals perception task information quality demand, and all use is final selecteed user per family, there is not user's selection course, the network platform is provided final award according to the consideration value of user's application to it.In addition, consider whether user is chosen as final contribution user's front and back, may change its bid price, and different user's behaviors of bidding is made to two kinds of modelings:

1) self adaptation changing is in proportion bidded:

$b_i=\left\{\begin{array}{c}(1+k)b_i,i\&Element;M,\\ (1-k)b_i,i\&NotElement;M,\end{array}\right.$

The κ >0 variation ratio that represents to bid wherein.

2) the fixing self adaptation changing is bidded: wherein χ is the variable quantity of bid, χ >0

$b_i=\left\{\begin{array}{c}{b}_i+\mathrm{\&chi;},i\&Element;M,\\ b_i-\mathrm{\&chi;},i\&NotElement;M,\end{array}\right.$

Observing Fig. 7 can find out, during the fixing self adaptation changing is bidded, because the level difference of bidding between each user is less, the advantage that user selects is not significantly embodied; And in the self adaptation of fixedly bidding and change is in proportion bidded, because the level difference of bidding between each user is larger, by implementing user's selection algorithm, can effectively reduce award total expenditure.

Finally should be noted that: above embodiment is only in order to illustrate the application's technical scheme but not restriction to its protection range; although the application is had been described in detail with reference to above-described embodiment; those of ordinary skill in the field are to be understood that: those skilled in the art still can carry out all changes, revise or be equal to replacement to the embodiment of application after reading the application; these change, revise or be equal to replacement, within the claim scope that it all awaits the reply in its application.

Claims (10)

1. the gunz aware network management method based on Gur Game, is characterized in that, comprises the steps:

Gunz aware services device sends the information quality demand of perception task to each user's smart machine by network platform information centre;

Each user's smart machine is sent to network platform information centre by contribute information amount initial value respectively, described network platform information centre carries out after use processing contribute information amount initial value, by Gur Game iterative process, calculate the contribute information amount of distributing to each user;

Each user's smart machine is sent to network platform information centre according to the contribute information amount of distributing by required remuneration, by described network platform information centre, is carried out participant's selection and selection result is fed back to each user's smart machine;

When selected user completes after this subtask, described network platform information centre is to user's payt.

2. the method for claim 1, is characterized in that, the method that described network platform information centre carries out use processing to contribute information amount is:

Contribute information amount is carried out to partition of the level, in each rank, include a stable state and at least one intermediateness;

Foundation is with the Gur Game structure of automatic chain;

By Gur Game structure determine each state respectively with the relation of Reward-Penalty Functions, described Reward-Penalty Functions comprises the reward function of following formula

with penalty P ^j _i(k):

In formula, I (k-1) is the information quality factor,

u ^a(k-1) be all user profile contribution amount initial value sums,

wherein, p _ifor information Effective Probability, u _ifor each user's contribute information amount initial value, N is total number of users, i ∈ N, j ∈ N, u ^rinformation quality demand for perception task.

3. method as claimed in claim 2, is characterized in that, described foundation with the method for the Gur Game structure of automatic chain is:

Determine the stable state of contribute information amount;

Determine intermediateness number corresponding with stable state in each rank;

All states connect successively by automatic chain;

The Reward-Penalty Functions migratory direction of stable state and intermediateness is set:

If the reward function of stable state, moves back this stable state; If the penalty of stable state, migrates to respectively all intermediatenesses adjacent with this stable state;

If the reward function of intermediateness, jumps to the corresponding stable state of this intermediateness; If the penalty of intermediateness, towards with this intermediateness the rightabout of corresponding stable state migrate to next adjacent stable state or intermediateness.

4. method as claimed in claim 2 or claim 3, is characterized in that, by Gur Game iterative process, dispensed is to the method for each user's contribute information amount:

First, M=N is set, M is for finally selecting the participation value number of carrying out this subtask; Make all users' contribute information amount surpass perception task information quality demand,

use parameter

replace the parameters u in Gur Game structure ^r;

Then, by Gur Game algorithm, calculate alternative parameter u ^rafter distribute to each user's contribute information amount

5. method as claimed in claim 4, is characterized in that, described Gur Game algorithm comprises the steps:

A1) initial value that current iteration number of times k is set is 0, and it is K that total iterations is set;

A2) initialization user smart machine;

A3) calculate all user profile contribution amount initial value sums

if

or

perform step A4), otherwise jump to steps A 6); ζ ∈ [0,0.2] wherein;

A4) by following various computing information quality factor respectively

reward function

with penalty P _i ^j(k):

A5) Gur Game structure generation random number seed ∈ [0,1], by judging the size of this random number and reward function, further determines state transition direction:

If

according to reward function, point to and migrate to NextState; Otherwise, by penalty, point to and migrate to NextState, jump to steps A 7);

A6) calculate the mean value that current all user's energy consume

and further judge state transition direction; In formula, the energy consumption of each user i is

in formula,

for dump energy, E _ifor primary power, γ is the transforming factor being consumed to energy by perception information,

contribute information amount initial value for each user;

A7) k=k+1 is set, if k>K exits this iterative computation; Otherwise, enter next iteration, jump to steps A 3);

A8) select k ^*the result of step is as the contribute information amount of distributing to each user,

If exist in iterative process

?

Otherwise wherein,

6. method as claimed in claim 5, is characterized in that, judges that the method for state transition direction is in described steps A 6:

If migration left; If

migration to the right; Otherwise, keep current state constant.

7. method as claimed in claim 4, is characterized in that, the method that described network platform information centre carries out participant's selection is:

B1) the remuneration b sending to network platform information centre according to user _in user sorted, make b ₁≤ b ₂≤ ... ≤ b _n, another i=N;

B2) if

order

re-execute this step B2; Otherwise, turn to step B3;

B3) finally selecting the participant who carries out this subtask is M user,

8. the gunz aware network management system based on Gur Game, this system comprises gunz aware services device, network platform information centre and user's intelligent terminal, between described gunz aware services device and user's intelligent terminal, by network platform information centre, carry out information interaction, it is characterized in that:

Described gunz aware services device, for gathering and store user's perception information around, also for send the information quality demand of perception task and the final perception information that receives user's intelligent terminal feedback to user's intelligent terminal;

Described network platform information centre, carries out use processing for the contribute information amount initial value to receiving;

Described user's intelligent terminal, the contribute information amount of giving each user for dispensed, also for sending the required remuneration of user to network platform information centre.

9. system as claimed in claim 8, is characterized in that, described network platform information centre comprises:

Perception task dispatch module, for being distributed to each user's intelligent terminal by wireless senser by the information quality demand of perception task;

Data fusion analysis platform, for carrying out use processing to the perception information receiving;

Described user's stimulating platform, for selecting the participant who carries out this subtask.

10. system as claimed in claim 8, is characterized in that, described user's intelligent terminal comprises:

Electric weight monitoring module, for the dump energy of supervisory user intelligent terminal;

Gur Game sensed activation recommended engine, for sending contribute information amount initial value and by determining with the interaction of the network platform information centre contribute information amount of distributing to each user to network platform information centre;

The module of bidding, recommends to send amount of money award to network platform information centre for the preliminary perception action of sending according to Gur Game sensed activation recommended engine, also for completing after this subtask as user, and the amount of money award that reception network platform information is returned;

Sensed activation decision module, selects information for receiving the participant that network platform information sends, if selected, informs that user carries out this subtask, final perception information is returned to gunz aware services device by network platform information centre simultaneously.

Images