CN101841915A - Method for scheduling and distributing resources in communication system and scheduler - Google Patents

Abstract

The invention discloses a method for scheduling and distributing resources in a communication system and a scheduler. The method comprises the following steps of: calculating utility promotion of each user through a constant risk aversion utility function according to the existing gain of the user and the available gain after the resources are distributed; marking the user with maximum utility promotion, corresponding to each resource unit, as a target user of the resource unit; selecting the resource unit serving as the target user and having maximum utility promotion as a target source unit through each user; and updating the existing gain of each user according to a resource scheduling and distributing result, and calculating the utility promotion capable of being realized after the resource units are distributed to different users according to the constant risk aversion utility function. The method for scheduling and distributing the resources in the communication system and the scheduler schedule and distribute the resources by using the utility function of a constant risk aversion type, and realize maximization of the system utility.

Description

Scheduling of resource in a kind of communication system and distribution method and scheduler

Technical field

The present invention relates to the resource management techniques of communication system, be meant scheduling of resource and distribution method and scheduler in a kind of communication system especially.

Background technology

Along with to the unprecedented demand that increases of personal mobile communication and to the dependence of personal communication, be that number of users or radio communication service amount are all in swift and violent increase.The finiteness of Internet resources and the unlimitedness of communication requirement are a pair of basic contradiction in the communication system, and the existence of this contradiction makes and makes full use of the business demand that limited resources satisfy the user and become present problem demanding prompt solution.

The contention of user that have its source in that the resource allocation technology occurs to limited resources, the resource allocation technology be user that request is inserted and the user of connecting system carry out that reasonable resources is distributed and scheduling, be intended to make that the resource utilization of system is the highest, also be the maximization of utility that resource is used, and the optimization of the every performance index of system.The validity of resource allocation and fairness are the focuses that the resource allocation problem is paid close attention to always, and Limited resources exists multiple resource allocation algorithm and scheduler at present better for the user provides service in the system in order to make.Existing deployment algorithm mainly comprises following several types: fair throughput (Throughput Fair), poll (Round Robin) and max carrier to interference basic concocting methods such as (Maximum C/I), and for seeking compromise equitable proportion (Proportional Fair) concocting method that proposes of throughput of system maximization and these two opposites of user fairness.

Summary of the invention

In view of this, the object of the present invention is to provide scheduling of resource and distribution method and scheduler in a kind of communication system, realize the maximization of system utility.

Scheduling of resource and distribution method based in above-mentioned purpose a kind of communication system provided by the invention comprise:

A. according to the existing income of user, distribute the income after the new resource, and after the new resource allocation to all the other users' income influence, the system utility of can realize for different user by risk aversion utility function computational resource allocation promotes;

B. the user ID that can realize the lifting of maximum system effectiveness with each Resource Unit correspondence is the targeted customer of this Resource Unit;

C. each user selects to promote maximum Resource Unit as the target resource unit as targeted customer and effectiveness certainly;

D. according to step C scheduling of resource and distribution result, upgrade bearing interest of each user, return steps A.

Optionally, take a step forward and comprise in the described steps A of returning among this method step D: judging whether that all Resource Units all are assigned with finishes, if, process ends then, otherwise continue;

Perhaps judge whether further scheduling and distribute no longer further elevator system effectiveness, if, process ends then, otherwise continuation;

The business demand that perhaps judges whether the user is met, if, process ends then, otherwise continue.

Optionally, the described risk aversion utility function of this method is that utility function is detested in the permanent wind danger.

Optionally, the described system utility of this method step A promotes by following formula and calculates:

$u_{i,n}=\left\{\begin{array}{cc}\mathrm{\Δ}{u}_{i,n}+{\mathrm{\Σ}}_{j\∈M_n}\mathrm{\Δ}{u}_{j,n},& {\mathrm{\γ}}_{i,n}=0\\ 0,& {\mathrm{\γ}}_{i,n}\≠0\end{array}\right.$

Wherein, u _{I, n}For user i takies the system utility lifting that RB-n can realize, Δ u _{I, n}Self effectiveness that takies for user i that RB-n can get promotes; Δ u _{J, n}Be other users of originally taking Resource Block RB-n variation, wherein γ because of the effectiveness resource RB-n being distributed to user i and cause _{I, n}Be used for identifying RB-n and whether distribute to user i, value is 1 if distribute then, otherwise is 0.

Optionally, the described user i of this method self effectiveness of taking that RB-n can get promotes by following formula and calculates:

$\mathrm{\Δ}{u}_{i,n}=1-e^{-c(R_i+r_{i,n})}-(1-e^{-c{R}_i})=e^{-c{R}_i}(1-e^{-c\·r_{i,n}})$

Described other users are because of the changes delta u of the effectiveness resource RB-n being distributed to user i and cause _{J, n}Calculate by following formula:

$\mathrm{\Δ}{u}_{j,n}=1-e^{-c{R}_j^{\′}}-(1-e^{-c{R}_j})=e^{-c{R}_j}-e^{-c{R}_j^{\′}}=e^{-c{R}_j}(1-e^{-c\·(r_{j,n}^{\′}-r_{j,n})}).,$

Wherein, γ _{J, n}For can be on RB-n for user j provides income, R _jBe user's total revenue that all shared Resource Block of user j can provide, r ' _{J, n}After RB-n is distributed to user i, the income that user j can get on RB-n, R _j' be along with Resource allocation and smoothing, the user j total revenue that can realize after the new round allotment.

Optionally, bearing interest of user described in this method step D calculated by following formula

$r_{i,n}=\frac{B}{N}\log (1+{\mathrm{\ψ}}_{i,n})$

$R_i=\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\γ}}_{i,j}{r}_{i,j}$

ψ wherein _{I, n}Expression user i takies the channel Signal to Interference plus Noise Ratio of RB-n, γ _{I, j}Be used for identifying RB-j and whether distribute to user i, value is 1 if distribute then, otherwise is 0.

Optionally, the described user's income of this method is the bit number of transmission rate or transmission.

Based on above-mentioned purpose, the present invention also provides the scheduler of scheduling of resource and distribution in a kind of communication system, comprising:

User's income calculation module is used to calculate the existing income of user, distributes the income after the new resource, and after the new resource allocation to all the other users' income influence;

Effectiveness promotes computing module, be used for according to the got income after the existing income of user, the resource allocation and the resource allocation income influence to all the other users, the system utility of can realize for different user by risk aversion utility function computational resource allocation promotes;

The targeted customer selects module, and the user ID that can realize the lifting of maximum system effectiveness that is used for each Resource Unit correspondence is the targeted customer of this Resource Unit;

Scheduling of resource and distribution module are used to each user to select to promote maximum Resource Unit as the target resource unit as targeted customer and effectiveness certainly;

User's income update module is used for scheduling of resource and distribution result according to scheduling of resource and distribution module, upgrades bearing interest of each user.

Optionally, the described risk aversion utility function of this scheduler is that utility function is detested in the permanent wind danger.

Optionally, the described system utility of the described effectiveness lifting of this scheduler computing module calculating promotes by following formula and calculates:

$u_{i,n}=\left\{\begin{array}{cc}\mathrm{\Δ}{u}_{i,n}+{\mathrm{\Σ}}_{j\∈M_n}\mathrm{\Δ}{u}_{j,n},& {\mathrm{\γ}}_{i,n}=0\\ 0,& {\mathrm{\γ}}_{i,n}\≠0\end{array}\right.$

Wherein, u _{I, n}For user i takies the system utility lifting that RB-n can realize, Δ u _{I, n}Self effectiveness that takies for user i that RB-n can get promotes; Δ u _{J, n}Be other users of originally taking Resource Block RB-n variation, wherein γ because of the effectiveness resource RB-n being distributed to user i and cause _{I, n}Be used for identifying RB-n and whether distribute to user i, value is 1 if distribute then, otherwise is 0.

Optionally, the described user i of this scheduler self effectiveness of taking that RB-n can get promotes by following formula and calculates:

$\mathrm{\Δ}{u}_{i,n}=1-e^{-c(R_i+r_{i,n})}-(1-e^{-c{R}_i})=e^{-c{R}_i}(1-e^{-c\·r_{i,n}})$

Described other users are because of the changes delta u of the effectiveness resource RB-n being distributed to user i and cause _{J, n}Calculate by following formula:

$\mathrm{\Δ}{u}_{j,n}=1-e^{-c{R}_j^{\′}}-(1-e^{-c{R}_j})=e^{-c{R}_j}-e^{-c{R}_j^{\′}}=e^{-c{R}_j}(1-e^{-c\·(r_{j,n}^{\′}-r_{j,n})}).$

Wherein, r _{J, n}For can be on RB-n for user j provides income, R _jBe user's total revenue that all shared Resource Block of user j can provide, r ' _{J, n}After RB-n is distributed to user i, the income that user j can get on RB-n, R _j' be along with Resource allocation and smoothing, the user j total revenue that can realize after the new round allotment.

Optionally, bearing interest of the described user of this scheduler user income calculation module calculating calculated by following formula

$r_{i,n}=\frac{B}{N}\log (1+{\mathrm{\ψ}}_{i,n})$

$R_i=\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\γ}}_{i,j}{r}_{i,j}$

ψ wherein _{I, n}Expression user i takies the channel Signal to Interference plus Noise Ratio of RB-n, γ _{I, j}Be used for identifying RB-j and whether distribute to user i, value is 1 if distribute then, otherwise is 0.

Optionally, the described user's income of this scheduler is the bit number of transmission rate or transmission.

Optionally, this scheduler of this scheduler is arranged in radio network controller (RNC) or base station controller (BSC) or base station or eNB (eNodeB) or the access point (AP).

From above as can be seen, scheduling of resource in the communication system provided by the invention and distribution method and scheduler utilize the utility function of risk aversion type resource is dispatched and to distribute, and have realized the maximization of system utility.

Description of drawings

Fig. 1 is used for the structural representation of the scheduler of scheduling of resource and distribution for the embodiment of the invention;

Fig. 2 is the schematic flow sheet of embodiment of the invention scheduling of resource and distribution method.

Embodiment

For making the purpose, technical solutions and advantages of the present invention clearer, below in conjunction with specific embodiment, and with reference to accompanying drawing, the present invention is described in more detail.

Scheduling of resource in the communication system provided by the invention and distribution method mainly comprise:

A. according to the existing income of user, distribute the income after the new resource, and after the new resource allocation to all the other users' income influence, the system utility of can realize for different user by risk aversion utility function computational resource allocation promotes;

B. the user ID that can realize the lifting of maximum system effectiveness with each Resource Unit correspondence is the targeted customer of this Resource Unit;

C. each user selects to promote maximum Resource Unit as the target resource unit as targeted customer and effectiveness certainly;

D. according to step C scheduling of resource and distribution result, upgrade bearing interest of each user, return steps A.

Wherein, take a step forward and comprise in the described steps A of returning among the step D: judging whether that all Resource Units all are assigned with finishes, if, process ends then, otherwise continue; Perhaps also can judge whether further scheduling and distribute no longer further elevator system effectiveness, if, process ends then, otherwise continuation; The business demand that perhaps judges whether the user is met, if, process ends then, otherwise continue.

System resource allotment entity of the present invention can be RNC, BSC, base station, access point (AP) and eNB etc.

Suppose that total N overlaps assignable Resource Unit, M user in the system.Wherein to arbitrary Resource Unit n, it is to all user i=1, and 2 ... M co-exists in M value of utility that calculates according to described utility function.

According to the proposed method, described scheduling of resource and assigning process are referring to shown in Figure 2.

Step 210, Resource Unit select target user.

Because the channel fading of user actual geographical position of living in difference, translational speed difference, experience is not equal, will cause each user's channel quality difference, therefore some Resource Units to the lifting of its effectiveness also with difference.

Calculate the system utility that various Resource Allocation Formulas can bring according to the risk aversion utility function in the present embodiment, each Resource Unit correspondence had user ID that maximum utility a promotes targeted customer for this Resource Unit according to the principle of maximize system resource overall utility.

Step 220, user's select target Resource Unit.

Each user selects to be the targeted customer and can to realize that it is the target resource unit that its effectiveness promotes maximum Resource Unit with it.When a user only is designated the targeted customer of described Resource Unit by a cover Resource Unit, give this user with described resource unit allocation; In scheduling of resource and assigning process overlapped Resource Unit when sign when a user takes turns one more, will make described user's effectiveness promote maximum resource unit allocation in these Resource Units and give described user.Wherein each user may obtain one or more sets Resource Units.

Step 230, the value of utility of Resource Unit correspondence upgrades.

After taking turns the resource allocation process through one, upgrade each user's current income according to the risk aversion utility function; And calculate when giving different user with each resource unit allocation according to described risk aversion utility function, calculate attainable system utility lifting values; Carry out the value of utility renewal of Resource Unit correspondence then according to described effectiveness lifting values, carry out the resource allocation process of next round more again.Wherein each user may obtain one or more sets Resource Units.

Step 240 is allocated the residual resource unit in an identical manner between the user, finish up to scheduling of resource and assigning process.Wherein, the decision condition of the scheduling and the end of assigning process can be one or several the combination in the following condition: all Resource Units all are assigned with and finish; Scheduling and distribute no longer further elevator system effectiveness (for example: current system utility value that calculates and original system value of utility are compared, if smaller or equal to the original system value of utility, then explanation can not be more further elevator system effectiveness, then process ends); User's business demand is met (for example: judge whether the current system utility value that calculates reaches the preset user requirements, if process ends is then calculated otherwise continue next round).

The utility function of correspondence was different when the different scheduling of resource were presented as mainly that with the difference of allocation criteria the user uses the different resource unit.The applicant finds to be more prone to the risk aversion type for communication system through research and experiment.The preferred permanent wind danger detest type utility function that adopts is used as utility function among the present invention, for example: u (x)=a one be ^-cx, a wherein, b, c are constant, c＞0; X represents the transmission rate that the user uses described Resource Unit to reach; The significant degree that the big or small representative system resource allocation strategy of u (x) is met consumers' demand, specifically can representative of consumer to the satisfaction of resulting transmission rate.

Embodiment one

Down link with ofdm system is an example, establishes to comprise L base station in the ofdm system, is respectively 1 ... l..., L represents; System comprises N Resource Block, is represented by RB, is respectively RB-1 ... RB-n..., RB-N; System bandwidth is B, and the number of users in each base station is M _l, so total in the system $M={\mathrm{\Σ}}_{l=1}^L{M}_l$ Individual user is respectively 1 ... m..., M represents; M _nExpression takies the user's of RB-n set, M _lThe user's of expression base station l service set, l (m) is expressed as user m serving base station, G _{M, l}Expression user m is to the channel gain between the l of base station, P _{L, n}Expression base station l distributes to the power of RB-n.γ _{M, n}Be used for identifying RB-n and whether distribute to user m, value is 1 if distribute then, otherwise is 0.If Resource Block n can be r for user i provides income _{I, n}, user's total revenue that all shared Resource Block of user can provide is respectively R ₁, R ₂..., R _M, along with Resource allocation and smoothing, user's total revenue that can realize after the new round allotment is respectively R ' ₁, R ' ₂..., R ' _MThe base station has N sub carrier group, RB-n wherein, and for all user i=1,2 ... M has M different system utility lifting values u _{I, n}Therefore, user i takies the Signal to Interference plus Noise Ratio of n Resource Block RB-n and is

${\mathrm{\ψ}}_{i,n}=\frac{P_{l\left(i\right),n}\·G_{i,l\left(i\right)}}{{\mathrm{\σ}}^2+{\mathrm{\Σ}}_{l\≠l\left(i\right)}{\mathrm{\Σ}}_{m\∈M_l}{\mathrm{\γ}}_{m,n}\·P_{l,n}\·G_{i,l}},---\left(1\right)$

If n Resource Block RB-n distributed to user i, this user income of obtaining of Resource Block thus then, i.e. the lifting of transmission rate is

$r_{i,n}=\frac{B}{N}\log (1+{\mathrm{\ψ}}_{i,n}).---\left(2\right)$

J Resource Block RB-j for distributing to user i satisfies γ _{I, j}=1, therefore, the total revenue that user i is current, promptly current transmission rate is

$R_i=\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\γ}}_{i,j}{r}_{i,j},---\left(3\right)$

Utilizing permanent wind danger to detest utility function in the embodiment of the invention calculates current effectiveness and is

$u_i=1-e^{-c{R}_i}---\left(4\right)$

If RB-n is distributed to user i, then γ _{I, n}The effectiveness that=1 user i can get at RB-n promotes and is

$\mathrm{\Δ}{u}_{i,n}=1-e^{-c(R_i+r_{i,n})}-(1-e^{-c{R}_i})=e^{-c{R}_i}(1-e^{-c\·r_{i,n}})---\left(5\right)$

If RB-n is distributed to user i, then for other use the user of RB-n, i.e. M in the system _nIn other users, will bring new interference, and then need upgrade its Signal to Interference plus Noise Ratio.For one of them other user j,

${\mathrm{\ψ}}_{j,n}^{\′}=\frac{P_{l\left(j\right),n}\·G_{j,l\left(i\right)}}{{\mathrm{\σ}}^2+{\mathrm{\Σ}}_{l\≠l\left(j\right)}{\mathrm{\Σ}}_{m\∈M_l}{\mathrm{\γ}}_{m,n}\·P_{l,n}\·G_{j,l}+P_{l\left(i\right)}\·G_{j,l\left(i\right)}}---\left(6\right)$

Accordingly, the income on RB-n of user j becomes:

$r_{j,n}^{\′}=\frac{B}{N}\log (1+{\mathrm{\ψ}}_{i,n}^{\′}),---\left(7\right)$

The income of user j becomes R ' _j=R _j-r _{J, n}+ r ' _{J, n}, (8)

The effectiveness of user j changes as follows:

$\mathrm{\Δ}{u}_{j,n}=1-e^{-c{R}_j^{\′}}-(1-e^{-c{R}_j})=e^{-c{R}_j}-e^{-c{R}_j^{\′}}=e^{-c{R}_j}(1-e^{-c\·(r_{j,n}^{\′}-r_{j,n})}).---\left(9\right)$

Wherein, r _{J, n}For can be on RB-n for user j provides income, R _jBe user's total revenue that all shared Resource Block of user j can provide, r ' _{J, n}After RB-n is distributed to user i, the income that user j can get on RB-n, R _j' be along with Resource allocation and smoothing, the user j total revenue that can realize after the new round allotment.

Therefore RB-n being distributed to user i system can get total utility and promote and be

$u_{i,n}=\left\{\begin{array}{cc}\mathrm{\Δ}{u}_{i,n}+{\mathrm{\Σ}}_{j\∈M_n}\mathrm{\Δ}{u}_{j,n},& {\mathrm{\γ}}_{i,n}=0\\ 0,& {\mathrm{\γ}}_{i,n}\≠0\end{array}\right.,---\left(10\right)$

γ wherein _{I, n}Be used for identifying RB-n and whether distribute to user i, value is 1 if distribute then, otherwise is 0.

At single service application, the calculating of the value of utility of user's correspondence uses identical utility function to calculate; At multiple services application, for each business designs utility function according to its QoS requirement, as adjusting the final realization of the c value overall utility value maximization in the permanent wind danger detest type utility function.For example: in streaming media service, can value be c=1.6 * 10 ^-6In ftp business, can value be c=1.6 * 10 ^-7

According to the proposed method, the described ofdm system flow process of carrying out resource allocation is as follows:

1) Resource Unit select target user:

Resource Unit calculates according to formula (10) it is distributed to different user available systems effectiveness changing value u _{I, n}

Detest utility function according to the permanent wind danger, be that formula (4) calculates the system utility that various Resource Allocation Formulas can bring, each Resource Unit correspondence had user ID that maximum utility a promotes targeted customer for this Resource Unit according to the principle of maximize system resource overall utility.

That supposes the 1st Resource Block RB-1 correspondence can realize that the user that maximum system effectiveness promotes is user-10, and then the targeted customer of RB-1 is user-10; The user that can realize the lifting of maximum system effectiveness of RB-2 correspondence is user-20, and then the targeted customer of RB-2 is user-20; The user that can realize the lifting of maximum system effectiveness of RB-3 correspondence is user-10, and then the targeted customer of RB-3 is user-10.

2) user's select target Resource Unit:

Each user selects to be the targeted customer and can to realize that it is the target resource unit that its effectiveness promotes maximum Resource Unit, promptly selects u with it _{I, n}When a user only is designated the targeted customer of described Resource Unit by a cover Resource Unit, give this user with described resource unit allocation; In resource allocation process overlapped Resource Unit when sign when a user takes turns one more, will make described effectiveness promote maximum resource unit allocation in these Resource Units and give described user.

Suppose that user-20 is designated the targeted customer by RB-2, then user-20 is designated RB-2 the target resource unit and takies this Resource Unit; Suppose user-10 simultaneously by RB-1, RB-3 is designated the targeted customer, but RB-1 can realize that than RB-3 higher effectiveness promotes, and then user-10 is designated RB-1 the target resource unit and takies this Resource Unit.

3) value of utility of Resource Unit correspondence upgrades:

After taking turns the resource allocation process through one, the current income that utility function is upgraded each user is detested according to the permanent wind danger by system, can get income R after the user distributes by formula (7,8) ₁, R ₂..., R _MFurther, detest utility function according to described permanent wind danger and consider at the same time on the basis of the effectiveness influence that resource allocation brings to different user, calculate attainable system utility lifting values { u when giving different user each resource unit allocation according to formula (10) _{I, n}| i ∈ [1, M], n ∈ [1, N] }, system carries out the value of utility renewal of Resource Unit correspondence according to described effectiveness lifting values then, revises the initial value of next round allotment.Carry out the resource allocation process of next round again, return 1).

Suppose user-10 simultaneously by RB-1, RB-3 is designated the targeted customer, but RB-1 can realize that than RB-3 higher effectiveness promotes, and then RB-3 is released in the resource collection to be allocated, and the value of utility that system carries out the RB-3 correspondence according to described effectiveness lifting values upgrades.

By that analogy, the residual resource unit is allocated between the user in an identical manner, till scheduling and distributing no longer further elevator system effectiveness, i.e. { u _{I, n}| i ∈ [1, M], n ∈ [1, N] } in all elements non-just till.

Embodiment two

Consider the time-division scheduling mode of WCDMA system descending shared channel, establish and comprise L base station in the WCDMA system, be respectively 1 ... l..., L represent that the number of users in each base station is M _lEach time slot (10ms), a user is given in the resource allocation of each base station, and the different different users of time slot base station scheduling serve.System adopts the AMC mode, selects suitable modulation coding mode according to the CSI value of user feedback for the user, and then has determined the bit value that user's current time can transmit, and promptly the income of user's current time is made as

So for current time slots n, all user i=1,2 ..., M _lM is arranged _lIndividual different system utility lifting values u _{I, n}

The current effectiveness of user i is

$u_i=1-e^{-c{R}_i}---\left(4\right)$

If current time slots n distributes to user i with current time slots, this user Ke De income then, i.e. the lifting of transmission rate is r _{I, n}, its value is calculated according to concrete AMC mode and is got.Therefore user i in the effectiveness lifting that time slot n can get is

$\mathrm{\Δ}{u}_{i,n}=1-e^{-c(R_i+r_{i,n})}-(1-e^{-c{R}_i})=e^{-c{R}_i}(1-e^{-c\·r_{i,n}})---\left(5\right)$

Different users is distributed to current time slots in each base station, the resource that the CU system is whole, therefore think with base station l that it is approximate identical that the current time slots of l ∈ L is distributed to the interference that user that different users connects other base stations causes, and ignores this part when calculating effectiveness.Therefore current time slots n distributes to the effectiveness lifting u of user i _{I, n}=Δ u _{I, n}

According to the proposed method, to carry out the flow process of resource allocation as follows in described WCDMA system:

It is the targeted customer that the base station will promote maximum user ID to current time slots effectiveness.

Because the channel fading of user actual geographical position of living in difference, translational speed difference, experience is not equal, the channel quality difference that will cause each user, suitable modulation coding mode is selected according to the CSI value of user feedback in the base station, determined the bit value that user's current time slots can be transmitted thus, i.e. current time slots income r _{J, n}

Utility function is detested according to the permanent wind danger in the base station, based on the existing income R of user _iAnd the income r of current time slots _{I, n}Calculate the effectiveness lifting u of different user to current time slots _{I, n}, it is the targeted customer that effectiveness is promoted maximum user ID.

Selecting effectiveness to promote maximum user for current time slots serves.

If current time slots has a plurality of users to be identified as the targeted customer, promptly their effectiveness lifting values is maximum and identical, and it is identical to the effectiveness lifting of system to serve different users, then selects a user to serve at random.If have only a user to be identified as the targeted customer, select this user to serve.

3) effectiveness is upgraded

Behind a time slot, the current income that utility function is upgraded each user is detested according to the permanent wind danger by system

Return 1) carry out the assigning process of next time slot.

Based on said method, the present invention also provides the scheduler of scheduling of resource and distribution in a kind of communication system, referring to shown in Figure 1, comprising:

User's income calculation module 101 is used to calculate the existing income of user, distributes the income after the new resource, and after the new resource allocation to all the other users' income influence;

Effectiveness promotes computing module 102, is used for according to the got income after the existing income of user, the resource allocation, detests the system utility lifting that the utility function computational resource allocation can realize for different user by the permanent wind danger;

The targeted customer selects module 103, and the user ID that can realize the lifting of maximum system effectiveness that is used for each Resource Unit correspondence is the targeted customer of this Resource Unit;

Scheduling of resource and distribution module 104 are used to each user to select to promote maximum Resource Unit as the target resource unit as targeted customer and effectiveness certainly;

User's income update module 105 is used for scheduling of resource and distribution result according to scheduling of resource and distribution module, upgrades bearing interest of each user.

This scheduler can be arranged in RNC or BSC or base station or eNB or the access point (AP).

Above-described specific embodiment is specific embodiments of the invention only, is not limited to the present invention, and is within the spirit and principles in the present invention all, any modification of being made, is equal to replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (14)

1. scheduling of resource and the distribution method in the communication system is characterized in that,

A. according to the existing income of user, distribute the income after the new resource, and after the new resource allocation to all the other users' income influence, the system utility of can realize for different user by risk aversion utility function computational resource allocation promotes;

B. the user ID that can realize the lifting of maximum system effectiveness with each Resource Unit correspondence is the targeted customer of this Resource Unit;

C. each user selects to promote maximum Resource Unit as the target resource unit as targeted customer and effectiveness certainly;

D. according to step C scheduling of resource and distribution result, upgrade bearing interest of each user, return steps A.

2. method according to claim 1 is characterized in that, takes a step forward and comprises in the described steps A of returning among the step D: judging whether that all Resource Units all are assigned with finishes, if, process ends then, otherwise continue;

Perhaps judge whether further scheduling and distribute no longer further elevator system effectiveness, if, process ends then, otherwise continuation;

The business demand that perhaps judges whether the user is met, if, process ends then, otherwise continue.

3. method according to claim 1 is characterized in that, described risk aversion utility function is that utility function is detested in the permanent wind danger.

4. method according to claim 3 is characterized in that, the described system utility of steps A promotes by following formula to be calculated:

$u_{i,n}=\left\{\begin{array}{cc}\mathrm{\Δ}{u}_{i,n}+{\mathrm{\Σ}}_{j\∈M_n}\mathrm{\Δ}{u}_{j,n},& {\mathrm{\γ}}_{i,n}=0\\ 0,& {\mathrm{\γ}}_{i,n}\≠0\end{array}\right.$

Wherein, u _{I, n}For user i takies the system utility lifting that RB-n can realize, Δ u _{I, n}Self effectiveness that takies for user i that RB-n can get promotes; Δ u _{J, n}Be other users of originally taking Resource Block RB-n variation, wherein γ because of the effectiveness resource RB-n being distributed to user i and cause _{I, n}Be used for identifying RB-n and whether distribute to user i, value is 1 if distribute then, otherwise is 0.

5. method according to claim 4 is characterized in that, self effectiveness that described user i takies that RB-n can get promotes by following formula to be calculated:

$\mathrm{\Δ}{u}_{i,n}=1-e^{-c(R_i+r_{i,n})}-(1-e^{-c{R}_i})=e^{-c{R}_i}(1-e^{-c\·r_{i,n}})$

Described other users are because of the changes delta u of the effectiveness resource RB-n being distributed to user i and cause _{J, n}Calculate by following formula:

$\mathrm{\Δ}{u}_{j,n}=1-e^{-c{R}_j^{\′}}-(1-e^{-c{R}_j})=e^{-c{R}_j}-e^{-c{R}_j^{\′}}=e^{-c{R}_j}(1-e^{-c\·(r_{j,n}^{\′}-r_{j,n})}).,$

Wherein, r _{J, n}For can be on RB-n for user j provides income, R _jBe user's total revenue that all shared Resource Block of user j can provide, r ' _{J, n}After RB-n is distributed to user i, the income that user j can get on RB-n, R _j' be along with Resource allocation and smoothing, the user j total revenue that can realize after the new round allotment.

6. method according to claim 1 is characterized in that, bearing interest of user described in the step D calculated by following formula

$r_{i,n}=\frac{B}{N}\log (1+{\mathrm{\ψ}}_{i,n})$

$R_i=\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\γ}}_{i,j}{r}_{i,j}$

ψ wherein _{I, n}Expression user i takies the channel Signal to Interference plus Noise Ratio of RB-n, γ _{I, j}Be used for identifying RB-j and whether distribute to user i, value is 1 if distribute then, otherwise is 0.

7. method according to claim 1 is characterized in that, described user's income is the bit number of transmission rate or transmission.

8. the scheduler of scheduling of resource and distribution in the communication system is characterized in that, comprising:

User's income calculation module is used to calculate the existing income of user, distributes the income after the new resource, and after the new resource allocation to all the other users' income influence;

Effectiveness promotes computing module, be used for according to the got income after the existing income of user, the resource allocation and the resource allocation income influence to all the other users, the system utility of can realize for different user by risk aversion utility function computational resource allocation promotes;

The targeted customer selects module, and the user ID that can realize the lifting of maximum system effectiveness that is used for each Resource Unit correspondence is the targeted customer of this Resource Unit;

Scheduling of resource and distribution module are used to each user to select to promote maximum Resource Unit as the target resource unit as targeted customer and effectiveness certainly;

User's income update module is used for scheduling of resource and distribution result according to scheduling of resource and distribution module, upgrades bearing interest of each user.

9. scheduler according to claim 8 is characterized in that, described risk aversion utility function is that utility function is detested in the permanent wind danger.

10. scheduler according to claim 9 is characterized in that, the described system utility that described effectiveness promotes computing module calculating promotes by following formula calculating:

$u_{i,n}=\left\{\begin{array}{cc}\mathrm{\Δ}{u}_{i,n}+{\mathrm{\Σ}}_{j\∈M_n}\mathrm{\Δ}{u}_{j,n},& {\mathrm{\γ}}_{i,n}=0\\ 0,& {\mathrm{\γ}}_{i,n}\≠0\end{array}\right.$

Wherein, u _{I, n}For user i takies the system utility lifting that RB-n can realize, Δ u _{I, n}Self effectiveness that takies for user i that RB-n can get promotes; Δ u _{J, n}Be other users of originally taking Resource Block RB-n variation, wherein γ because of the effectiveness resource RB-n being distributed to user i and cause _{I, n}Be used for identifying RB-n and whether distribute to user i, value is 1 if distribute then, otherwise is 0.

11. scheduler according to claim 10 is characterized in that, self effectiveness that described user i takies that RB-n can get promotes by following formula to be calculated:

$\mathrm{\Δ}{u}_{i,n}=1-e^{-c(R_i+r_{i,n})}-(1-e^{-c{R}_i})=e^{-c{R}_i}(1-e^{-c\·r_{i,n}})$

Described other users are because of the changes delta u of the effectiveness resource RB-n being distributed to user i and cause _{J, n}Calculate by following formula:

$\mathrm{\Δ}{u}_{j,n}=1-e^{-c{R}_j^{\′}}-(1-e^{-c{R}_j})=e^{-c{R}_j}-e^{-c{R}_j^{\′}}=e^{-c{R}_j}(1-e^{-c\·(r_{j,n}^{\′}-r_{j,n})}).$

Wherein, r _{J, n}For can be on RB-n for user j provides income, R _jBe user's total revenue that all shared Resource Block of user j can provide, r ' _{J, n}After RB-n is distributed to user i, the income that user j can get on RB-n, R _j' be along with Resource allocation and smoothing, the user j total revenue that can realize after the new round allotment.

12. scheduler according to claim 8 is characterized in that, bearing interest of the described user that user's income calculation module is calculated calculated by following formula

$r_{i,n}=\frac{B}{N}\log (1+{\mathrm{\ψ}}_{i,n})$

$R_i=\underset{j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\γ}}_{i,j}{r}_{i,j}$

ψ wherein _{I, n}Expression user i takies the channel Signal to Interference plus Noise Ratio of RB-n, γ _{I, j}Be used for identifying RB-j and whether distribute to user i, value is 1 if distribute then, otherwise is 0.

13. scheduler according to claim 8 is characterized in that, described user's income is the bit number of transmission rate or transmission.

14. scheduler according to claim 8 is characterized in that, this scheduler is arranged among RNC or BSC or base station or eNB or the access point AP.

Images