WO2011097908A1

WO2011097908A1 - Scheduling method, device, base station and system for collaboration resources

Info

Publication number: WO2011097908A1
Application number: PCT/CN2010/078899
Authority: WO
Inventors: 张弓; 杨讯
Original assignee: 华为技术有限公司
Priority date: 2010-02-11
Filing date: 2010-11-19
Publication date: 2011-08-18
Also published as: CN102158974A; CN102158974B

Abstract

A scheduling method for collaboration resources, comprising: collecting the corresponding information about time average rates during the first time period of all the users under the jurisdiction; obtaining the gradient of the utility function for the time average rates during the first time period of all the users according to the corresponding information about the time average rates during the first time period of each user; updating division scheme of each collaboration cluster on the division probabilities according to the gradient of the utility function for the time average rates during the first time period of all the users; updating all the base stations on the division of the collaboration cluster according to the updated division probabilities; informing all the base stations of the result for updating the division of the collaboration cluster, so that each base station in the collaboration cluster selects corresponding users within the cluster during the second time period and allocates resources to the users within the cluster. Correspondingly, a scheduling method, device, base station and system for collaboration resources are also disclosed by the present embodiments, achieving better network performance.

Description

Cooperative resource scheduling method, device, base station and system The present application is filed on February 11, 2010, the Chinese Patent Office, the application number is 201010113635.4, and the Chinese patent application entitled "Collaborative Resource Scheduling Method, Apparatus, Base Station and System" Priority is hereby incorporated by reference in its entirety. Technical field

The present invention relates to the field of communications, and in particular, to a cooperative resource scheduling method, apparatus, base station, and system.

Background technique

Cooperative MIMO (Co-MIMO) technology has been recognized as a key physical layer technology for the IMT-Advanced standard. In a globally frequency-multiplexed cellular system, inter-cell interference has become a major factor limiting mobile communication performance, and the basic idea of Co-MIMO is to coordinate the signal transmission of multiple base stations to mitigate inter-cell interference. As shown in FIG. 1, in a Co-MIMO system, a plurality of base stations (BSs) will simultaneously provide communication services for a plurality of mobile terminals (MSs).

In the base station, the wireless environment can be continuously adjusted by SDMA (Space Division Multiple Access) to provide a good downlink signal for each user. In the network, this advanced base station performance can be used to increase base station coverage, reduce network cost, increase system capacity, and ultimately improve frequency utilization. SDMA can be compatible with any spatial modulation method or frequency band, so it has great practical value. The SDMA after the base station cooperation can effectively overcome the interference problem at the edge of the cell, and the interference becomes a useful signal, which can further improve the spectrum efficiency. Base station cooperation is generally defined as the sharing of data between base stations and joint operations, and channel information can be shared, partially shared, or not shared. Base station cooperation brings more overhead to the network while bringing huge gains. This also brings some new problems to the existing cellular network: The cooperation of the entire network can fully utilize interference, but its complexity varies with the number of users. The exponential growth, and the signaling overhead of network and user resource scheduling is too high; and the collaborative network always has edge effects, and scalability is always a problem.

According to the cluster (cluster), the base stations in the cluster cooperate to effectively solve the above problems. A cluster is defined as a group of cooperating base stations participating in shared data and joint computing. The size of the cluster in the network depends on the network. Backhaul (backhaul) capacity and cluster computing power. The base station clustering scheme in the current network, such as fixed cluster partitioning and dynamic cluster partitioning, requires a large signaling overhead, so that the overall performance loss of the network is larger than the global cooperation, and the overall performance of the network. not tall. Summary of the invention

Embodiments of the present invention provide a cooperative resource scheduling method, apparatus, base station, and system to improve overall network performance.

An embodiment of the present invention provides a cooperative resource scheduling method, including:

Collecting information about the average time rate of all users in the first time period; obtaining the average time rate of all users in the first time period according to the information about the average time rate of each user in the first time period The gradient of the utility function;

Updating the partitioning probability of each of the cooperative clustering schemes according to a gradient of the utility function of the time average rate of all users in the first time period;

Performing cooperative clustering update on all base stations according to the updated partitioning probability;

All base stations are notified of the result of the cooperative cluster partition update, so that the base stations in each cooperative cluster select the corresponding intra-cluster users in the second time period, and perform resource allocation for the users in the cluster.

And during the second time period, sharing data with other base stations in the cooperative cluster according to the cooperative cluster partitioning result, where the cooperative cluster scheduling result is related by the cooperative resource scheduling apparatus according to the time average rate of all users in the first time period. Made by information;

In order to maximize the utility function of the collaborative cluster, select the intra-users in conjunction with other base stations in the cooperative cluster;

And determining, by the other base stations in the cooperative cluster, an uplink-downlink communication mode of the user in the cluster; and performing, in downlink communication, with other base stations in the cooperative cluster to allocate power to the users in the cluster. An embodiment of the present invention provides a cooperative resource scheduling apparatus, including:

a collecting module, configured to collect information about a time average rate of all users under the jurisdiction of the device during a first time period; a gradient obtaining module, configured to obtain a gradient of a utility function of a time average rate of all users in a first time period according to information about a time average rate of each user in a first time period; a probability update module, configured to pass the The gradient obtained by the gradient acquisition module updates the division probability of each cooperative cluster division scheme;

a partitioning update module, configured to perform cooperative clustering update on all base stations according to the updated partitioning probability;

The informing module is configured to notify all base stations of the result of the cooperative cluster partition update, so that the base stations in each cooperative cluster select the corresponding intra-cluster users in the second time period, and perform resource allocation on the users in the cluster.

An embodiment of the present invention provides a base station, including:

a shared collaboration module, configured to share data with other base stations in the cooperative cluster to which the base station belongs according to the cooperative cluster partitioning result in the second time period, where the cooperative cluster partitioning result is performed by the cooperative resource scheduling apparatus according to all users The information about the time average rate in the first time period is made; the user scheduling module is configured to maximize the utility function of the cooperation cluster, and jointly select the users in the cluster with other base stations in the cooperation cluster;

a communication mode determining module, configured to determine, by the other base stations in the cooperative cluster, an uplink and downlink communication mode of the user in the cluster;

And a power allocation module, configured to jointly allocate power to users in the cluster in downlink communication and other base stations in the cooperative cluster.

An embodiment of the present invention provides a cooperative resource scheduling system, including the foregoing cooperative resource scheduling apparatus and the plurality of base stations.

According to the foregoing technical solution, the base station is dynamically clustered according to the gradient of the time average rate utility function in the network in the first time period, and the intra-cluster user selection and resource allocation are performed in the cluster in the second time period; The clustering scheme and resource allocation are performed separately, and the computing tasks are separated, which greatly reduces the difficulty of implementing the entire system, reduces system signaling overhead, and achieves superior network performance. DRAWINGS

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following description will be made on the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are incorporated in the drawings These figures take additional drawings.

FIG. 1 is a structural diagram of a cooperative MIMO system according to an embodiment of the present invention;

2 is a flowchart of a method for scheduling a cooperative resource according to an embodiment of the present invention;

FIG. 3 is a flowchart of a method for scheduling a cooperative resource according to an embodiment of the present invention;

4 is a structural diagram of a cooperative resource scheduling apparatus according to an embodiment of the present invention;

FIG. 5 is a structural diagram of a probability update module according to an embodiment of the present invention;

FIG. 6 is a structural diagram of a partitioning update module according to an embodiment of the present invention;

FIG. 7 is a structural diagram of a base station according to an embodiment of the present invention;

FIG. 8 is a structural diagram of a user scheduling module according to an embodiment of the present invention;

FIG. 9 is a structural diagram of a communication mode determining module according to an embodiment of the present invention;

FIG. 10 is a structural diagram of a collaborative resource scheduling system according to an embodiment of the present invention. detailed description

BRIEF DESCRIPTION OF THE DRAWINGS The technical solutions in the embodiments of the present invention will be described in detail below with reference to the accompanying drawings. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without creative work are within the scope of the present invention.

As shown in FIG. 2, an embodiment of the present invention provides a cooperative resource scheduling method, where the method includes:

S110. The cooperative resource scheduling apparatus collects information about the average time rate of all users under its jurisdiction during the first time period;

The cooperative resource scheduling apparatus mentioned in this embodiment may be a central controller, a gateway, or other network element having similar functions. In this embodiment, the central controller is taken as an example for description.

In one embodiment, the user's time average rate is the average of the user's data rate over time.

In an embodiment, the related information of the time average rate of the user may include a cumulative average rate of the user under all cooperative cluster division schemes, a cumulative average probability of the user under the current cluster division scheme, and a user's accumulation under the current cluster division. Average rate. In an embodiment, the central controller may obtain the correlation of the time average rate of all users under its jurisdiction in the first time period by receiving information about the time average rate of the user in the first time period fed back by all the base stations. information;

In another embodiment, the central controller may also obtain time information of all users under its jurisdiction during the first time period by receiving information about the time average rate of the user during the first time period fed back by all base station controllers. Information about the average rate.

S120. The cooperative resource scheduling apparatus obtains a gradient of a utility function of time average rates of all users in a first time period according to information about time average rates of respective users in a first time period;

Still described by the central controller:

In one embodiment, the utility function of the user's time average rate should be a measurable monotonically increasing function. In one embodiment, the sum of the utility functions of the time average rates of all users may be the average throughput of the network.

In an embodiment, the central controller may calculate the gradient of the utility function of each user in parallel according to the time average rate information of each user in the first time period, and sum the calculated gradients of the utility functions of the respective users, A gradient of the utility function of the time average rate of all users over the first time period is obtained. In one embodiment, the gradient of the utility function of each user is calculated in parallel, and the utility function of each user may be calculated in parallel to calculate the gradient of each utility function.

In one embodiment, the central controller may calculate a utility function of the time average rate of all users in the first time period according to the time averaged speed information of each user in the first time period, and then calculate the utility function of all users. The gradient gives the gradient of the utility function of all users' time averaged speeds in the first time period.

In one embodiment, the central controller sums the gradients of the utility functions of the time average rates of the various users over a first time period to obtain a gradient of the utility function of the time average rate of all users over the first time period.

S130. The cooperative resource scheduling apparatus updates the division probability of each cooperative cluster division scheme by using the gradient in S120.

Still illustrated with a central controller, in one embodiment step S 130 can include: 1) updating the gradient of the current cluster partitioning scheme according to the gradient of the utility function of the time average rate of all users in the first time period;

2) updating the gradient set of all the cooperative cluster partitioning schemes by using the gradient of the updated current collaborative cluster partitioning scheme;

3) Using the formula: π + ^D) to update the partitioning probability of each cooperative cluster partitioning scheme,

D is the gradient set of all the collaborative clustering schemes after updating, ^ is the gradient weighting value, ^ ^{> 0} , is the w-dimensional probability vector, N is the number of all clustering schemes, and the coordinates of the probability vector π correspond to the clustering scheme The division probability, ∑ _{χ π Ν} represents the sum of the coordinates of the probability vector is 1.

That is to say, 71 is added to the weighted gradient D, and the obtained sum is projected on the space of π to obtain the updated probability vector π, thereby updating the probability of occurrence of the cooperative cluster division in the restricted interval.

Here, it should be noted that the restricted interval is such that π satisfies the condition that the sum of its coordinates is 1, that is, (11).

In one embodiment, the central controller may weight the calculated gradients, update the gradient sets of all cooperative cluster partitioning schemes using the weighted gradients, and update the partitioning probability of the cooperative cluster partitioning scheme.

S140, performing cooperative cluster division update on the base station according to the foregoing division probability;

In an embodiment, the central controller may randomly select a cooperative cluster partitioning scheme corresponding to the partitioning probability according to the partitioning probability of the cooperative clustering scheme, and perform uplink clustering update on the base station; in one embodiment, the central control The device may select a cooperative clustering scheme corresponding to the largest partitioning probability according to the partitioning probability of the cooperative clustering scheme, and perform uplink clustering update on the base station; in one embodiment, the central controller may also divide the two cooperative clusters. The updated schemes are combined to perform cooperative clustering update on the base station.

S150. The cooperative resource scheduling apparatus notifies all base stations of the result of the cooperative cluster partition update, so that the base station in each cooperative cluster selects a corresponding intra-cluster user (ie, performs user scheduling) in the second time period, and performs intra-cluster The user performs resource allocation.

In one embodiment, since the real-time requirements of the users in the corresponding clusters are higher, and the real-time requirements for clustering of the base stations are relatively low, the first time period can be much larger than the second time period. In one embodiment, the first time period may be 30 times the second time period; in one embodiment, the first time period may be 50 times, or more than 100 times, the second time period.

According to the foregoing technical solution, the base station is dynamically clustered according to the gradient of the time average rate utility function in the network in the first time period, and is used in the cluster in the second time period. User scheduling (ie, selecting users within the cluster) and resource allocation; this clustering scheme and resource allocation are performed separately, and the computing tasks are separated, which greatly reduces the difficulty of implementing the entire system, reduces system signaling overhead, and achieves superior performance. Network performance.

As shown in FIG. 3, an embodiment of the present invention provides a method for scheduling a cooperative resource, where the method includes: S220, in a second time period, a base station shares data with other base stations in a collaboration cluster according to a cooperative cluster division result; The collaborative clustering result is that the central controller makes information based on the time average rate of all users in the first time period;

In one embodiment, sharing data with other base stations within the associated cooperating cluster may include sharing channel information; in one embodiment, the channel information may also be shared.

S230. The base station maximizes the utility function of the coordinating cluster to which it belongs, and selects the users in the cluster jointly with other base stations in the coordinating cluster to which it belongs;

In one embodiment, the base station and other base stations in its cooperating cluster can compare the channel capacity of all user combinations in the cooperative cluster, and select the user combination with the largest capacity as the intra-cluster user.

For example, in one embodiment, assuming that there are a total of five users, and the current collaborative cluster can only serve three users, then the collaborative cluster needs to select three users from among the five users as intra-cluster users. Selecting 3 users from 5 users and having a total of C (20) options, the current collaboration cluster needs to compare the channel capacity of all 20 options (ie, all 20 user combination schemes). The largest number of user combinations are used as users within the cluster.

In one embodiment, the base station and other base stations in the coordinating cluster to which it belongs may also compare the channel matrices of all users in the cluster, and select the user combination with the largest number of minimum conditions of the channel matrix as the intra-cluster user. Here, the minimum condition number refers to the ratio of the minimum eigenvalue to the maximum eigenvalue of the channel matrix of a user combination.

For example, in one embodiment, assuming that there are a total of five users, and the current collaborative cluster can only serve three users, then the collaborative cluster needs to select three users from the five user types as intra-cluster users. Selecting 3 users from 5 users and having a total of (20) selection schemes, then the current collaboration cluster needs to compare the channel matrix of all 20 options (ie, all 20 user combination schemes) in all channel matrices. In the middle, select the user combination with the largest number of conditions as the user in the cluster.

In an embodiment, the base station and other base stations in the coordinating cluster to which it belongs may also pass the proportional The flat scheduling algorithm schedules all users in the collaborative cluster and selects users in the cluster.

S240, the base station and other base stations in the cluster to which it belongs determine the communication mode of the user in the cluster selected in S230;

In one embodiment, a joint pre-coded communication mode may be employed in downlink communications with other base stations within the cooperating cluster; in one embodiment, linear precoding may be employed for joint precoding. For example, ZF (zero forcing) linear precoding, MMSE (least mean square error) linear precoding, block orthogonal linear precoding, or oblique projection based linear precoding may be used for joint precoding; in one embodiment Joint precoding can also be performed using nonlinear precoding. For example, joint precoding may be performed using nonlinear precoding such as THP (Tomlinson-Harashima Precoding) or DPC (Dirty Paper Coding). It should be noted that the above pre-coding method is only an example of the embodiment of the present invention, and the embodiment of the present invention is not particularly limited.

In one embodiment, the communication mode of the joint detection may be employed for the user in the uplink communication. In one embodiment, a linear detection algorithm may be used for joint detection, such as a linear detection algorithm such as ZF or MMSE; in one embodiment, a non-linear detection algorithm may be used for joint detection, for example, SIC (Successive Interference Cancellation, String) Non-linear algorithms such as line interference cancellation), PIC (Parallel Interference Cancellation), or QR decomposition based algorithms; in one embodiment, an optimal detection algorithm can also be used for joint detection, for example, ML (Maximum Likelihood) , max (), SD (spherical decoding, Sphere Decoding) or LG (Lattice Reduction) algorithm.

It should be noted that the above algorithm is only an example of the embodiment of the present invention, and the embodiment of the present invention is not particularly limited.

S250. In the downlink communication, the base station and all other base stations in the coordinating cluster to which it belongs jointly allocate power to the intra-cluster users in S230.

In one embodiment, the water injection algorithm may be used to assign a rate to users within the collaborative cluster; in one embodiment, a water injection power allocation algorithm that does not consider inter-cluster interference may be employed; in one embodiment, inter-cluster considerations may be considered An iterative water injection power allocation algorithm based on game theory of interference. In an embodiment, a Greedy algorithm and a bit power allocation algorithm may be used to allocate a rate to a user within a collaborative cluster. It should be noted that the above algorithm is only an example of the embodiment of the present invention, and the embodiment of the present invention is not particularly limited.

As shown by the dashed box in FIG. 3, in one embodiment, the method further includes:

S210, in the second time period, the base station receives a cooperative cluster division result sent by the central controller, where the coordinated cluster division result is that the central controller performs, according to information about time average rates of all users in the first time period;

In one embodiment, the time average rate information for all users is fed back by the various base stations under the jurisdiction of the central controller; in one embodiment, the time average rate information for all users may also be fed back by all users themselves.

S260. The base station collects information about a time average rate of users under its jurisdiction;

In an embodiment, the related information of the time average rate of the user may include a cumulative average rate of the user under all cooperative cluster division schemes, a cumulative average probability of the user under the current cluster division scheme, and a user's accumulation under the current cluster division. Average rate.

In one embodiment, the cumulative average rate of user t under all cooperative clustering schemes is: 3⁄4 (r) = (1 - A ) 3⁄4 (/ - 1) + (0 3⁄4 ( ' where Α is the rate of user & Adjust the weight, A > o, i _k (t) to identify whether the user t is scheduled at the time, 3⁄4 (ή indicates the rate allocated to the user at the time of the clustering scheme Ω.

In one embodiment, the cumulative average probability of the user under the current cluster partitioning scheme Ω is:

3⁄4 ( = ( ¹ -Α)3⁄4('-ΐ)+Α ( ' where Α is the probability of user & adjustment weight, A > o.

In one embodiment, the cumulative average rate of the user under the current cluster partitioning scheme Ω is:

For the user to divide the cluster

r _ka t - l , otherwise rate adjustment weight, A > o.

S270. The base station feeds back to the central controller information about the time average rate of the users it governs.

According to the foregoing technical solution, the base station is dynamically clustered according to the gradient of the average rate utility function in the network in the first time period, and the user scheduling and resource allocation are performed in the cluster in the second time period; Separate program and resource allocation, separate computing tasks, and greatly reduce The implementation of the entire system is difficult, the system signaling overhead can be reduced, and better network performance can be achieved. In an embodiment of the present invention, a more detailed description is made on the cooperative cluster division of the base station and the selection of users in the cluster:

a. The utility function of all users in the network is represented by U:

Where is the time average rate of user A, (a utility function for the time average rate of user A: in one embodiment, a strictly monotonically increasing concave function is introduced.

In one embodiment, to maximize the performance of the entire network, it is desirable to maximize the utility functions of all users in the network. In one embodiment, the S lytolyar gradient algorithm (Stolyar's Gradient algorithm) can be used to maximize the utility function as:

ρ

p

· Η

H

Among them, equations (3) and (4) represent constraints, that is, (2) two constraints (3) and (4) need to be satisfied. The derivative of 1 is represented. W indicates the cluster division scheme

The Ω rate at which Ω appears at time ί.

/ ) is the user; t is the probability of being scheduled at time ί, 4(ί) = 1 means that the user is scheduled at time, /^) = 0 means that user 1 is not scheduled at time ί. /^) = _{Ω Ω} ( Μ Μ is the rate at which time is allocated to the user. Where W is the rate assigned to user A at the time of the clustering scheme Ω. 3⁄4 ) =

= 3⁄4 -! + ( - 3⁄4 - !)]; respectively for clustering scheme Ω, cluster/intra-cluster user set and base station set; ^ is the number of base station antennas, ^ is the number of user terminal antennas. On the other side _'^ _¾ = 0§ _£ ^ 5

+ N _n Bl This formula represents the channel capacity of all scheduled users (both clusters/all intra-cluster users) in the cluster/sump clustering scheme Ω. Where S is the bandwidth, 11^, and intra-cluster channels, intra-cluster precoding, and power allocation for all scheduled users of the clustering scheme Ω under clusters respectively; H _3⁄4 „, the cluster is divided into Ω lower clusters/'to Cluster/inter-cluster interference channel; N. is the noise power Ϊ 密度 density. \3⁄4 is the conjugate transpose, Η _Ω conjugate transpose, \^ is ¥,,, conjugate transpose, !! ^ is the conjugate transpose of !!^ b. The maximization utility function can be decomposed into two problems: Collaborative cluster partitioning selection: Ω*(?) (6)

User selection within the cluster: (t) = argmaxt; (R _k {t-\))r _k (f) (7)

M. Thus, the selection problem of the cooperative cluster partitioning for the first time period, that is, the equation (6) can be solved by the Gradient Projection Algorithm in one embodiment:

. ∑π _Ω =\ (9)

( ⁹ ) The formula is the constraint of ( ⁸ ). Here =^, where ^ is the probability that the user is scheduled when the cluster partitioning scheme is Ω, and the probability that the clustering scheme Ω appears, which is the average rate when the user k cluster partitioning scheme is Ω. Let _Ω , for (8),

∑u _k (R _k ) = ^u _k (∑ partial derivative:

Da =^∑ _k {R _k ) =∑U _k {R _t )^ _m ( 10) The obtained gradient vector for updating the corresponding cluster partitioning scheme in the gradient vector to Ω. The gradient vector of the corresponding cluster partitioning scheme Ω is updated to obtain D. D is the gradient set of all the collaborative clustering schemes after the update. The probability vector of cluster partitioning is updated in the restricted interval, and π is added to the weighted gradient D, and the obtained sum is projected on the space of 7Γ to obtain the updated probability vector 7t:

Where Pr / is the projection function and y is the gradient weight value, where 0. D is the gradient set of all cluster partitioning schemes, "is the N-dimensional probability vector, N is the number of all cluster partitioning schemes, the coordinates of the probability vector correspond to the partitioning probability of each cluster partitioning scheme, and the sum of the coordinates representing the probability vector ^π is 1 .

Here, it should be noted that the restricted interval is to let 71 satisfy the condition that the sum of its elements is 1, that is, (11) where ∑ ^.

After updating the partitioning probability of all the partitioning schemes of the cluster, the central controller may randomly select a cooperative clustering scheme corresponding to the partitioning probability according to the partitioning probability of the cooperative clustering scheme, and perform update of the cooperative cluster partitioning on the base station; The central controller may select a cooperative clustering scheme corresponding to the largest partitioning probability according to the partitioning probability of the cooperative clustering scheme, and perform uplink clustering update on the base station; in an embodiment, the central controller may also perform the foregoing two A cooperative cluster division update scheme is combined to perform cooperative cluster division update on the base station.

B2. For the intra-cluster user selection problem in the second time period, the following conditions are required for (7):

S-t. ∑ {t)≤ ■N N,

(4) The formula is the constraint of (7). (4) The meanings of the relevant parameters in the equations and (7) have been described in detail above and will not be mentioned here.

c. The average time rate and related information of the base station in the cooperative cluster to update the user:

The cumulative average rate of users & under all collaborative clustering schemes is:

A is the rate adjustment weight of user A:, and identifies the probability that user A: is scheduled at time, >3⁄4>0.

The cumulative average probability of user t under the cluster partitioning scheme Ω is: 3⁄4( = ( ¹ -A)3⁄4(fi)+ ( ( 13 ) where A is the probability adjustment weight of the user, A > 0.

The cumulative average rate of user k under the cluster partitioning scheme Ω is:

F ( = i ⁽¹ D. ^{( 1) +} , ^{if ( = 1} ( 14 ) \r _m (ί-l), otherwise

Where 3⁄4 is the probability adjustment weight of the user t under the cluster division Ω, >0. d. The base station in the cooperative cluster sends the updated related information to the central controller, and the central controller calculates the gradient of the utility function of the average time rate of all users in the network according to the relevant information according to the related information, and then according to (11) The partitioning probability of the cooperative clustering scheme is updated to update the partitioning of the cooperative cluster.

According to the foregoing technical solution, the base station is dynamically clustered according to the gradient of the average rate utility function in the network in a first time period, and user scheduling and resource allocation are performed in the cluster in the second time period; The scheme and resource allocation are performed separately, and the computing tasks are separated, which greatly reduces the difficulty of implementing the entire system, reduces the system signaling overhead, and can achieve better network performance.

As shown in FIG. 4, an embodiment of the present invention provides a cooperative resource scheduling apparatus, including:

The collecting module 310 is configured to collect information about a time average rate of all users under the jurisdiction of the device during a first time period;

In an embodiment, the collecting module 310 can receive information about the time average rate of the user in the first time period fed back by all the base stations, and obtain the correlation of the time average rate of all users under the jurisdiction of the device in the first time period. information.

In an embodiment, the collecting module 310 can receive information about the time average rate of the user in the first time period fed back by all the base station controllers, and obtain the time average rate of all users under the jurisdiction of the device in the first time period. Related information.

The gradient obtaining module 320 is configured to obtain, according to the collected information about the time average rate of each user in the first time period, a gradient of the utility function of the time average rate of all users in the network in the first time period;

In an embodiment, the gradient obtaining module 320 may calculate the gradient of the utility function of each user in parallel according to the time average rate information of each user in the first time period, and calculate each of the calculated The gradient of the user's utility function gives a gradient of the utility function of the time average rate of all users over the first time period. In one embodiment, the gradient of the utility function of each user is calculated in parallel, and the utility function of each user may be calculated in parallel to calculate the gradient of each utility function.

In an embodiment, the gradient obtaining module 320 may calculate a utility function of the time average rate of all users in the first time period according to the time averaged speed information of each user in the first time period, and then calculate the utility of all users. The gradient of the function yields a gradient of the utility function of all users' time averaged snapshots over the first time period. In one embodiment, the gradient acquisition module 320 may also sum the gradients of the utility functions of the time average rates of the respective users during the first time period to obtain a utility function of the time average rate of all users in the first time period. Gradient.

The probability update module 330 is configured to update the partitioning probability of the cooperative clustering scheme by the gradient obtained in the gradient obtaining module 320;

a partitioning update module 340, configured to perform cooperative clustering update on the base station according to the updated partitioning probability of the probability update module 330;

The informing module 350 is configured to notify all base stations of the result of the cooperative cluster partition update, so that the base stations in each cooperative cluster select corresponding intra-cluster users (ie, perform user scheduling) in the second time period, and perform intra-cluster users. Make resource allocations.

As shown in FIG. 5, in one embodiment, the probability update module 330 can include:

The first probability update unit 331, is configured to update the gradient of the current cluster division scheme according to the gradient obtained by the gradient acquisition module 320;

a second probability update unit 332, configured to update a gradient set of all the cooperative clustering schemes by using a gradient of the updated current collaborative clustering scheme;

The third probability update unit 333 is configured to update the partition probability of each cooperative cluster partitioning scheme by using the formula 71 Pr _∑ +, which is a projection function, and D is an updated all cooperative cluster partitioning scheme. The gradient set is a gradient weighted value, ^ ^{> 0} , π is a w-dimensional probability vector, N is the number of all cluster partitioning schemes, and the coordinates of the probability vector correspond to the partitioning probability of each cluster partitioning scheme, ∑^ indicates The sum of the coordinates of the probability vector is 1; the formula indicates that 71 is added to the y-weighted D, and the resulting sum is projected on the space of 71 to obtain the updated probability vector.

Here, it is noted _{that, π Pro 7Σ (π + rD} ) in Σ _N, [pi] represents the updated probability vector is to meet this, i.e., the communication probability vector within a confined space under conditions of 2 ^; Γ.

As shown in FIG. 6, in one embodiment, the partition update module 340 can include;

The first update unit 341 is configured to: in the split probability that is updated by the probability update module 330, randomly select a cooperative cluster partitioning scheme corresponding to the partitioning probability, and perform an update of the cooperative clustering on the base station; and the second updating unit 342 is configured to Among the partitioning probabilities updated by the probability update module 330, the cooperative clustering scheme corresponding to the largest partitioning probability is selected, and the cooperative clustering is updated for the base station.

It should be noted that the cooperative resource scheduling apparatus mentioned in this embodiment may be a central controller, a gateway, or other network element having similar functions.

According to the foregoing technical solution, the base station is dynamically clustered according to the gradient of the time average rate utility function in the network in the first time period, and the user scheduling is performed in the cluster in the second time period (ie, selecting the intra-cluster user) And resource allocation; this clustering scheme and resource allocation are performed separately, and the computing tasks are separated, which greatly reduces the implementation difficulty of the whole system, can reduce the system signaling overhead, and achieve better network performance.

As shown in FIG. 7, an embodiment of the present invention provides a base station, including:

The sharing cooperation module 410 is configured to share data with other base stations in the cooperation cluster to which the base station belongs according to the cooperative cluster division result in the second time period; the cooperative cluster division result is that the cooperative resource scheduling apparatus is in the first time period according to all users. Made by information about the time average rate within;

In one embodiment, the shared collaboration module 410 shares data with other base stations within the cooperative cluster to which the base station belongs, and may include shared channel information; in one embodiment, the channel information may also be shared.

The user scheduling module 420 is configured to maximize the utility function of the cooperative cluster to which the base station belongs, and jointly select the users in the cluster together with other base stations in the cooperative cluster to which the base station belongs;

The communication mode determining module 430 is configured to determine, with the other base stations in the cluster to which the base station belongs, and the uplink and downlink communication modes of the users in the cluster selected by the user scheduling module 420;

The power distribution module 440 is configured to perform all the other in the downlink communication and the cooperation cluster to which the base station belongs. The base station jointly allocates power to users within the cluster.

In one embodiment, the water injection algorithm may be used to allocate a rate to users within the cooperative cluster; in one embodiment, a water injection power allocation algorithm that does not consider inter-cluster interference may be employed; in one embodiment, inter-cluster consideration may be considered An iterative water injection power allocation algorithm based on game theory of interference. In one embodiment, a Greedy algorithm and a bit power allocation algorithm may also be employed to allocate rates for users within the collaborative cluster.

As shown in the dashed box in FIG. 7, in an embodiment, the apparatus may further include: a statistics module 450, configured to collect a correlation signal sending module 460 for calculating a time average rate of users controlled by the base station, Information about the average time rate of the user under the jurisdiction of the base station is sent to the cooperative resource scheduling device.

As shown in FIG. 8, in one embodiment, the user scheduling module 420 can include:

The first scheduling unit 421 is configured to compare, with other base stations in the cooperative cluster, channel capacity of all user combinations in the cooperative cluster, and select a user combination with the largest capacity as the intra-cluster user;

The second scheduling unit 422 is configured to compare channels of all users in the cluster with other base stations in the cooperative cluster, and select a user combination with the largest number of conditions as the intra-cluster user; the minimum condition number refers to a channel matrix of a user combination. The ratio of the minimum eigenvalue to the largest eigenvalue.

The third scheduling unit 433 is configured to, with other base stations in the cooperative cluster, schedule all users in the cooperative cluster by using a proportional fair scheduling algorithm, and select a user in the cluster.

As shown in FIG. 9, in an embodiment, the communication mode determining module 430 may include: a downlink communication determining unit 431, configured to use a pre-coding communication mode with other base stations in the cooperative cluster in downlink communication;

In one embodiment, joint precoding may be performed using linear precoding. For example, ZF linear precoding, MMSE linear precoding, block orthogonal linear precoding, or oblique projection based linear precoding may be used for joint precoding; in one embodiment, nonlinear precoding may also be used for joint precoding. coding. For example, joint precoding can be performed using nonlinear precoding such as THP or DPC.

The uplink communication determines a single far 432, which is used for communication in conjunction with other base stations in the cooperative cluster in uplink communication.

In one embodiment, a joint detection can be performed using a linear detection algorithm, such as ZF or Linear detection algorithm such as MMSE; In one embodiment, a non-linear detection algorithm may be used for joint detection, for example, a non-linear algorithm such as SIC, PIC, or a QR decomposition based algorithm; in one embodiment, an optimal method may also be employed. The detection algorithm performs joint detection, for example, ML, SD, or a subtractive algorithm.

According to the foregoing technical solution, the base station is dynamically clustered according to the gradient of the average rate utility function in the network in the first time period, and the user scheduling and resource allocation are performed in the cluster in the second time period; The scheme and resource allocation are performed separately, and the computing tasks are separated, which greatly reduces the difficulty of implementing the entire system, reduces the system signaling overhead, and can achieve better network performance.

As shown in FIG. 10, an embodiment of the present invention provides a cooperative resource scheduling system, including multiple base stations 10 and a cooperative resource scheduling apparatus 20.

The cooperative resource scheduling device 20 is configured to collect information about a time average rate of all users under the jurisdiction of the device in a first time period; and according to the collected information about the time average rate of each user in the first time period, Obtaining a gradient of the utility function of the time average rate of all users in the network in the first time period; updating the division probability of the cooperative cluster division scheme by the obtained gradient; performing cooperative cluster division update on the base station according to the updated division probability; The result of the cooperative cluster division update notifies all base stations, so that the base stations in each cooperative cluster select the corresponding intra-cluster users in the second time period, and perform resource allocation on the users in the cluster.

The specific structure and function of the cooperative resource scheduling apparatus 20 have been described in detail in the foregoing embodiments, and are not described herein again.

The base station 10 is configured to share data with other base stations in the cooperation cluster to which the base station 10 belongs according to the cooperative cluster division result; the cooperative cluster division result is that the cooperative resource scheduling apparatus performs information according to the time average rate of all users in the first time period. The utility function of the cooperative cluster to which the station 10 belongs is the largest target, and the users in the cluster are jointly selected with other base stations in the cooperative cluster to which the base station belongs; and the decision of the other base stations in the cluster to which the base station 10 belongs and the user scheduling module 420 selects The uplink and downlink communication mode of the user in the cluster; in the downlink communication, all other base stations in the cooperation cluster to which the base station 10 belongs are jointly allocated power for the users in the cluster.

In an embodiment, the base station 10 can also be used to count the time flat of the user under the control of the base station. Information about the average rate; information about the time average rate of the users under the control of the base station is sent to the cooperative resource scheduling apparatus.

The specific structure and function of the base station 10 have been described in detail in the foregoing embodiments, and details are not described herein again. In the present embodiment, the base station in Fig. 10 is divided into five clusters as shown by the curves in the figure. Of course, it can be understood that these base stations may be divided into 6 or 4 clusters in the next first time period.

A person skilled in the art can understand that all or part of the process of implementing the above embodiment method can be completed by a computer program to instruct related hardware, and the program can be stored in a computer readable storage medium, the program When executed, the flow of an embodiment of the methods as described above may be included. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

The above is only a few embodiments of the present invention, and those skilled in the art can make various changes or modifications to the invention without departing from the spirit and scope of the invention.

Claims

Rights request

A cooperative resource scheduling method, which is characterized in that:

Collecting information about the average time rate of all users in the first time period; obtaining the average time rate of all users in the first time period according to the information about the time average rate of each user in the first time period The gradient of the utility function;

2. The cooperative resource scheduling method according to claim 1, wherein the time average rate of all users in a first time period is obtained according to related information of time average rates of respective users in a first time period. The gradient of the utility function, including:

Calculating the gradient of each user's function in parallel according to the information about the time average rate of each user in the first time period;

The gradients of the various utility functions are summed to obtain a gradient of the utility function of the time average rate of all users over the first time period.

The cooperative resource scheduling method according to claim 1, wherein the time average rate of all users in the first time period is obtained according to the information about the time average rate of each user in the first time period. The gradient of the utility function, including:

Calculating a utility function of the time average rate of all users in the first time period according to the time averaged record information of each user in the first time period;

The gradient of the utility function of all users is calculated, and the gradient of the utility function of the time average of all users in the first time period is obtained.

The cooperative resource scheduling method according to claim 1, wherein the partitioning probability of each cooperative clustering scheme is updated according to a gradient of a utility function of the time average rate of all users in a first time period, include: Updating the gradient of the current cluster partitioning scheme according to the gradient of the utility function of the time average rate of all users in the first time period;

Updating the gradient sets of all collaborative cluster partitioning schemes using the gradient of the updated current collaborative clustering scheme;

The partitioning probability of each cooperative cluster partitioning scheme is updated by the formula π Proj _{∑ π} ^ _χ (π + ^D), Proj is the projection function, and D is the gradient set of all the cooperative clustering schemes after updating, which is the gradient weighting value, r ^{> 0} , "is an N-dimensional probability vector, W is the number of all cluster partitioning schemes, the coordinates of the probability vector π correspond to the partitioning probability of each cluster partitioning scheme, ∑ _Λ , _Ν - represents the sum of the coordinates of the probability vector" It is 1; the formula indicates that π is added to the weighted D, and the obtained sum is projected on the space of π to obtain the updated probability vector π.

The cooperative resource scheduling method according to claim 1, wherein the cooperative clustering update is performed on all the base stations according to the updated partitioning probability, including:

In the updated partitioning probability, a cooperative clustering scheme corresponding to the partitioning probability is randomly selected, and the cooperative cluster is updated and updated to the base station;

Or, in the updated partitioning probability, the cooperative clustering scheme corresponding to the largest partitioning probability is selected, and the cooperative cluster is updated and updated.

6. The cooperative resource scheduling method according to claim 1, wherein the utility function is a steerable monotonically increasing function.

A cooperative resource scheduling method, the method comprising:

And determining, by the other base stations in the cooperative cluster, an uplink-downlink communication mode of the user in the cluster; and performing, in downlink communication, with other base stations in the cooperative cluster to allocate power to the users in the cluster.

The cooperative resource scheduling method according to claim 7, wherein the selecting the utility function of the cooperative cluster to maximize the target, and jointly selecting the users in the cluster with other base stations in the cooperative cluster, includes: Comparing the channel capacity of all user combinations in the cooperative cluster with other base stations in the cooperative cluster, and selecting the user combination with the largest capacity as the intra-cluster user;

Or comparing, with other base stations in the cooperative cluster, a channel matrix of all user combinations in the cooperative cluster, and selecting a user combination with a largest minimum number of channel matrices as a user in the cluster, and the minimum condition number of the channel matrix is The ratio of the minimum eigenvalue to the largest eigenvalue of the channel matrix;

Or, with other base stations in the cooperative cluster, all users in the cooperative cluster are scheduled by a proportional fair scheduling algorithm, and users in the cluster are selected.

The cooperative resource scheduling method according to claim 7, wherein the determining, by the other base stations in the cooperation cluster, the uplink and downlink communication manners of the users in the cluster, the method includes:

In downlink communication, a joint precoding communication mode is adopted with other base stations in the cooperative cluster; in uplink communication, a joint detection communication mode is adopted for the user with other base stations in the cooperative cluster.

The cooperative resource scheduling method according to claim 7, wherein the other base stations in the downlink communication communicate with the other intra-cluster clusters to allocate power to the intra-cluster users, including: And the other base stations in the associated cooperative cluster adopt a water injection power allocation algorithm that does not consider inter-cluster interference, and allocate power to users in the cluster;

Alternatively, other base stations in the downlink communication and in the associated cooperative cluster employ a game theory based iterative water injection power allocation algorithm that considers inter-cluster interference to allocate power to users within the cluster.

The cooperative resource scheduling method according to claim 7, wherein the method further comprises:

Information on the average rate of households under statistical jurisdiction;

The statistically related information of the average speed record of the user of the jurisdiction is transmitted to the cooperative resource scheduling apparatus.

The cooperative resource scheduling method according to claim 11, wherein the information about the average speed of the user under the jurisdiction includes:

The cumulative average rate of the user under the jurisdictional clustering scheme, the cumulative average probability of the user under the current cluster partitioning scheme, and/or the user of the jurisdiction in the current cluster partitioning The cumulative average rate under the case.

13. A cooperative resource scheduling apparatus, comprising:

a collecting module, configured to collect information about a time average rate of all users under the jurisdiction of the device during a first time period;

a gradient obtaining module, configured to obtain a gradient of a utility function of a time average rate of all users in a first time period according to information about a time average rate of each user in a first time period; a probability update module, configured to pass the The gradient obtained by the gradient acquisition module updates the division probability of each cooperative cluster division scheme;

The notification module is configured to notify all base stations of the result of the cooperative cluster division update, so that the base stations in each cooperation cluster select the corresponding intra-cluster users in the second time period, and perform resource allocation on the users in the cluster.

The cooperative resource scheduling apparatus according to claim 13, wherein the probability update module comprises:

a first probability updating unit, configured to update a gradient of the current clustering scheme according to the gradient obtained by the gradient acquiring module;

a second probability update unit, configured to update a gradient set of all cooperative cluster partitioning schemes by using a gradient of the updated current collaborative cluster partitioning scheme;

a third probability updating unit, configured to update a partitioning probability of each cooperative cluster partitioning scheme by using a formula π Proj _∑ π _Η _ {π + γΌ, / is a projection function, and D is a gradient set of all the cooperative clustering schemes after updating, Gradient weighting value, > o , π is a w-dimensional probability vector, N is the number of all clustering schemes, the coordinates of the probability vector correspond to the division probability of each cluster division scheme, and the sum of the coordinates of the probability vector ^π Is 1; the formula indicates that 71 is added to the weighted D, and the resulting sum is projected onto the space of r; to obtain an updated probability vector.

The cooperative resource scheduling apparatus according to claim 13, wherein the partitioning module comprises a first updated unit or a second updating unit;

The first updating unit is configured to: randomly select a cooperative cluster partitioning scheme corresponding to the partitioning probability, and perform cooperative clustering update on the base station; The second updating unit is configured to select a cooperative clustering scheme corresponding to the largest partitioning probability in the updated partitioning probability, and perform cooperative clustering update on the base station.

16. A base station, comprising:

The base station according to claim 16, wherein the user scheduling module comprises a first scheduling unit, a second scheduling sheep element, or a third scheduling unit.

The first scheduling unit is configured to compare channel capacity of all user combinations in the cooperative cluster with other base stations in the cooperative cluster, and select a user combination with the largest capacity as a user in the cluster;

The second scheduling unit is configured to compare, with other base stations in the cooperative cluster, a channel matrix of all user combinations in the cooperative cluster, and select a user combination with a minimum number of minimum condition numbers of the channel matrix as a user in the cluster, The minimum condition number of the channel matrix is the ratio of the minimum eigenvalue to the maximum eigenvalue of the channel matrix;

The third scheduling unit is configured to schedule, by using a proportional fair scheduling algorithm, all users in the cooperative cluster to select other users in the cluster.

The base station according to claim 16, wherein the base station further comprises: a statistics module, configured to collect information about an average rate of the occupied households;

And a sending module, configured to send, to the cooperative resource scheduling device, the related information of the average speed record of the user of the jurisdiction.

19. A cooperative resource scheduling system, comprising: any one of claims 13-15 A coordinated resource scheduling apparatus as claimed in any of the preceding claims and a plurality of base stations according to any of claims 16-18.