CN115866789A

CN115866789A - Wireless network interference coordination and resource scheduling method and device based on hierarchical clustering algorithm

Info

Publication number: CN115866789A
Application number: CN202211581860.XA
Authority: CN
Inventors: 彭涛; 郭异辰; 赵誉洁; 牛怡静
Original assignee: Beijing University of Posts and Telecommunications
Current assignee: Beijing University of Posts and Telecommunications
Priority date: 2022-12-09
Filing date: 2022-12-09
Publication date: 2023-03-28

Abstract

The application provides a wireless network interference coordination and resource scheduling method based on a hierarchical clustering algorithm, and relates to the technical field of wireless communication, wherein the method comprises the following steps: acquiring a user equipment set with unsatisfied requirements; classifying the devices with mutual interference smaller than a threshold value in the user equipment set into the same class according to a hierarchical clustering algorithm, classifying the devices with mutual interference larger than the threshold value into different classes, and obtaining an optimal non-overlapping clustering result; adjusting and optimizing the optimal non-overlapping clustering result to enable different classes to overlap and obtain an optimal clustering result; and allocating wireless resource units for each type according to the optimal clustering result, and stopping allocation if the requirements of all the devices are met or the resources are exhausted to obtain a final resource allocation result. The invention adopting the scheme realizes that the wireless resource scheduling satisfies the QoS of all the devices and effectively reduces the consumption of the wireless resources at the same time.

Description

Wireless network interference coordination and resource scheduling method and device based on hierarchical clustering algorithm

Technical Field

The present application relates to the field of wireless communication technologies, and in particular, to a method and an apparatus for wireless network interference coordination and resource scheduling based on a hierarchical clustering algorithm.

Background

With the rapid increase of network capacity demand, wireless networks nowadays often adopt a mode of increasing the deployment density of network devices to meet the demand of network capacity. In the current 5G wireless network, no matter in Enhanced Mobile Broadband (eMBB), or in scenarios such as high-reliable Low-latency Communications (URLLC), the more intensive deployment of base stations can significantly reduce the transmission distance between the base station and the Mobile terminal, thereby reducing transmission loss, and further improving network capacity and reliability. However, the dense deployment of small base stations will cause severe Co-channel Interference (CCI), which brings new challenges to wireless communication networks.

In the existing cellular network, the radio resource allocation function is performed by the base station, and each cell basically performs management and allocation of radio resources independently. As one of core functions of a Media Access Control (MAC) layer, a scheduler in a base station performs scheduling of radio resources according to a result given by a radio resource scheduling algorithm by obtaining feedback of information such as a retransmission state, a buffer state, and a channel Quality state of each device in a cell with respect to a Hybrid Automatic Repeat reQuest (HARQ) process, and a buffer state at a base station side, a historical throughput of each device, and a Quality of Service (QoS) requirement. As can be seen, the radio resource scheduling algorithm is the core of radio resource scheduling.

Common resource scheduling algorithms in the existing network are mainly classified into the following three types:

maximum throughput class: such algorithms are primarily concerned with the maximum throughput that can be achieved by the wireless network. Base stations using such algorithms tend to try to allocate resources to several devices with better channel quality, while allocating few resources to other devices. Therefore, although higher network throughput can be achieved by using the algorithm, the network fairness is very poor;

polling type: such algorithms focus primarily on network fairness. The base station using such an algorithm allocates resources to each device in turn, with the devices being fairly equal. The algorithm can ensure better fairness, but the network performance is lower, and when the number of users in a cell is more, the time delay caused by polling is often more serious, and the QoS requirements of certain services cannot be met;

proportional fair class: the algorithm takes fairness and performance into consideration, and is the most common algorithm. The core idea of the algorithm is that in each scheduling period, a priority index is calculated according to the historical throughput, the channel quality and the like of each device, and each device is scheduled according to the index. Generally, the lower the historical throughput of the device, the better the channel quality, and the higher the priority. In addition, besides historical throughput and channel quality, the priority index of the proportional fair algorithm can be calculated by combining other information so as to match the requirements of the wireless network.

The resource allocation algorithm in the cell adopted in the existing network does not consider the CCI condition between cells, so the performance is seriously influenced in the environment of dense networking. Therefore, an additional inter-cell interference coordination mechanism, such as ICIC, eICIC or CoMP, is often adopted in the existing network to make up for the defect.

However, ICIC and eICIC techniques rely heavily on signaling exchange, resulting in poor timeliness of interference information and severely limited data size. Moreover, in a densely deployed network, the number of neighboring cells is extremely large, and signaling exchange will cause extremely serious overhead; in CoMP, a large amount of computing resources are required for signal processing, and frequent channel measurement is required, which occupies a large amount of pilot resources and affects normal operation of the network. Therefore, current wireless networks cannot handle the problem of inter-cell interference well.

In academic circles, most of the existing research on multi-cell resource allocation algorithms is based on the assumption that the channel information of the whole network is completely and timely known, but the assumption is not true in the real wireless network. Based on this assumption, the academic community also ignores the influence of small-scale fading in the research on the resource allocation algorithm, so that the existing algorithm cannot guarantee the QoS well in the scenario with higher QoS requirement, such as URLLC.

Disclosure of Invention

The present application is directed to solving, at least to some extent, one of the technical problems in the related art.

Therefore, a first objective of the present application is to provide a method for coordinating wireless network interference and scheduling resources based on a hierarchical clustering algorithm, which solves the technical problem that the existing method cannot well handle inter-cell interference, and achieves that the QoS of all devices is met and the consumption of wireless resources is effectively reduced when scheduling wireless resources.

The second objective of the present application is to provide a wireless network interference coordination and resource scheduling device based on a hierarchical clustering algorithm.

In order to achieve the above object, an embodiment of the first aspect of the present application provides a method for wireless network interference coordination and resource scheduling based on a hierarchical clustering algorithm, including: acquiring a user equipment set with unsatisfied requirements; classifying the devices with mutual interference smaller than a threshold value in the user equipment set into the same type according to a hierarchical clustering algorithm, and classifying the devices with mutual interference larger than the threshold value into different types to obtain an optimal non-overlapping clustering result; adjusting and optimizing the optimal non-overlapping clustering result to enable different classes to overlap and obtain an optimal clustering result; and allocating wireless resource units for each type according to the optimal clustering result, and stopping allocation if the requirements of all the devices are met or the resources are exhausted to obtain a final resource allocation result.

According to the wireless network interference coordination and resource scheduling method based on the hierarchical clustering algorithm, on the premise that accurate and complete channel measurement information does not need to be mastered, the interference relation among network devices is obtained by using the existing interference modeling technology, and interference avoidance, coordination and resource scheduling are carried out on the network devices according to the interference relation by means of algorithms such as hierarchical clustering, so that the QoS of all devices is met, and meanwhile, the consumption of wireless resources is reduced as much as possible.

Optionally, in an embodiment of the present application, the hierarchical clustering algorithm is an aggregation-type hierarchical clustering algorithm, and classifying the set of user equipments according to the hierarchical clustering algorithm includes:

clustering the user equipment set according to the difference between the classes by an agglomeration type hierarchical clustering algorithm to obtain a binary convergence tree, and setting different heights for the binary convergence tree to obtain a clustering result set, wherein the difference between the classes is determined by a link function, the equipment in each class contained in each clustering result uses the same resource, and the equipment in different classes uses different resources;

and screening the clustering result set according to the expected resource occupation amount and the achievable total rate of each clustering result to obtain the optimal non-overlapping clustering result, wherein the expected resource occupation amount is the sum of the expected maximum resource amount of all classes in the clustering result, the achievable total rate is the sum of the rates which can be achieved by all the classes in the clustering result, and the sum of the rates which can be achieved by each class is the sum of the rates which can be achieved by all the devices contained in the class.

Optionally, in an embodiment of the present application, the determining process of the link function is:

defining a distance function determining the difference between the user equipments, and determining a link function based on the distance function,

wherein the distance function is represented as:

the link function is represented as:

wherein the content of the first and second substances,

presentation device U _p As signalling means, slave unit U _q Reciprocal signal to interference ratio, based on the signal to interference ratio>

Presentation device U _q As mailNumber device, receiving device U _p Inverse signal-to-interference ratio, j, of interference _p Presentation device U _p Associated base station number j _q Presentation device U _q Associated base station number j _q Presentation device U _q Associated base station number, <' > in>

Represents the mth class in the clustering result, is greater than or equal to>

Represents the nth class in the clustering result, and d (p, q) represents the difference between device p and device q.

Optionally, in an embodiment of the present application, the adjusting and optimizing the optimal non-overlapping clustering result includes:

obtaining all classes with heights smaller than a preset height in the binary convergence tree to obtain a class set to be merged, wherein the preset height is the height corresponding to the optimal non-overlapping clustering result, and each element in the class set to be merged is assigned with a subscript in a descending order according to the corresponding height;

judging whether the current class to be merged in the class set to be merged meets a preset condition, if so, merging the current class to be merged with the optimal non-overlapping clustering result, and updating the optimal non-overlapping clustering result by using the merged optimal non-overlapping clustering result;

continuously updating the optimal non-overlapping clustering result until all classes in the class set to be merged are judged, and obtaining the final optimal non-overlapping clustering result as the optimal clustering result;

wherein the preset conditions are as follows: the intersection of the current updated optimal non-overlapping clustering result and the current class to be merged is empty, all devices in the merged clusters use the same resource, and the resource utilization rate after merging is greater than that before merging, wherein the resource utilization rate after merging is greater than that before merging, and the method comprises the following steps: the combined expected resource occupancy is greater than the combined expected resource occupancy or the combined expected resource occupancy is equal to the combined achievable total rate before and after the combination.

Optionally, in an embodiment of the present application, allocating a radio resource unit to each class according to the optimal clustering result, and stopping allocation if the demands of all the devices are met or the resources are exhausted, includes:

allocating 1 wireless resource unit for each class in the optimal clustering result, and calculating the reachable rate of each user equipment after allocation, wherein if the user equipment exists in a plurality of classes, the reachable rate is the sum of the reachable rates of the user equipment in each class;

updating the rate requirement of each user equipment according to the reachable rate of each user equipment, and adding the currently completed resource allocation into a resource allocation result list;

and if the updated rate requirements of all the user equipment are reduced to 0 or the resources are exhausted, scheduling the resources according to the current resource allocation result list, and otherwise, re-executing the wireless network interference coordination and resource scheduling method based on the hierarchical clustering algorithm.

Optionally, in an embodiment of the present application, if there is a traffic with a QoS requirement greater than a threshold in the ue set, then, applying a small-scale fading compensation mechanism to the wireless network interference coordination and resource scheduling method based on the hierarchical clustering algorithm, including:

when a hierarchical clustering algorithm is used for classifying a user equipment set and adjusting and optimizing an optimal non-overlapping clustering result, the sum of the reachable rates of all user equipment to which a fading protection boundary is applied is used as a reachable total rate, the optimal transmission times are determined, and the product of the optimal transmission times and the expected maximum required resource amount is used as the expected resource occupation amount;

when the wireless resource units are allocated to each type according to the optimal clustering result, the allocated wireless resource units are changed from 1 to the optimal transmission times, and the reachable rate of the user equipment after the application of the fading protection boundary is used as the reachable rate of the user equipment.

In order to achieve the above objects, a second aspect of the present invention provides a wireless network interference coordination and resource scheduling apparatus based on hierarchical clustering algorithm, including an obtaining module, a clustering optimization module, and a resource allocation module, wherein,

the acquisition module is used for acquiring a user equipment set with unsatisfied requirements;

the clustering module is used for classifying the devices with mutual interference smaller than a threshold value in the user equipment set into the same type according to a hierarchical clustering algorithm, and classifying the devices with mutual interference larger than the threshold value into different types to obtain an optimal non-overlapping clustering result;

the clustering optimization module is used for adjusting and optimizing the optimal non-overlapping clustering result to enable different classes to overlap and obtain the optimal clustering result;

and the resource allocation module is used for allocating wireless resource units to each type according to the optimal clustering result, and stopping allocation if the requirements of all the devices are met or the resources are exhausted to obtain the final resource allocation result.

Optionally, in an embodiment of the present application, the hierarchical clustering algorithm is an aggregation-type hierarchical clustering algorithm, and the clustering module is specifically configured to:

clustering a user equipment set according to the difference between classes by using an agglomeration type hierarchical clustering algorithm to obtain a binary convergence tree, and setting different heights for the binary convergence tree to obtain a clustering result set, wherein the difference between the classes is determined by a link function, the equipment in each class contained in each clustering result uses the same resource, and the equipment in different classes uses different resources;

Optionally, in an embodiment of the present application, the cluster optimization module is specifically configured to:

Optionally, in an embodiment of the present application, the resource allocation module is specifically configured to:

Additional aspects and advantages of the present application will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the present application.

Drawings

The above and/or additional aspects and advantages of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:

fig. 1 is a schematic diagram of BMT obtained by hierarchical clustering according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a 3GPP dual-strip model according to an embodiment of the present application;

fig. 3 is a schematic flowchart of a method for wireless network interference coordination and resource scheduling based on a hierarchical clustering algorithm according to an embodiment of the present application;

FIG. 4 is a graph comparing performance of a proposed scheme and a comparison scheme in a small scale network according to an embodiment of the present application;

FIG. 5 is a graph comparing performance of proposed and comparative schemes in a large scale network according to an embodiment of the present application;

FIG. 6 is a schematic diagram illustrating the probability of decoding errors generated by the proposed scheme and the comparison scheme according to the embodiment of the present application under different types of average transmission times;

FIG. 7 is a schematic diagram illustrating the probability of decoding errors generated by the proposed scheme and the comparison scheme according to the embodiment of the present application under different types of average transmission times;

FIG. 8 is a schematic diagram illustrating the probability of decoding errors generated under different RU occupancy for the proposed scheme and the comparative scheme of the embodiment of the present application;

fig. 9 is a schematic structural diagram of a wireless network interference coordination and resource scheduling device based on a hierarchical clustering algorithm according to an embodiment of the present application.

Detailed Description

Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the accompanying drawings are illustrative and intended to explain the present application and should not be construed as limiting the present application.

The techniques used in the solution of the present application are briefly described below:

and (3) clustering algorithm: the Clustering algorithm is one of unsupervised machine learning, and is to divide a data set into different classes or clusters according to a certain criterion (such as distance), so that the similarity of data objects in the same cluster is as large as possible, and the difference of data objects not in the same cluster is also as large as possible. That is, after clustering, the data of the same class are gathered together as much as possible, and the data of different classes are separated as much as possible. The clustering method can be mainly divided into methods such as division type clustering, density-based clustering, hierarchical clustering and the like.

Agglomeration-type Hierarchical Clustering (AHC) is mainly used in the present application. Set of data points to be clustered S = { S = { S } _i I =1,2, \8230;, N }, defining a "leaf" as a class containing only 1 data point and a "root" as a class containing all data points. In AHC, each data point is initially considered a leaf. In each step of the algorithm, the two most similar classes are combined into a new larger class until all data points are contained in one class, i.e. a Binary Merge Tree (BMT) is formed, which is an example of a BMT with N =6, as shown in fig. 1.

AHC relies on a non-negative link function (Linkage function) to determine the inter-class variability.

Step 1, the rest classes are collected

Initialise to S, i.e. all classes are leaves.

Let the remaining classes in the k step be collected as

Class->

And/or>

Note of the linkage function betweenIs composed of

Then in the clustering process, two classes with the highest similarity (i.e. the smallest difference) are found at each step:

and merge the two classes into a new class

Define its height as->

Updating the residual class set:

iteration is continued until only 1 class (i.e., root) remains in the remaining class set, and the algorithm stops. Note that the set formed by all new classes (called nodes) generated by merging other classes is N = { N = ₁ ,…,N _N-1 H = { h } for the corresponding height set ₁ ,…,h _N-1 }. Obviously, the height sets are arranged in ascending order (i.e., with h) ₁ ≤h ₂ ≤…≤h _N-1 ). Further, the height of the leaf is defined as 0.

From the above introduction, the link function directly determines the final result of the AHC. Therefore, the link function should be defined reasonably according to the problem to be solved.

Wireless network interference and system model:

fig. 2 is a schematic diagram of a 3GPP dual-strip model, as shown in fig. 2, in a small-range wireless service hot spot area, such as an office, a movie theater, or an intelligent factory with a large number of sensors and industrial devices, in order to meet throughput or QoS requirements, an operator will deploy a large number of small base stations. Two rows of rooms are arranged at two sides of the corridor, and each row consists of N _r And (4) forming rooms. There is one small base station and q devices in each room, thus sharing R =4N _r Small base station and M =4N _r q devices. In addition, the system has K allocatable radio Resource Units (RUs) in common. Recording the set of small base stations as C = { C ₁ ,…,C _R The equipment set is U = { U = } ₁ ,…,U _M }, any equipment U _n Has a minimum rate requirement of R _n, 。

The solution proposed in the present application performs interference coordination and resource scheduling according to a wireless interference model, and therefore, the interference model required in the present application needs to be described.

In the uplink direction of the mobile communication network, the serving device U _m Reach home cell C using uplink signals transmitted by kth RU _jm The signal receiving power of the base station is:

wherein

For the k RU upper equipment U _m To home cell C _jm The channel gain of the base station includes large scale fading substantially independent of RU and small scale fading associated with RU.

Interference device U for transmitting signals also via the k-th RU _n To cell C _jm The interference signal power of the base station is:

wherein

For the k RU upper equipment U _n To cell C _jm The channel gain of the base station also includes large scale fading that is substantially independent of RU and small scale fading that is dependent on RU.

Thereby, the deviceOn the k RU, service device U _m Is in the home cell C _jm The Uplink signal-to-interference-plus-noise ratio (UL-SINR) at the base station end is:

wherein, the first and the second end of the pipe are connected with each other,

as and service device U _m Set of interfering devices, σ, multiplexing the kth RU ² Is the noise power.

The above formula can be obtained by further transformation:

wherein the content of the first and second substances,

for service device U _m Received Signal-to-noise ratio (SNR) at the kth RU, -, (R) in the receiver>

For service device U _m With interfering device U multiplexing the kth RU _n Signal-to-interference ratio (SIR) between.

In practical wireless networks, the conditions of small-scale fading are not available in real time and cannot be predicted. Therefore, in the proposed scheme, only statistical channel gains involving large-scale fading will be used for the estimation of network performance. Corresponding to device U _m Statistical UL-SINR, SNR and U of _m And U _n Are respectively recorded as gamma _m ，

And gamma _m, 。

Order to

Then the above equation reduces to:

recording and service device U _m The set of interfering devices occupying a same radio resource is N _m . As can be seen from the above equation, the reciprocal of the statistical UL-SINR

Can be served by a service device U _m With each interfering device U occupying the same radio resources _n Statistical inverse SIR in between>

And serving device statistical SNR reciprocal &>

And linear addition. Further, densely deployed wireless networks are typically interference limited, with the effects of noise being negligible compared to interference. Therefore, the proposed scheme will coordinate and optimize the interference based on the statistical SIR. The interference model used in this application should also include the statistical SIR information of all pairs of devices that may cause interference, as well as the statistical SNR information of each device.

It should be noted that the configuration of the set of interfering devices in different orthogonal resource sharing systems may be different. In systems such as Orthogonal Frequency Division Multiplexing (OFDMA) or Time Division Multiplexing (TDMA), each device located in the same cell does not have interference, and therefore the set of interfering devices only includes devices that are not in the same cell as the serving device; however, in Code Division Multiple Access (CDMA) and Non-Orthogonal multiplexing (NOMA) systems, interference may also exist between devices located in the same cell, and thus the set of interfering devices includes all devices in the system except the serving device. However, since the derivation of the interference relationship is not specific to a specific orthogonal resource multiplexing method, the interference coordination and resource allocation scheme proposed in the present application can be applied to various wireless systems using different orthogonal resource multiplexing methods by constructing different sets of interfering devices. Similarly, the interference relationship derivation is not dependent on the networking scenario shown in fig. 2, so the interference coordination and resource allocation scheme proposed in the present application is also applicable to any wireless communication networking model.

To evaluate network performance, under the scenario of low QoS requirements and best effort service dominance, any device U is usually calculated by shannon's formula _m Instantaneous rate achievable at the kth RU:

wherein, W ^(k) Is the bandwidth of the kth RU.

In an actual wireless network, the bandwidth of each RU is typically the same, denoted as W. Therefore, excluding unpredictable small-scale fading, the present application estimates system performance using statistical SINR:

R _m ＝min{Wlog ₂ (1+ _m ), _m,req }

here the QoS demand rate is taken as an estimated rate cap to avoid allocating resources far in excess of its demand to the same device during the resource allocation process.

For scenarios with high QoS requirements, such as URLLC, reliability needs to be taken into account in particular. In this case, any device U _m The instantaneous rates that can be achieved on the kth RU are:

wherein the content of the first and second substances,

for channel dispersion, τ is the duration of the signal transmission, Q ^-1 (. Cndot.) is the inverse of the Gaussian Q function, and ε is the decoding error probability. In this case, it is not appropriate to estimate the system performance using statistical SINR, since this will result in real-time variation of the decoding error probability, making it unable to meet the QoS requirements. In order to solve the problems that performance cannot be estimated and resources are difficult to schedule due to small-scale fading, the scheme provided by the application makes additional limitation on the achievable rate to avoid performance fluctuation on the basis of utilizing statistical SINR, and a specific method is set forth in detail in the section of "small-scale fading compensation".

The wireless network interference coordination and resource scheduling method based on the hierarchical clustering algorithm comprises four parts of interference coordination, resource optimization, resource scheduling and small-scale fading compensation.

Based on the interference model provided by the modeling scheme for uplink interference among wireless network users, the interference coordination part of the scheme uses a hierarchical clustering algorithm to cluster the devices according to the interference strength relationship among the devices provided by the interference model, so that the devices with weak mutual interference can be divided into the same class and reuse the same resources, and the devices with strong mutual interference are divided into different classes and use orthogonal resources, thereby achieving the purposes of reducing network interference and reducing wireless resource occupation.

Because the hierarchical clustering algorithm belongs to a non-overlapping clustering algorithm, each device only belongs to one class in a classification result obtained by the interference coordination part of the scheme, namely, each device only multiplexes the same group of wireless resources with one group of devices. In practice, however, each device has the capability to use multiple sets of radio resources. Therefore, the resource optimization part of the scheme of the application is adjusted on the basis of the classification result obtained by the interference coordination part, so that one device can use multiple groups of different resources, and the resource utilization efficiency of the system is further improved.

After the interference coordination and the resource optimization are completed, the resource scheduling part of the scheme of the application performs actual resource scheduling according to the classification result adjusted by the resource optimization part, and the method comprises the following steps: and allocating resources for each type in the optimal clustering result among the devices after the resource optimization, and integrating the allocation result into a resource allocation result list. If the QoS of all the devices is met or the resources are completely exhausted after the allocation, the scheme finishes the operation and obtains a final resource allocation result; otherwise, repeating the above process for the device whose QoS requirement is not met until the scheme is finished to run.

In addition, in the case of a scenario with a high QoS requirement (such as URLLC), the solution described in the present application combines a small-scale fading compensation method to better guarantee the QoS requirement of each device.

It can be seen from the above description that the operation of the present solution does not depend on a specific networking scenario and a multiple access method, and is simultaneously adaptable to various scenarios and service models that are common in the present 5G network. Therefore, the scheme has very wide application prospect. The applicable scenarios include but are not limited to: resource scheduling of a public commercial 5G eMBB wireless network and authorization-free resource scheduling of an industrial 5G URLLC wireless network.

The method and the device for wireless network interference coordination and resource scheduling based on hierarchical clustering algorithm according to the embodiments of the present application are described below with reference to the accompanying drawings.

Fig. 3 is a schematic flowchart of a method for wireless network interference coordination and resource scheduling based on a hierarchical clustering algorithm according to an embodiment of the present application.

As shown in fig. 3, the method for wireless network interference coordination and resource scheduling based on hierarchical clustering algorithm includes the following steps:

step 301, acquiring a user equipment set with unsatisfied requirements;

step 302, classifying the devices with mutual interference smaller than a threshold value in the user equipment set into the same type according to a hierarchical clustering algorithm, and classifying the devices with mutual interference larger than the threshold value into different types to obtain an optimal non-overlapping clustering result;

303, adjusting and optimizing the optimal non-overlapping clustering result to enable different classes to overlap and obtain an optimal clustering result;

and 304, allocating wireless resource units for each class according to the optimal clustering result, and stopping allocation if the requirements of all equipment are met or the resources are exhausted to obtain a final resource allocation result.

Optionally, in an embodiment of the present application, the interference coordination portion uses a hierarchical clustering algorithm as a core, classifies user equipments with small mutual interference into the same class, and allows the same resource to be reused to improve the resource utilization rate; the user equipments with large mutual interference are classified into different classes, so that the user equipments occupy orthogonal resources to avoid that the strong interference seriously affects the system performance. Specifically, the hierarchical clustering algorithm used in the present application is an aggregation-type hierarchical clustering algorithm AHC, and since the BMT shown in fig. 1 is finally obtained by the AHC, the BMT may generate a plurality of groups of different clustering results according to different given heights, and the classifying the set of user equipments according to the hierarchical clustering algorithm includes:

screening the clustering result set according to the expected resource occupation amount and the achievable total rate of each clustering result to obtain an optimal non-overlapping clustering result, wherein the expected resource occupation amount is the sum of the expected maximum resource amount of all classes in the clustering result, the achievable total rate is the sum of the rates which can be achieved by all the classes in the clustering result, and the sum of the rates which can be achieved by each class is the sum of the rates which can be achieved by all the devices contained in the class;

taking the OFDMA wireless system based on the 3GPP dual-strip model densely deployed as shown in fig. 2 as an example of a typical system scenario, the method for determining and generating the optimal non-overlapping clustering result proposed in the present solution is summarized as follows:

step 1: the result of the initialization clustering is

Indicating variable i =1 and expected minimum resource occupation amount is E _min = + ∞, expected maximum total rate is R _max =0, optimal clustering result is ÷>

Corresponding height ^ of optimal clustering result>

Step 2: order to

The devices included in each class in (a) use the same resources, and different classes use different resources.

And 3, step 3: calculated according to the above-described rate estimation method in the wireless network interference and system model technique

The achievable rates of each device in each class. To facilitate description, make->

Then to any one class->

Each device estimates an achievable rate of { R } _m1 ,…,R _mK Get the resource amount needed by the class }>

The rate which it can achieve is->

Therefore, the clustering manner->

The resulting expected resource occupancy is ` based>

Can achieve a total rate of >>

And 4, step 4: if E<E _min Or E = E _min And R is>R _max Then update

E _min ＝E，R _max ＝R，/>

And 5, step 5: i = i +1. If i<N, and h _i <If + infinity, returning to the step 2; otherwise, then the flow is ended and the process is finished,

the obtained optimal clustering result is determined, and the height of the result is greater than or equal to>

Optionally, in an embodiment of the present application, in order to enable the AHC to better solve the interference coordination problem, the present solution needs to define a link function according to a characteristic of the interference coordination problem. Further, since the link function describes the difference between classes, and the difference between classes is determined by the difference between data points located between different classes, the present solution first defines a function describing the difference between data points (in this application, each data point corresponds to one device) -the distance function is:

i.e. if two devices U _p And U _q The greater the interference between them, the greater the distance between them. If the two devices cannot multiplex resources due to the limitation of the multiple access method used by the system, the distance between the two devices is set to be infinite, so as to ensure that the two devices are not classified into the same class.

From the definition of the distance function, naturally, the link function can be defined as:

this link function is also called Complete link (Complete link). When the complete link function is used, the construction of the BMT can be quickly completed through the CLINK algorithm.

Optionally, in an embodiment of the present application, the optimal non-overlapping clustering results are non-overlapping, that is, an intersection of any two classes in the clustering results is empty. However, in an actual wireless network, each device may occupy multiple resources simultaneously. Therefore, the optimal non-overlapping clustering result needs to be adjusted and optimized, and the device is allowed to occupy multiple resources at the same time, so as to further improve the resource utilization rate of the system. Specifically, the adjusting and optimizing of the optimal non-overlapping clustering result includes:

wherein the preset conditions are as follows: the intersection of the current updated optimal non-overlapping clustering result and the current class to be merged is empty, all devices in the merged clusters use the same resource, and the resource utilization rate after merging is greater than that before merging, wherein the resource utilization rate after merging is greater than that before merging, and the method comprises the following steps: the combined expected resource occupation is greater than the combined expected resource occupation or the combined expected resource occupation is equal to the combined achievable total rate before and after the combination;

taking the OFDMA wireless system with dense deployment and based on 3GPP dual-strip model shown in FIG. 2 as an example of a typical system scenario, the optimal non-overlapping clustering result is recorded as

The method for determining the optimal clustering result is summarized as follows:

step 1: sorting classes that may be used. The BMT constructed in the process of generating the optimal non-overlapping clustering result can be used for simplifying the flow of resource optimization. All heights in BMT are less than

Class (including height less than ÷ greater than +)>

Node and all leaves) can be merged into a reconciliation process>

Any class of objects. Let the set of these classes be S = { S = { (S) } ₁ ,…,S _|S| }. Each element in the set is subscripted in descending order of its corresponding height, i.e., the higher the height, the smaller the value of the subscript.

Step 2: initialization i =1,j =1.

And 3, step 3: according to the method used in the process of generating the optimal non-overlapped clustering result, calculating

Is expected to take up the resource>

And the rate that can be achieved>

And initializes an optimal expected resource occupancy->

Optimal achievable rate

Optimization of class->

And 4, step 4: try to make

And S of S _j Are combined to obtain->

If the following conditions are simultaneously met, for +>

In other words, S _j The most preferred terms for their incorporation:

1)

2)

all devices in (1) may use the same resource;

3) After combination, the resource utilization rate can be improved. I.e. calculated according to the method used in the process of generating the optimal non-overlapping clustering result

Is expected to take up the resource>

And an achievable rate->

Satisfy +>

Or at the same time satisfy

And/or>

And 5, step 5: if S _j Is composed of

The current optimal merge option, then update>

And 6, step 6: j = j +1. And if j is less than or equal to | S |, returning to the step 4. Otherwise, will

In>

Is replaced by>

And 7, step 7: i = i +1. If it is

And returning to the step 3. Otherwise, is taken off>

Namely the adjusted optimal clustering result.

Optionally, in an embodiment of the present application, allocating a radio resource unit to each class according to the optimal clustering result, and stopping allocation if the requirements of all the devices are met or the resources are exhausted, includes:

updating the rate requirement of each user equipment according to the reachable rate of each user equipment, and adding the currently finished resource allocation into a resource allocation result list;

if the updated rate requirements of all the user equipment are reduced to 0 or the resources are exhausted, scheduling the resources according to the current resource allocation result list, otherwise, re-executing the wireless network interference coordination and resource scheduling method based on the hierarchical clustering algorithm;

taking the OFDMA wireless system based on the 3GPP dual-strip model, which is densely deployed as shown in fig. 2, as an example of a typical system scenario, the resource scheduling process proposed in the present solution is summarized as follows:

step 1: for optimal clustering results

Is assigned 1 RU, in accordance with the wireless network interference and system model portion described aboveThe manner described estimates the achievable rate @ of each device after this round of assignment>

If a device exists in multiple classes, its estimated achievable rate is the sum of the estimated achievable rates of the device in the classes.

Step 2: the rate requirements of each device are updated. For device U _i With the rate requirement updated to

And adding the resource allocation completed in the current round into a resource allocation result list. If the rate requirements of all the devices are reduced to 0 after updating or the resources are exhausted, scheduling the resources according to the current resource allocation result list; otherwise, the wireless network interference coordination and resource scheduling method based on the hierarchical clustering algorithm is carried out again.

Optionally, in an embodiment of the present application, the method is mainly suitable for services such as eMBB with low QoS requirements. For the service types with higher QoS requirements such as URLLC, the performance fluctuation caused by small-scale fading cannot be compensated only by the above process, so that the QoS meeting condition of the services is influenced.

Therefore, depending on the characteristics of the small-scale fading, the present application will be directed to 2 small-scale fading compensation methods: the fading protection boundaries are reserved and the transmission is repeated a number of times, giving them the best configuration. Meanwhile, the present application will also provide a mechanism for combining the 2 methods with the wireless network interference coordination and resource scheduling scheme based on hierarchical clustering algorithm proposed by the present application, so as to further expand the applicability of the wireless network interference coordination and resource scheduling scheme based on hierarchical clustering algorithm proposed by the present application.

In the present application, the small-scale fading analysis is mainly performed for the Nakagami-m channel fading model. The Nakagami-m channel model has good representativeness because the model has been proved to have good fitting performance to the measured data in a plurality of wireless communication system measured experiments. The way to combine multiple repeated transmissions is the maximum-ratio combining (MRC), which is very common in today's wireless networks.

Fading protection boundaries

The reserved fading protection boundary is actually a way of reducing the amount of data transmitted by the user on the resource, so as to reduce the minimum SINR required for ensuring the decoding success, i.e. a certain boundary is reserved for the small-scale fading. In order to save resources while satisfying QoS, the size of the protection boundary needs to be decided reasonably. The expression for the optimum protection limit value will be given in this application.

In a Nakagami-m fading channel, all useful signals and interfering signals follow the Nakagami-m distribution within the same resource. Provided with useful signals

Wherein->

Is distributed as Nakagami-m, L is signal transmission frequency, m _x E [0.5, + ∞) is the shape factor of the Nakagami-m distribution, whose value depends on the channel characteristics, and the larger the value, the flatter the fading; omega _x Is the average power of the signal x.

Similarly, if there are J interference sources, then the jth interference signal

For the analysis convenience, J interference sources are combined into an equivalent interference y, then->

And has->

Thus, the decoding error probability can be expressed as

Wherein gamma isInstantaneous SINR->

The minimum SINR required to ensure reliability is obtained by dividing the statistical SINR calculated using the interference model by the guard margin G.

Note that densely deployed wireless networks are typically interference limited, i.e., the effect of noise is small, so the SINR can be well estimated from the SIR. Thereby having

For a statistical SIR of the current useful signal, it is apparent from the above derivation that there is->

Which can be obtained directly by the interference model. The decoding error probability is required to be not more than the element, and the expression of the protection boundary value can be obtained through derivation:

wherein B (·,. Cndot.) is a Beta function.

Similarly, if the current resource is used by only one device and there is no interference, the SINR is degenerated to SNR. At this time, the expression of the guard boundary value is:

obviously, the value of G is related to L. Therefore, the following note that the guard boundary value is G _L . After applying the fading protection boundary, any device U can be estimated using the following equation _m The achievable rate of (c):

the value of L may be arbitrarily specified when the method is used alone, or may be determined in conjunction with the method described in the multiple retransmission section below.

Multiple repeat transmission

Multiple repeated transmissions utilize diversity and combining to avoid system reliability degradation due to deep fading of the signal. If the number of transmissions is too small, it is difficult to ensure the reliability of the system, and if the number of transmissions is too large, the resource utilization efficiency of the system is affected. Therefore, the method and the device for determining the class under the premise of guaranteeing the QoS (quality of service) will be provided

Method for optimal transmission times.

Step 1: initializing the optimum number of transmissions L ^* ＝1，L＝1；

Step 2: calculate either device as follows

Estimated average rate per resource occupancy when L transmissions are employed:

in the same way, calculate any equipment

Estimating average rate per resource occupancy R when using L +1 transmissions _m,L+1 ；

And 3, step 3: the total achievable average rate for this class with L transmissions is found as follows:

similarly, the total achievable average rate for the class when L +1 transmissions are used can be calculated

And 4, step 4: the lower SINR limit is calculated using L transmissions better than L +1 transmissions by:

wherein the content of the first and second substances,

and 5, step 5: if the following 2 conditions are all satisfied:

①

②/>

then there is L ^* = L is the optimal transmission number of times sought; otherwise, L = L +1, and the step 2 is returned.

The 2 small-scale fading compensation methods referred to in this application can be applied separately, but in this application, both will be applied to the aforementioned scheme at the same time, which will affect the estimation values of the aforementioned scheme on the rate and the resource occupation amount, in order to achieve better effect.

Specifically, after applying the small-scale fading compensation mechanism, the above-mentioned wireless network interference coordination and resource scheduling scheme based on hierarchical clustering algorithm needs to be changed as follows:

an interference coordination part:

in the step 3 of generating the optimal non-overlapping clustering result, the expression given by the fading protection boundary part is used to estimate any device U _mk Rate R of _mk ；

In step 3, generating the optimal non-overlapping clustering result, estimating any class

When the resource occupation amount is increased, the method of the repeated transmission part is firstly used for determining the optimal transmission times L ^* And estimating the required resource quantity of the class by using the following formula:

and a resource optimization part:

following the change of the interference coordination part, the method for estimating the reachable rate and the resource occupation amount in the 3 rd step and the 4 th step of the part synchronously carries out the same change;

and a resource scheduling part:

in step 3 of the part, pair

Is selected based on any one of the classes>

The number of RUs allocated to it is changed from 1 to L ^* Wherein L is ^* For determined for a method using the above-described multiple repeat transmission section>

The optimal number of transmissions;

in step 3 of this section, the device achievable rate is estimated by changing to the expression given by the fading protection boundary section.

The performance of the wireless network interference coordination and resource scheduling scheme (hereinafter abbreviated as "HCSA") based on hierarchical clustering algorithm proposed in the present application will be described below. In this section, existing SGRA, stM, TSDR are used as comparison schemes to demonstrate the performance advantage of the HCSA scheme proposed in this application over the existing closest and most effective scheme.

(1) Setting simulation parameters and scenes.

In order to fully show the performance of the HCSA scheme proposed in the present application, a URLLC scenario is adopted for simulation. 2 systems of different sizes were considered in the simulation:

1) Small-scale: the system has 16 cells, and each cell has 3 devices and 48 devices;

2) Large scale: the system has 40 cells, and each cell has 6 devices and 240 devices.

The delay limit of each device is 1ms, and 700 distributable RUs are provided in the corresponding system. And each device has a transmission rate requirement of 288 to 720kbps. The following table summarizes the detailed parameters used in the simulation.

TABLE 1 simulation parameter Table

(2) The performance of the scheme is proposed without applying a small-scale fading compensation mechanism.

In this subsection, small-scale fading is not considered to evaluate the performance of the HCSA scheme itself without applying a small-scale fading compensation mechanism. This applies to most situations in wireless networks where average performance is required, rather than instantaneous performance. In order to evaluate the influence of the interference relationship predictability on the performance of the resource allocation algorithm, the SGRA algorithm is also simulated in this section for the cases where the interference relationship is known and unknown, respectively.

Fig. 4 is a performance diagram of a proposed scheme and a comparative scheme in a small-scale network, and it is apparent from fig. 4 that the HCSA scheme proposed in the present application can save certain resource consumption in the small-scale network compared with the SGRA scheme with a known interference relationship, and in the case of different device rate requirements, the HCSA scheme can save 14.99% on average, and the maximum saving is 19.5%. For the StM scheme, the HCSA scheme proposed by the present application can save 49.76% of resources on average, and can save 56.58% at most.

Furthermore, the perceptibility of interference relationships can also greatly affect the performance of the algorithm. Compared with the situation that the interference relationship cannot be obtained, the SGRA algorithm can save 10-30% of resource occupation by means of the interference relationship provided by the interference model, and the difference is obvious.

Fig. 5 is a performance diagram of the proposed solution and the comparative solution in a large-scale network, in which the amount of resources that can be saved by the HCSA solution proposed by the present application is more considerable. As can be seen from fig. 5, compared to the SGRA scheme when the interference relationship is known, the HCSA scheme can save 23.24% and at most 31.5% of resources occupation on average under different device rate requirement situations. Compared with the StM scheme, the method can save resources by 56.19% on average.

The excellent performance of the HCSA scheme proposed in the present application can be significantly demonstrated from fig. 4 and 5.

(3) When a small-scale fading compensation mechanism is applied, the performance of the scheme is provided.

The part carries out simulation by taking small-scale fading into consideration so as to evaluate the performance of the HCSA scheme which can be achieved by combining with a small-scale fading compensation mechanism under the URLLC scene with the most severe requirement.

In this section, a small scale network is employed, with a rate requirement R _REQ Simulations were performed for a 288kbps scene.

Fig. 6 and fig. 7 show the decoding error probability generated by the proposed scheme and the comparison scheme under different types of average transmission times under different statistical calibers. Where fig. 6 shows the decoding error probability when only the multiplexed RUs are statistically multiplexed, and fig. 7 shows the decoding error probability when all RUs are statistically multiplexed. As can be seen from the great similarity of the two figures, decoding errors occur mainly on the multiplexed RUs.

As can be seen from fig. 6 and 7, the HCSA proposed by the present application is greatly improved compared with all the comparative schemes. Under the condition of also meeting the decoding error probability limit, the average transmission times required by various types obtained by the HCSA scheme are less than 1.5 times, and the resource saving effect is very obvious compared with more than 4 times required by the SGRA scheme introducing a 3dB protection boundary, wherein the average transmission times are more than 60%.

In addition, in order to demonstrate the effectiveness of the small-scale fading compensation method referred to in this application as "reserved guard boundaries", this section also employs the SGRA scheme for comparison. Under the condition that the interference relationship can be known, the decoding error probability of at least 1 magnitude can be reduced by introducing a 3dB protection boundary into the SGRA scheme, and the effectiveness of the reserved protection boundary on guaranteeing QoS is fully shown.

Fig. 8 shows the decoding error probability generated by the proposed scheme and the comparative scheme under different RU occupation amounts when counting all RUs, and fig. 8 shows the relationship between RU occupation amounts and decoding error probabilities of each scheme more directly. By combining a small-scale fading compensation mechanism, compared with an SGRA (generalized block random access) scheme introducing a 3dB protection boundary with optimal performance in a contrast scheme, the HCSA scheme provided by the application can use less resources by more than 55% under the condition of also meeting the limit condition of decoding error probability.

Undoubtedly, the HCSA scheme proposed by the present application exhibits excellent adaptability to the small-scale fading compensation mechanism, and combines the advantages of the HCSA scheme itself, so as to achieve superior performance in this section, and clearly demonstrate the advantages of the scheme proposed by the present application.

In order to implement the above embodiments, the present application further provides a wireless network interference coordination and resource scheduling device based on hierarchical clustering algorithm.

As shown in fig. 9, the wireless network interference coordination and resource scheduling apparatus based on hierarchical clustering algorithm includes an obtaining module, a clustering optimization module, and a resource allocation module, wherein,

the clustering module is used for classifying the devices with mutual interference smaller than a threshold value in the user equipment set into the same class and classifying the devices with mutual interference larger than the threshold value into different classes according to a hierarchical clustering algorithm to obtain an optimal non-overlapping clustering result;

the clustering optimization module is used for adjusting and optimizing the optimal non-overlapping clustering result to ensure that different classes are overlapped to obtain the optimal clustering result;

wherein the preset conditions are as follows: the intersection of the current updated optimal non-overlapping clustering result and the current class to be merged is empty, all devices in the merged clusters use the same resource, and the resource utilization rate after merging is greater than that before merging, wherein the resource utilization rate after merging is greater than that before merging, and the method comprises the following steps: the combined expected resource occupation amount is larger than that before combination or the combined expected resource occupation amount is equal to that before combination and the achievable total rate after simultaneous combination is larger than that before combination.

It should be noted that the explanation of the embodiment of the method for coordinating wireless network interference and scheduling resources based on the hierarchical clustering algorithm is also applicable to the device for coordinating wireless network interference and scheduling resources based on the hierarchical clustering algorithm in this embodiment, and is not described herein again.

In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example" or "some examples" or the like are intended to mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Moreover, various embodiments or examples and features of various embodiments or examples described in this specification can be combined and combined by one skilled in the art without being mutually inconsistent.

Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present application, "plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.

Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing steps of a custom logic function or process, and alternate implementations are included within the scope of the preferred embodiment of the present application in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present application.

The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.

It should be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.

It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.

In addition, functional units in the embodiments of the present application may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.

The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc. Although embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application, and that variations, modifications, substitutions and alterations may be made to the above embodiments by those of ordinary skill in the art within the scope of the present application.

Claims

1. A wireless network interference coordination and resource scheduling method based on a hierarchical clustering algorithm is characterized by comprising the following steps:

acquiring a user equipment set with unsatisfied requirements;

classifying the devices with mutual interference smaller than a threshold value in the user equipment set into the same class according to a hierarchical clustering algorithm, classifying the devices with mutual interference larger than the threshold value into different classes, and obtaining an optimal non-overlapping clustering result;

adjusting and optimizing the optimal non-overlapping clustering result to enable different classes to overlap and obtain an optimal clustering result;

and allocating wireless resource units for each type according to the optimal clustering result, and stopping allocation if the requirements of all the devices are met or the resources are exhausted to obtain a final resource allocation result.

2. The method of claim 1, wherein the hierarchical clustering algorithm is an aggregation-type hierarchical clustering algorithm, and wherein classifying the set of user devices according to the hierarchical clustering algorithm comprises:

clustering the user equipment set according to the difference between classes through the aggregation type hierarchical clustering algorithm to obtain a binary convergence tree, and setting different heights for the binary convergence tree to obtain a clustering result set, wherein the difference between the classes is determined by a link function, the equipment in each class contained in each clustering result uses the same resource, and different resources are used for different classes;

and screening the clustering result set according to the expected resource occupation amount and the achievable total rate of each clustering result to obtain the optimal non-overlapping clustering result, wherein the expected resource occupation amount is the sum of the maximum resource amounts expected and needed by all the classes in the clustering result, the achievable total rate is the sum of the rates which can be achieved by all the classes in the clustering result, and the sum of the rates which can be achieved by each class is the sum of the rates which can be achieved by all the devices contained in the class.

3. The method of claim 2, wherein the link function is determined by:

defining a distance function determining differences between user equipments, and determining the link function from the distance function,

wherein the distance function is represented as:

the link function is represented as:

wherein the content of the first and second substances,

Presentation device U _q As signalling means, slave unit U _p Inverse signal-to-interference ratio, j, of interference _p Presentation device U _p Associated base station number j _q Presentation device U _q Associated base station number j _q Presentation device U _q Associated base station number, <' > in>

Represents the mth class, </or in the clustering result>

Indicates the nth class in the clustering result, d (p, q) represents the difference between device p and device q.

4. The method of claim 2, wherein the adjusting and optimizing the optimal non-overlapping clustering result comprises:

continuously updating the optimal non-overlapping clustering result until all the classes in the class set to be merged are judged, and obtaining the final optimal non-overlapping clustering result as the optimal clustering result;

wherein the preset conditions are as follows: the intersection of the current updated optimal non-overlapping clustering result and the current class to be merged is empty, all devices in the merged cluster use the same resource, and the resource utilization rate after merging is greater than that before merging, wherein the resource utilization rate after merging is greater than that before merging, and the method comprises the following steps: the combined expected resource occupancy is greater than the combined expected resource occupancy or the combined expected resource occupancy is equal to the combined achievable total rate before and after the combination.

5. The method of claim 4, wherein the allocating radio resource units for each class according to the optimal clustering result, and stopping allocation if the demands of all devices are met or the resources are exhausted, comprises:

allocating 1 radio resource unit to each class in the optimal clustering result, and calculating the reachable rate of each user equipment after allocation, wherein if the user equipment exists in a plurality of classes, the reachable rate is the sum of the reachable rates of the user equipment in each class;

6. The method of claim 5, wherein if the set of user equipment has traffic with a QoS requirement greater than a threshold, applying a small-scale fading compensation mechanism to the wireless network interference coordination and resource scheduling method based on the hierarchical clustering algorithm, including:

when the hierarchical clustering algorithm is used for classifying the user equipment set and adjusting and optimizing the optimal non-overlapping clustering result, the sum of the reachable rates of all user equipment to which a fading protection boundary is applied is used as the reachable total rate, the optimal transmission times are determined, and the product of the optimal transmission times and the expected maximum required resource amount is used as the expected resource occupation amount;

and when allocating the radio resource units for each class according to the optimal clustering result, changing the allocated radio resource units from 1 to optimal transmission times, and taking the reachable rate of the user equipment after the application of the fading protection boundary as the reachable rate of the user equipment.

7. A wireless network interference coordination and resource scheduling device based on a hierarchical clustering algorithm is characterized by comprising an acquisition module, a clustering optimization module and a resource allocation module, wherein,

the acquisition module is used for acquiring a user equipment set with an unsatisfied demand;

the clustering module is used for classifying the devices with mutual interference smaller than a threshold value in the user equipment set into the same class according to a hierarchical clustering algorithm, classifying the devices with mutual interference larger than the threshold value into different classes, and obtaining an optimal non-overlapping clustering result;

the cluster optimization module is used for adjusting and optimizing the optimal non-overlapping cluster result to enable different clusters to overlap and obtain an optimal cluster result;

8. The apparatus of claim 7, wherein the hierarchical clustering algorithm is an agglomerative hierarchical clustering algorithm, and the clustering module is specifically configured to:

9. The apparatus of claim 8, wherein the cluster optimization module is specifically configured to:

10. The apparatus of claim 9, wherein the resource allocation module is specifically configured to:

allocating 1 wireless resource unit to each class in the optimal clustering result, and calculating the reachable rate of each allocated user equipment, wherein if the user equipment exists in a plurality of classes, the reachable rate of the user equipment is the sum of the reachable rates of the user equipment in each class;