CN106304307A - A kind of resource allocation methods under heterogeneous network converged - Google Patents

Abstract

Resource allocation methods under a kind of heterogeneous network converged, including step 1: initializing systematic parameter, be averagely allocated to each RAN including the UE being provided with line return bandwidth and overlapping region, the radio channels distributing to each RAN is 0；Step 2: determine the allocation strategy of first channel resource, increases a wireless channel to each RAN, solves the time delay accessed about wireless channel, wired bandwidth, user and reduces problem, obtains RAN subscript i that this channel is corresponding^*, and by this radio channel allocation to i^*Number RAN；Step 3: determine the allocation strategy of remaining wireless channel, according to step 2, obtains the RAN that each wireless channel is corresponding, and these channels is distributed to the RAN of correspondence；Step 4: update systematic parameter, wireless channel distributes to each RAN according to the optimal distribution strategy that step 2 obtains, and the UE of overlapping region is averagely allocated to each RAN, and the wired bandwidth unit number distributing to each RAN is 0.

Description

A kind of resource allocation methods under heterogeneous network converged

Technical field

The invention belongs to communication technical field, particularly to the resource allocation methods under a kind of heterogeneous network converged.

Background technology

Along with quickly increasing and the fast development of mobile communication and information technology, multiple wireless access skill of business demand Art is arisen at the historic moment.Most mobile terminal is equipped with multiple wave point to access different wireless access networks (RAN), example Such as Cellular Networks, Wi-Fi, bluetooth etc..HetNet has become as the trend of future wireless network development.Different wireless networks is connecing Entering the aspects such as technology, terminal, business and all have isomerism, this adds difficulty to the fusion between heterogeneous networks.In order to complete Supporting that under HetNet magnanimity low power consuming devices accesses, and expands coverage rate, improve reliability, reduce the targets such as delay, design is many Net merge under resource allocation policy and optimized algorithm just become the focus of research.

Wherein, add caching technology and can be obviously improved the overall performance of heterogeneous network.In all of service traffics content In, the content that only sub-fraction is popular can be the most accessed by the user.Therefore caching some popular contents at RAN can To substantially reduce the access of the redundancy to core net and some identical contents repeated transmission.

The access mechanism of heterogeneous wireless network is studied by existing research worker.The fusion of heterogeneous network is built by they Found into a multiobject decision model.Consider in terms of network, business, user three, optimization aim is created as one Utility function, purpose is then so that the utility function value of whole network is maximum.But this technical sophistication degree is higher, calculate power consumption relatively Greatly, and when relating to concrete decision factor such as network delay, network reliability, user-friendliness, reality is not provided Formula, but rule of thumb oneself definition.In the case of cost is limited, this brings the most negative obviously can to the expense of user Load.

English abbreviations and the Key Term that the present invention relates to include:

HetNet (heterogeneous network) heterogeneous network,

RAN (radio access network) wireless access network,

SDN (software defined network) software defined network,

UE (user end) user terminal,

AS (application server) application server,

BS (base station) base station.

Summary of the invention

The present invention proposes the resource allocation methods under a kind of heterogeneous network converged, is a kind of efficient with caching capabilities Isomery coalescing network system under to user's end-to-end time delay optimize algorithm.For under isomery UNE, prior art Algorithm in the selection and resource allocation policy of access mechanism is the most complicated, proposes under multi-internet integration model, with network End-to-end time delay is the resource allocation algorithm of optimization aim.

The technical scheme is that, the resource allocation methods under a kind of heterogeneous network converged there is heterogeneous wireless and access Network RAN1, RAN2 ... RANK, UE access application server by wireless link and wired backhaul, and the BS of different RAN is Have caching capabilities, be owned by the spatial cache of formed objects, and all store the content that the same popularity is the highest, including step Rapid:

Step 1: initialize systematic parameter, be averagely allocated to each including the UE being provided with line return bandwidth and overlapping region RAN, the radio channels distributing to each RAN is 0；

Step 2: determine the allocation strategy of first channel resource, to each RAN increase a wireless channel, solve about The time delay that wireless channel, wired bandwidth, user access reduces problem, obtains RAN subscript i that this channel is corresponding^*, and this is wireless Channel distributes to i^*Number RAN；

Step 3: determine the allocation strategy of remaining wireless channel, according to step 2, obtains each wireless channel corresponding RAN, and these channels are distributed to the RAN of correspondence；

Step 4: update systematic parameter, wireless channel distributes to each RAN according to the optimal distribution strategy that step 2 obtains, The UE of overlapping region is averagely allocated to each RAN, and the wired bandwidth unit number distributing to each RAN is 0；

Step 5: determine the allocation strategy of each wired bandwidth resources quantization unit, it is assumed that increase a list to each RAN Wired bandwidth of position, solves the time delay accessed about wireless channel, wired bandwidth, user and reduces problem, and obtain this unit has RAN subscript i that tape width is corresponding^*, and wired bandwidth of this unit is distributed to i^*Number RAN；

Step 6: determine the allocation strategy of all wired bandwidth unit, to remaining wired bandwidth unit, repeats step 5, Obtain the RAN that wired bandwidth of each unit is corresponding, and wired bandwidth of these units is distributed to the RAN of correspondence；

Step 7: update systematic parameter, wireless channel distributes to each RAN according to the optimal distribution strategy that step 2 obtains, Wired bandwidth unit distributes to each RAN according to the optimal distribution strategy that step 5 obtains, it is assumed that the UE of each RAN overlapping region Number is 0；

Step 8: the UE to each overlapping region, calculates the average delay being connected under its each RAN that can connect, Choosing the minimum RAN of time delay is its RAN finally connected.

Described step 2 includes: according to formulaObtain the RAN of correspondence Subscript i^*。

Described step 5 includes: according to formulaObtain correspondence RAN subscript i^*。

The present invention dynamically can connect according to the real-time performance of network for being in the overlapping covered interior user of many nets Enter different RAN.Wired bandwidth, radio spectrum resources can be dynamically assigned to different RAN, to reach different real-time network Performance requirement.

The invention provides a kind of the building of user's end-to-end time delay under the isomery coalescing network system with caching capabilities Mode division analysis method, and provide the resource allocation optimization algorithm minimized it；The time delay optimization algorithm that the present invention proposes is relative Little in force search Algorithms T-cbmplexity, and have carrying of significant network performance relative to other resource allocation algorithms Rise；The caching technology introduced for heterogeneous network system is obvious for the lifting of delay performance.The algorithm that the present invention proposes is complicated Spend relatively low, calculate power consumption less, implement facilitate understandable.

Accompanying drawing explanation

The different radio access network schematic diagram of covering area overlapping in Fig. 1 prior art.

The heterogeneous wireless network schematic diagram of universal meaning in Fig. 2 prior art.

The resource allocation optimization algorithm FB(flow block) based on isomery UNE that Fig. 3 present invention provides.

Detailed description of the invention

Below in conjunction with specific embodiment, the present invention is described in detail.Following example will assist in the technology of this area Personnel are further appreciated by the present invention, but limit the present invention the most in any form.It should be pointed out that, the ordinary skill to this area For personnel, without departing from the inventive concept of the premise, it is also possible to make some deformation and improvement.These broadly fall into the present invention Protection domain.

Heterogeneous wireless network system is collectively formed by multiple heterogeneous network element, such as macrocell, Microcell, slightly Community and Home eNodeB etc..They through-put power, overlay area, the backbone network of connection and transmission characteristics each other are all Likely different.They have employed different Radio Transmission Technologys.If two wireless access systems have employed identical wireless Transmission technology is still belonging respectively to different operators, it is also possible to be considered as isomery.Fig. 1 may be considered certain integration of three networks The system model of heterogeneous wireless network.Different RAN has the base station of different coverages and different structure, and they are by each From backhaul network be connected to same core backbone.

In heterogeneous network, when a user is covered by multiple Radio Access Network simultaneously, he can be according to current The factor such as network state and business demand, changes the selection of oneself at any time.User can access or be switched to be best suitable at any time In certain wireless access system that oneself or Consumer's Experience are best, to meet the personalized communication service of oneself flexible transformation Demand.Can also cooperate with each other between each wireless access system, the resource such as wireless frequency spectrum that reasonable distribution is limited so that whole net The performance of network system reaches optimal.

Under isomery UNE, a UE may be covered by multiple different RAN simultaneously.It is different from traditional mechanism, UE in overlapping region can be linked into any one RAN covering it neatly.As it is shown in figure 1, UE1 and UE2 can select Being connected to any one in three kinds of RAN, UE3 and UE4 only has access RAN1 and RAN2, and UE5 only has access RAN3.Separately Outward, radio spectrum resources total in whole system is bHZ, and total wired bandwidth resources are Bbps.These resources are dynamically distributed To different RAN.

For above-mentioned model, the resource allocation optimization algorithm that the present invention uses comprises the steps:

Step one: distribute wireless channel to each RAN.Assume wire-line back haul bandwidth resources and overlapping region UE mean allocation To each RAN, and the wireless channel number initializing each RAN is 0.During first channel of distribution, it is assumed that increase to each RAN One channel, calculates the decrement of UE average delay under the most each RAN, and the RAN distribution choosing decrement most obtains one Channel.For remaining channel, by that analogy, said process is repeated.

Step 2: be assigned tape width to each RAN.Radio channel resource is distributed to according to step one gained allocative decision Each RAN, it is assumed that overlapping region UE is averagely allocated to each RAN, and the wired bandwidth number initializing each RAN is 0.Distribution the During one unit bandwidth, it is assumed that the bandwidth increasing a unit to each RAN, calculate UE average delay under the most each RAN Decrement, chooses the most RAN distribution of decrement and obtains the bandwidth of a unit.For remaining unit bandwidth, by that analogy, Repeat said process.

Step 3: overlapping region UE accesses RAN process.Radio channel resource and wired bandwidth resources are according to step one and step Rapid two gained allocative decisions distribute to each RAN, and the UE number being connected to each RAN in initializing overlapping region is 0.To each The UE of overlapping region, calculates the average delay being connected under its each RAN that can connect, and the RAN choosing time delay minimum is Its RAN finally connected.

Consider as Fig. 2 comprises K layer heterogeneous radio access networks network RAN1, RAN2 ... the hybrid system of RANK.UE passes through Wireless link and wired backhaul access the content in application server (ASs, ApplicationServers).I-th layer of nothing The set expression of the user terminal that line access network covers is M_i, i=1,2 ..., K, the user terminal number in set is designated as | M_i |.As it is shown in figure 1, the overlay area of different Radio Access Networks can be overlapping, and assumeRepresent whole system User terminal sum in system.

The BS of all different RAN is owned by caching capabilities, is owned by the spatial cache of formed objects, and as all storing The highest content of popularity.The popularity of content presses certain regular distribution.

In order to closer to practical situation, it is believed that the time that user load is minimum in the middle of one day, be in general exactly At midnight, controller can update once their popularity according to the use frequency of different pieces of information on the same day, then new popularity Those high contents are broadcast to all of BS.When a UE request service, if the content of its request is at caching local for BS Space is exist, then corresponding content is handed down to UE by base station immediately；If the content of request is not deposited in spatial cache , then BS can first pass through wired backhaul and take out corresponding content at AS, then these contents is handed down to UE again.

Based on model above, the present invention is modeled as the queuing process of a M/M/c/m type, obtains the flat of all users All end-to-end time delay, to each RAN limited resource of distribution and determine that user accesses criterion with its minimum target.

When calculating end-to-end time delay, ignore the time and coded modulation expended in up-link and read from caching The time of data.Being limited to limited radio channel resource and wire-line back haul bandwidth, and synthetic user selection strategy, isomery merges The optimization problem of network end-to-end time delay can be expressed as:

$\begin{array}{cc}\underset{\stackrel{\→}{B},\stackrel{\→}{b},\stackrel{\→}{m}}{\min }& T=\frac{\underset{i=1}{\overset{K}{\Σ}}{t}_i\left|m_i\right|}{\left|M\right|}\end{array}$

$\begin{array}{ccc}subject& to& \underset{i=i}{\overset{K}{\Σ}}\left|m_i\right|=\left|M\right|\end{array}$

$\underset{i=1}{\overset{K}{\Σ}}{B}_i\≤B$

$\underset{i=1}{\overset{K}{\Σ}}{b}_i\≤b$

Wherein K represents total number of RAN,

For i-th RAN The average delay of lower user, | m_i| representing the number of users under i-th RAN, | M | represents total number of users, λ_iRepresent under i-th RAN Average user request speed, μ_iRepresenting the service speed of each channel under i-th RAN, it is with wire-line back haul bandwidth and slow Deposit space size relevant, be specifically represented by:Wherein N_iCaching for i-th RAN is empty Between size, Z_jFor the popularity of corresponding cache contents j, α_iAnd β_iIt is that wired bandwidth of i-th RAN and caching capabilities are to it respectively The weighter factor of service speed impact.B represents available wire-line back haul bandwidth, B_iRepresent the wire-line back haul distributing to i-th RAN Bandwidth, b represents total number of available channels, b_iRepresent the number of channel distributing to i-th RAN.

WithRepresent the allocation strategy of wired bandwidth and wireless frequency spectrum respectively,Represent that user accesses selection strategy.Wireless frequency spectrum is quantified as transmission channel number, constraintsCan be equivalent toWherein c is the number of available channels quantified, c_iRepresent the channel distributing to i-th RAN Number.Correspondingly, when asking for optimal resource allocation strategy and user accesses criterion, wire-line back haul bandwidth is also quantified as integer Unit bandwidth resource.

By solving above-mentioned optimization problem, the optimal strategy of resource distribution, Yi Jiyong under isomery UNE can be obtained The optiaml ciriterion selected is accessed at family.

For above-mentioned optimization problem, the present invention proposes a kind of associating greedy algorithm, has obtained a kind of optimum of this problem Majorization of solutions algorithm.Specific algorithm comprises the steps:

Step 1: initialize systematic parameter.Assume that the UE of wire-line back haul bandwidth and overlapping region is averagely allocated to each RAN, The radio channels distributing to each RAN is 0.

Step 2: determine the allocation strategy of each channel resource.Assume to increase a wireless channel to each RAN, solve pass The time delay accessed in wireless channel, wired bandwidth, user reduces problem, obtains RAN subscript i that this channel is corresponding^*, and by this nothing Line channel distributes to i^*Number RAN.

Step 3: determine the allocation strategy of all wireless channels.To remaining wireless channel, repeat step 2, obtain each The RAN that wireless channel is corresponding, and these channels are distributed to the RAN of correspondence.

Step 4: update systematic parameter.The optimal distribution strategy that wireless channel obtains according to above-mentioned steps distributes to each RAN, the UE of overlapping region are averagely allocated to each RAN, and the wired bandwidth unit number distributing to each RAN is 0.

Step 5: determine the allocation strategy of wired bandwidth resources of each unit.Assume to increase a unit to each RAN Wired bandwidth, solve about wireless channel, wired bandwidth, user access time delay reduce problem, obtain the wired of this unit RAN subscript i that bandwidth is corresponding^*, and wired bandwidth of this unit is distributed to i^*Number RAN.

Step 6: determine the allocation strategy of all wired bandwidth unit.To remaining wired bandwidth unit, repeat step 5, Obtain the RAN that wired bandwidth of each unit is corresponding, and wired bandwidth of these units is distributed to the RAN of correspondence.

Step 7: update systematic parameter.The optimal distribution strategy that wireless channel obtains according to above-mentioned steps distributes to each RAN, wired bandwidth unit distributes to each RAN according to the optimal distribution strategy that above-mentioned steps obtains.Assume each RAN overlay region The UE number in territory is 0.

Step 8: the UE to each overlapping region, calculates the average delay being connected under its each RAN that can connect, Choosing the minimum RAN of time delay is its RAN finally connected.

Preferably, described step 2 includes: according to formulaIt is right to obtain RAN subscript i answered^*。

Preferably, described step 5 includes: according to formulaIt is right to obtain RAN subscript i answered^*。

Can be obtained under isomery UNE by optimization algorithm above to minimize the time delay wire-line back haul money as target Source and the optimal case of radio channel resource distribution, and user load access criterion.

Claims (3)

1. the resource allocation methods under heterogeneous network converged, has heterogeneous radio access networks network RAN1, RAN2 ... RANK, UE accesses application server by wireless link and wired backhaul, and the BS of different RAN is owned by caching capabilities, is owned by phase With the spatial cache of size, and all store the content that the same popularity is the highest, including step:

Step 1: initialize systematic parameter, be averagely allocated to each RAN including the UE being provided with line return bandwidth and overlapping region, point The radio channels of each RAN of dispensing is 0；

Step 2: determine the allocation strategy of first channel resource, increases a wireless channel to each RAN, solves about wireless The time delay that channel, wired bandwidth, user access reduces problem, obtains RAN subscript i that this channel is corresponding^*, and by this wireless channel Distribute to i^*Number RAN；

Step 3: determine the allocation strategy of remaining wireless channel, according to step 2, obtain the RAN that each wireless channel is corresponding, And these channels are distributed to the RAN of correspondence；

Step 4: update systematic parameter, wireless channel distributes to each RAN according to the optimal distribution strategy that step 2 obtains, overlapping The UE in region is averagely allocated to each RAN, and the wired bandwidth unit number distributing to each RAN is 0；

Step 5: determine the allocation strategy of each cable broadband resource quantization unit, it is assumed that increase a unit to each RAN Wired bandwidth, solves the time delay accessed about wireless channel, wired bandwidth, user and reduces problem, and obtain this unit has tape Wide corresponding RAN subscript i^*, and wired bandwidth of this unit is distributed to i^*Number RAN；

Step 6: determine the allocation strategy of all wired bandwidth unit, to remaining wired bandwidth unit, repeats step 5, obtains The RAN that wired bandwidth of each unit is corresponding, and wired bandwidth of these units is distributed to the RAN of correspondence；

Step 7: updating systematic parameter, wireless channel distributes to each RAN according to the optimal distribution strategy that step 2 obtains, wired Bandwidth unit distributes to each RAN according to the optimal distribution strategy that step 5 obtains, it is assumed that the UE number of each RAN overlapping region is 0；

Step 8: the UE to each overlapping region, calculates the average delay being connected under its each RAN that can connect, chooses RAN that RAN is its final connection that time delay is minimum.

2. the resource allocation methods under heterogeneous network converged as claimed in claim 1, it is characterised in that described step 2 includes: root According to formulaObtain RAN subscript i of correspondence^*。

3. the resource allocation methods under heterogeneous network converged as claimed in claim 1, it is characterised in that described step 5 includes: root According to formulaObtain RAN subscript i of correspondence^*。