CN104640227A - Downlink resource scheduling method for hybrid service in LTE (long term evolution) system - Google Patents

Abstract

The invention discloses a kind of downlink resource scheduling methods of mixed service in LTE system, i.e. blended service exponential factor (HSEF, Hybrid Service Exponential Factor) algorithm. This dispatching method is the following steps are included: the information reported in base station side according to base station bottom, user is calculated in effective signal-to-noise ratio SINRi, the j of each resource block, then according to shannon formula calculate user this when the instantaneous transmission speed ri that engraves, j, and acquire the instantaneous spectrum factor According to instantaneous transmission speed ri, the j currently obtained, user's average throughput is calculated And acquire the average throughput factor According to the type of service of each service and mutually it should ensure that bit rate (GBR, Guranteed Bit Rate) requires calculating business weight factor Wi, j; User's dispatch weight ω i, j are finally calculated according to aforementioned obtained each factor, system then carries out scheduling of resource according to the dispatch weight of user. Present invention improves the performances of fairness between user and ensures bit rate business, to improve LTE system overall performance.

Description

The downlink resource scheduling method of mixed service in LTE system

Technical field

The present invention relates to mobile communication technology field, be specifically related to scheduling downlink resource technical field in LTE (Long Term Evolution) system.

Background technology

In recent years, along with the fast development of hybrid multimedia business in the Internet access demand and wireless communication technology widely, there is explosive growth in wireless cellular network technology and market.The mixed service scheduling downlink resource algorithm of next generation mobile communication system LTE in many service scenarios effectively can promote real-time service performance, meet user to the demand ensureing bit rate service in the mixed service demand, particularly real time business under high bandwidth, high mobile context.

The key meeting the service quality QoS demand of the real time business such as data, video and non-real-time service is that wireless communication system has Resourse Distribute and scheduling by service priority algorithm flexibly and effectively.Dispatching algorithm has the status of core domination in the resource management of wireless communication system, dispatching method will consider the content of two aspects: the complexity that method realizes and the impact on system performance index, as QoSs such as fairness, time delay, throughputs.At present, main wireless packet scheduling algorithm can be divided into two classes, one class considers the packet scheduling algorithm of physical layer (PHY-layer) channel quality, as max carrier to interference (Max C/I) dispatching algorithm, poll (RR, Round Robin) dispatching algorithm and equitable proportion (PF, Proportional Fairness) dispatching algorithm; Another kind of is the cross-layer packet scheduling algorithm combining physical layer channel quality and media access control layer (MAC-layer) buffer queue, as M-LWDF (Modified Largest Weighted Delay First) dispatching algorithm and EXP dispatching algorithm.Last class algorithm does not consider the QoS demand of customer service, makes it not be well positioned to meet the performance requirement of real time business; Although then a class algorithm meets the QoS demand of real time business, do not consider that fairness between different business and different business are applied to the systematic function in mixing scene.

Summary of the invention

The present invention to overcome prior art exist the problems referred to above, the downlink resource scheduling method of mixed service in a kind of LTE system is proposed, can distinguish different brackets business, and make its dispatching priority weight change with its type of service and change, in mixed service scene, promote the performance ensureing bit rate service, meet present stage user for the demand ensureing bit rate service in real time business.

For achieving the above object, the present invention takes following technical scheme:

The downlink resource scheduling method of mixed service in LTE system, i.e. blended service exponential factor (HSEF, Hybrid Service Exponential Factor) algorithm, performing step is as follows:

Step 1, according to the effective signal-to-noise ratio SINR of calculated signals user on each Resource Block (RB, Resource Block) that user receives _i,j, and according to shannon formula calculate user this time the instantaneous transmission speed r that engraves _i,j, calculate the instantaneous spectrum factor

Step 2, according to the instantaneous transmission speed r of current acquisition _i,j, and calculate user's average throughput calculate the average throughput factor of user

Step 3, according to type of service and the corresponding GBR requirement computing service weight factor W of each business _i,j;

Step 4, according to the scheduling priority factor ω of the result of calculation calculating user that above-mentioned steps one, step 2 and step 3 are tried to achieve _i,j;

Step 5, according to the dispatching priority of users factor ω that step 4 is tried to achieve _i,jcarry out scheduling of resource.

Wherein, in step 1, calculate user's instantaneous spectrum factor according to shannon formula

$E_{i,j}^S=\min (r_{i,j},Q)$

Wherein, r _i,j=log ₂(1+SINR _i,j) represent the instantaneous transmission speed that user engraves at this time, SINR _i,jrepresent the effective signal-to-noise ratio of user i on Resource Block j, Q represents that momentary rate contrasts parameter;

And, in step 2, calculate the average throughput factor of user

$E_{i,j}^A=e^{{\stackrel{\‾}{R}}_{i,j}\left(n\right)}$

Wherein user's average throughput r _i,jrepresent the instantaneous transmission speed that user engraves at this time, T _cfor window update time, R _i,j(n-1) average transmission rate in a previous n-1 Transmission Time Interval (TTI, Transmission Time Interval) is represented.

And in step 3, described type of service is for ensureing that bit rate GBR requires that business and non-ensured bit rate require business.

For the user ensureing bit-rate requirements business, the operational authority repeated factor calculating user is as follows,

$W=e^{\mathrm{\α}({\∂}_i\·D_{\mathrm{HOL},i}\·\max (1,\mathrm{GBR}/{\stackrel{\‾}{R}}_{i,j}\left(n\right)))}$

Wherein e represents mathematics constant, and α represents the adjustable preferential level factor, D _{hOL, i}for the HOL packet delay of user i in buffer queue; represent QoS rank, namely wherein τ _ithe delay threshold of representative of consumer i, δ _iit is the time delay D of team's head packet _{hOL, i}exceed the maximum probability of delay threshold, GBR represents guarantee bit-rate parameters, represent user's average throughput.

For the user of non-ensured bit rate service, the operational authority repeated factor calculating user is as follows,

$W=e^{\mathrm{\β}\·\frac{T_D}{T_i}}$

Wherein β represents Self-adjustment Factor, T _drepresent traffic scheduling time delay, T _irepresent time delay normalized parameter.

And, in step 4, calculate the scheduling priority factor ω of user according to type of service _i,j.

For the user ensureing bit rate service, calculate its scheduling priority factor as follows:

$\begin{array}{c}{\mathrm{\ω}}_{i,j}=\frac{E_{i,j}^S}{E_{i,j}^A}\·W\\ =\frac{\min ({\log }_2(1+{\mathrm{SINR}}_{i,j}),Q)}{e^{(1-\frac{1}{T_c})\·{\stackrel{\‾}{R}}_{i,j}(n-1)+\frac{1}{T_c}\·r_{i,j}}}\×e^{\mathrm{\α}({\∂}_i\·D_{\mathrm{HOL},i}\·\max (1-\frac{\mathrm{GBR}}{{\stackrel{\‾}{R}}_{i,j}\left(n\right)}))}\end{array}$

For the user of non-ensured bit rate service, calculate its scheduling priority factor as follows:

$\begin{array}{c}{\mathrm{\ω}}_{i,j}=\frac{E_{i,j}^S}{E_{i,j}^A}\·W\\ =\frac{\min ({\log }_2(1+{\mathrm{SINR}}_{i,j}),Q)}{e^{(1-\frac{1}{T_c})\·{\stackrel{\‾}{R}}_{i,j}(n-1)+\frac{1}{T_c}\·r_{i,j}}}\×e^{\mathrm{\β}\·\frac{T_D}{T_i}}\end{array}$

The present invention and existing dispatching algorithm ratio, have the following advantages:

This method has considered the impact of service traffic class (ensureing bit rate service and non-ensured bit rate service), traffic scheduling time delay, Speed Guarantee factor pair dispatching priority, and take different computational methods for different service types and different guarantee bit rate user, distinguish the service of different brackets.

First whether the present invention ensure bit rate service person according to type of service, selects corresponding formulae discovery priority factors.If when type of service is for ensureing bit rate service, consider the GBR requirement of business, the business required for not meeting GBR is given priority parameters and is compensated. item ensures to meet the weight factor ensureing bit rate service, when a customer service Mean Speed when meeting the GBR of its rate guarantee requirement, this value is 1; When a customer service Mean Speed when being less than the GBR of its rate guarantee requirement, this value so there is no to meet the customer service of minimum speed limit requirement by priority allocation Resource Block, and obtain compensation. larger with the gap of GBR, weight factor is larger, and the priority that user obtains Resource Block is higher, thus solves the guarantee bit-rate requirements of different business.

This method gets the index of e respectively to user's average throughput and business weight, these two kinds of calculated factor is obtained in magnitude unified, overcomes the weight that may cause due to different magnitude unbalanced.

Advantage of the present invention is: be applicable to mixed service and have the different scene required to different business GBR, under the prerequisite improving fairness, sacrifice a small amount of non-ensured bit rate service performance and exchange the service quality greatly improving and ensure bit rate service for, thus improve LTE system overall performance.

Accompanying drawing explanation

Fig. 1 scheduling flow figure of the present invention

Fig. 2 priority calculation flow chart of the present invention

The simulation result figure of Fig. 3 the present invention under mixed service scene, wherein,

Fig. 3 a is the fairness curve of non-ensured bit rate service

Fig. 3 b is the fairness curve ensureing bit rate service

Fig. 3 c is the packet loss curve of non-ensured bit rate service

Fig. 3 d is the packet loss curve ensureing bit rate service

Fig. 3 e is the throughput curve of non-ensured bit rate service

Fig. 3 f is the throughput curve ensureing bit rate service

Embodiment

Below in conjunction with accompanying drawing, enforcement of the present invention is further illustrated, accompanying drawing described herein is used to provide a further understanding of the present invention, form a application's part, illustrative examples of the present invention and its explanation, for explaining the present invention, do not form inappropriate limitation of the present invention.

Step 1, the effective signal-to-noise ratio SINR of calculated signals user on each Resource Block received according to user _i,j, and according to shannon formula calculate user this time the instantaneous transmission speed r that engraves _i,j, calculate the instantaneous spectrum factor

$E_{i,j}^S=\min (r_{i,j},Q)$

Wherein, r _i,j=log ₂(1+SINR _i,j) represent the instantaneous transmission speed that user engraves at this time, SINR _i,jrepresent the effective signal-to-noise ratio of user i on Resource Block j, Q represents that momentary rate contrasts parameter, can according to LTE system order of modulation and actual conditions value.

Step 2, according to the instantaneous transmission speed r of current acquisition _i,j, and calculate user's average throughput calculate the average throughput factor of user

$E_{i,j}^A=e^{{\stackrel{\‾}{R}}_{i,j}\left(n\right)}$

Wherein user's average throughput r _i,jrepresent the instantaneous transmission speed that user engraves at this time, T _cfor window update time, represent the average transmission rate in a previous n-1 TTI.

Step 3, according to type of service and the corresponding GBR requirement computing service weight factor W of each business _i,j;

For the user ensureing bit-rate requirements business, the operational authority repeated factor calculating user is as follows,

$W=e^{\mathrm{\α}({\∂}_i\·D_{\mathrm{HOL},i}\·\max (1,\mathrm{GBR}/{\stackrel{\‾}{R}}_{i,j}\left(n\right)))}$

Wherein e represents mathematics constant, and α represents the adjustable preferential level factor, and its value can adjust according to practical application scene in the specific implementation, D _{hOL, i}for the HOL packet delay of user i in buffer queue; represent QoS rank, namely wherein τ _ithe delay threshold of representative of consumer i, δ _iit is the time delay D of team's head packet _{hOL, i}exceed the maximum probability of delay threshold, GBR represents guarantee bit-rate parameters, represent user's average throughput.

For the user of non-ensured bit rate service, the operational authority repeated factor calculating user is as follows,

$W=e^{\mathrm{\β}\·\frac{T_D}{T_i}}$

Wherein β represents Self-adjustment Factor, and its value can adjust according to practical application scene in the specific implementation, T _drepresent traffic scheduling time delay, T _irepresent time delay normalized parameter.

Step 4, according to the scheduling priority factor ω of the result of calculation calculating user that above-mentioned steps one, step 2 and step 3 are tried to achieve _i,j.

For the user ensureing bit rate service, calculate its scheduling priority factor as follows:

$\begin{array}{c}{\mathrm{\ω}}_{i,j}=\frac{E_{i,j}^S}{E_{i,j}^A}\·W\\ =\frac{\min ({\log }_2(1+{\mathrm{SINR}}_{i,j}),Q)}{e^{(1-\frac{1}{T_c})\·{\stackrel{\‾}{R}}_{i,j}(n-1)+\frac{1}{T_c}\·r_{i,j}}}\×e^{\mathrm{\α}({\∂}_i\·D_{\mathrm{HOL},i}\·\max (1-\frac{\mathrm{GBR}}{{\stackrel{\‾}{R}}_{i,j}\left(n\right)}))}\end{array}$

For the user of non-ensured bit rate service, calculate its scheduling priority factor as follows:

$\begin{array}{c}{\mathrm{\ω}}_{i,j}=\frac{E_{i,j}^S}{E_{i,j}^A}\·W\\ =\frac{\min ({\log }_2(1+{\mathrm{SINR}}_{i,j}),Q)}{e^{(1-\frac{1}{T_c})\·{\stackrel{\‾}{R}}_{i,j}(n-1)+\frac{1}{T_c}\·r_{i,j}}}\×e^{\mathrm{\β}\·\frac{T_D}{T_i}}\end{array}$

Step 5, according to the dispatching priority of users factor ω that step 4 is tried to achieve _i,jcarry out scheduling of resource.

This implementation method effect:

The downlink resource scheduling method of mixed service in the LTE system that this implementation method proposes, i.e. blended service exponential factor (HSEF, Hybrid Service Exponential Factor) algorithm, reduce the packet loss and time delay that ensure bit rate service in mixed service scene, and improve the throughput and fairness that ensure bit rate service.PF algorithm and MLWDF algorithm algorithm as a comparison, simulation analysis is carried out to these three kinds of algorithms.As shown in Fig. 3 a-f, support mixed service in the community that radius is 1km and focus on ensureing in the LTE system of bit rate service, along with the increase (number of users is greater than 30) of load, now PF algorithm and MLWDF algorithm can not ensure user performance, as being greater than 30, this two kinds of algorithm packet loss and the control breakdown of user fairness aspect as user.And HSEF algorithm is under the prerequisite improving non-ensured bit rate service fairness, sacrifice a small amount of non-ensured bit rate service throughput and packet loss meets the performance improved and promote guarantee bit rate service.

Following emulation is adopted to prove beneficial effect of the present invention:

Simulating scenes is the LTE system of a multiple cell, and the user participating in scheduling is randomly dispersed in center cell.The center of each community arranges an eNodeB and controls all available Resource Block RB, and all users share 50RB, and namely system bandwidth is 10MHz.The LTE system simulation parameter that the present invention adopts arranges as shown in table 1:

Table 1LTE downlink system parameter

The present invention is directed to mixed service scene, the type of service that emulation adopts is as shown in table 2:

Table 2 artificial service parameter

Wherein Type1 business represents non-ensured bit rate service, and Type2 represents guarantee bit rate service.

The guarantee bit rate service time delay of table 3 three kinds of algorithms compares

Table 3 gives the HSEF algorithm that the present invention carries and compares with the guarantee bit rate service time delay of classical PF algorithm and MLWDF algorithm, can find out and to be sharply deteriorated when number of users is greater than the time delay of classical PF algorithms after 30, and the time delay of MLWDF algorithm also reaches more than 0.04 second, the time delay of relative HSEF of the present invention only has about 1/3rd of MLWDF algorithm and more stable.Find with these two kinds of method comparison, the HSEF algorithm that the present invention proposes is far superior to all the other two kinds of algorithms in guarantee bit rate service time delay, and under identical scenario, the delay performance of HSEF algorithm more satisfies the demands.

Fig. 3 a and Fig. 3 b sets forth HSEF algorithm proposed by the invention and classical PF algorithm, MLWDF algorithm non-ensured bit rate service and ensure bit rate service user between fairness index contrast figure.Can find other two kinds of algorithms are far superior to for the user fairness sex index of HSEF algorithm non-ensured bit rate service user from these two figure.And for guarantee bit rate service user, when number of users is less, (number of users is less than 30) three kinds of algorithm fairness indexes are more or less the same, along with increasing of user, the fairness index of classical PF algorithm sharply declines, MLWDF algorithm is relative with the change of HSEF algorithm slowly, and the fairness index of HSEF algorithm is slightly better than MLWDF algorithm.The HSEF algorithms to improve fairness index of user.

Fig. 3 c and Fig. 3 d sets forth HSEF algorithm proposed by the invention and classical PF algorithm, the non-ensured bit rate service of MLWDF algorithm and the packet loss comparison diagram of guarantee bit rate service.Although HSEF algorithm packet loss is all higher than classical PF algorithm and MLWDF algorithm in non-ensured bit rate service, gap is also little, and along with the increase of number of users, packet loss decline and close to classic algorithm and MLWDF algorithm.In the user ensureing bit rate service, when number of users is less, the packet loss of (number of users is less than 30) three kinds of algorithms is close, along with the increase of number of users, the packet loss of classical PF algorithm and MLWDF algorithm sharply increases, make the control breakdown of LTE system performance, thus possibly cannot meet consumers' demand, although and the packet loss of HSEF algorithm also obviously increases after number of users is greater than 60, its amplitude is much smaller than classical PF algorithm and MLWDF algorithm.Therefore can find out that HSEF algorithm proposed by the invention can greatly improve the performance ensureing bit rate service user under the prerequisite of sacrificing a small amount of non-ensured bit rate service user packet loss, and along with user's increase and the demand increase ensureing bit rate service, non-ensured bit rate service packet loss fall is substantially close to classical PF algorithm and MLWDF algorithm.

Fig. 3 e and Fig. 3 f gives HSEF algorithm that the present invention carries and classical PF algorithm, MLWDF algorithm at non-ensured bit rate service and the throughput of system ensureing bit rate service.As can be seen from the figure along with the increase of number of users, non-ensured bit rate service throughput of system declines, and ensure that bit rate service throughput of system rises, and the guarantee bit rate service throughput of system of HSEF algorithm is far above classical PF algorithm and MLWDF algorithm.Although the throughput performance of HSEF algorithm is a little less than classical PF algorithm and MLWDF algorithm in non-ensured bit rate service, but difference amplitude is little, and these three kinds of algorithms are when number of users increases, throughput of system all declines, this is because ensure in LTE system scene that the increase of bit rate service demand occupies the Main Resources of LTE system.Compare with MLWDF algorithm with classical PF algorithm, the throughput of system performance of a small amount of non-ensured bit rate service of HSEF algorithm sacrifice that the present invention proposes has exchanged the raising greatly ensureing the throughput performance of bit rate service for, makes to ensure most of resource that can occupy LTE system of bit rate service thus meets the demand of user.

As can be seen from above simulating scheme, a kind of LTE system downlink resource scheduling method based on MLWDF that the present invention proposes can under the prerequisite promoting fairness between user, sacrifice a small amount of non-ensured bit rate service performance and exchange the larger lifting ensureing bit rate service throughput of system for, and significantly reduce the time delay and packet loss that ensure bit rate service, thus promote and improve the service quality ensureing bit rate service.

Claims (5)

1. the downlink resource scheduling method of mixed service in LTE system, comprises the following steps:

Step 1, according to the effective signal-to-noise ratio SINR of calculated signals user on each Resource Block that user receives _i,j, according to shannon formula calculate user this time the instantaneous transmission speed r that engraves _i,j, calculate the instantaneous spectrum factor

Step 2, according to the instantaneous transmission speed r of current acquisition _i,j, and calculate user's average throughput calculate the average throughput factor of user

Step 3, requires computing service weight factor W according to the type of service of each business and corresponding guarantee bit rate (GBR, Guranteed Bit Rate) _i,j;

Step 4, according to above-mentioned steps 1, step 2 and step 3 try to achieve result and calculate the scheduling priority factor ω of user _i,j;

Step 5, according to the dispatching priority of users factor ω that step 4 is tried to achieve _i,jcarry out scheduling of resource.

2. the downlink resource scheduling method of mixed service in LTE system according to claim 1, is characterized in that: calculate user's instantaneous spectrum factor according to shannon formula in described step 1

$E_{i,j}^S=\min (r_{i,j},Q)$

Wherein, r _i,j=log ₂(1+SINR _i,j) represent the instantaneous transmission speed that user engraves at this time, SINR _i,jrepresent the effective signal-to-noise ratio of user i on Resource Block j, Q represents that momentary rate contrasts parameter.

3. the downlink resource scheduling method of mixed service in LTE system according to claim 1, is characterized in that: the average throughput factor calculating user in step 2

$E_{i,j}^A=e^{{\stackrel{\‾}{R}}_{i,j}\left(n\right)}$

Wherein user's average throughput r _i,jrepresent the instantaneous transmission speed of user in this moment, T _cfor window update time, R _i,j(n-1) average transmission rate in a previous n-1 Transmission Time Interval (TTI, Transmission Time Interval) is represented.

4. the downlink resource scheduling method of mixed service in LTE system according to claim 1, is characterized in that: in step 3, and described type of service is for ensureing that bit rate GBR requires that business and non-ensured bit rate require business;

For the user ensureing bit-rate requirements business, the operational authority repeated factor calculating user is as follows:

$W=e^{\mathrm{\α}({\∂}_i\·D_{\mathrm{HOL},i}\·\max (1,\mathrm{GBR}/{\stackrel{\‾}{R}}_{i,j}\left(n\right)))}$

Wherein e represents mathematics constant, and α represents the adjustable preferential level factor, D _{hOL, i}for the time delay of the queue head of user i in buffer queue (HOL, Head of Line) packet; represent service quality (QoS, Quality of Service) rank, namely wherein τ _ithe delay threshold of representative of consumer i, δ _iit is the time delay D of queue heads packet _{hOL, i}exceed the maximum probability of delay threshold, GBR represents guarantee bit-rate parameters, represent user's average throughput;

For the user of non-ensured bit rate service, the operational authority repeated factor calculating user is as follows:

$W=e^{\mathrm{\β}\·\frac{T_D}{T_i}}$

Wherein β represents Self-adjustment Factor, T _drepresent traffic scheduling time delay, T _irepresent time delay normalized parameter.

5. the downlink resource scheduling method of mixed service in LTE system according to claim 1, is characterized in that: in step 4, calculates the scheduling priority factor ω of user according to type of service _i,j;

For the user ensureing bit rate service, calculate its scheduling priority factor as follows:

$\begin{array}{c}{\mathrm{\ω}}_{i,j}=\frac{E_{i,j}^S}{E_{i,j}^A}\·W\\ =\frac{\min ({\log }_2(1+{\mathrm{SINR}}_{i,j}),Q)}{e^{(1-\frac{1}{T_c})\·{\stackrel{\‾}{R}}_{i,j}(n-1)+\frac{1}{T_c}\·r_{i,j}}}\×e^{\mathrm{\α}({\∂}_i\·D_{\mathrm{HOL},i}\·\max (1,\frac{\mathrm{GBR}}{{\stackrel{\‾}{R}}_{i,j}\left(n\right)}))}\end{array}$

For the user of non-ensured bit rate service, calculate its scheduling priority factor as follows:

$\begin{array}{c}{\mathrm{\ω}}_{i,j}=\frac{E_{i,j}^S}{E_{i,j}^A}\·W\\ =\frac{\min ({\log }_2(1+{\mathrm{SINR}}_{i,j}),Q)}{e^{(1-\frac{1}{T_c})\·{\stackrel{\‾}{R}}_{i,j}(n-1)+\frac{1}{T_c}\·r_{i,j}}}\×e^{\mathrm{\β}\·\frac{T_D}{T_i}}\end{array}$