CN100512503C - Multi-mode configuration and scheduling method of proportional fair scheduling algorithm - Google Patents

Abstract

The invention discloses a proportional fair scheduling algorithm multi-mode configuration and scheduling method, which enables the proportional fair scheduling algorithm to have multiple scheduling modes and provides operators with multiple choices between sector throughput and service fairness. The method comprises the following steps: (a) determining the configuration parameters of various modes of the proportional fair scheduling algorithm; (b) storing the configuration parameters corresponding to the various modes of the proportional fair scheduling algorithm in the base station; (c) through the remote maintenance station Or the near-end maintenance station sends the scheduling mode of the selected proportional fair scheduling algorithm to the base station; (d) after the base station receives the scheduling mode of the modified proportional fair scheduling algorithm, it immediately or periodically makes the specified mode take effect.

Description

A Proportional Fair Scheduling Algorithm Multi-mode Configuration and Scheduling Method

technical field

The invention relates to a WCDMA HSDPA (high-speed downlink packet access) system, in particular to a proportional fair scheduling algorithm of a MAC-hs scheduler in the WCDMA HSDPA.

Background technique

In the MAC-hs scheduler of the WCDMA HSDPA system, the scheduling algorithms include: fair service time, maximum carrier-to-interference ratio and proportional fair scheduling algorithm, among which:

In the fair service time (Round Robin, RR) algorithm, the service time (scheduling times) obtained by each UE is almost exactly the same. This scheduling algorithm is the fairest, but the throughput rate is the lowest.

In the maximum carrier-to-interference ratio scheduling algorithm (MAX-C/I), the MAC-hs scheduler of the base station provides services for UEs with the best signal quality first, and only serves UEs with poorer signal quality if there are remaining resources. Therefore, the sector can get the maximum throughput, but this algorithm is the most unfair.

Proportional Fair Scheduling (PF) algorithm is a compromise between the fair service time algorithm and the maximum load-to-interference ratio algorithm. This algorithm can obtain relatively large throughput and better service fairness.

However, the existing proportional fair scheduling algorithm has only one scheduling mode, which cannot meet the different requirements of operators for scheduling fairness and cell throughput.

Contents of the invention

The technical problem to be solved by the present invention is to provide a multi-mode configuration method for a proportional fair scheduling algorithm, so that the proportional fair scheduling algorithm has multiple scheduling modes, and provides operators with multiple choices between sector throughput and service fairness.

In order to solve the above-mentioned technical problems, the present invention provides a multi-mode configuration method of a proportional fair scheduling algorithm, comprising the following steps:

(a) Determine the configuration parameters of various modes of the proportional fair scheduling algorithm;

(b) saving the configuration parameters corresponding to various modes of the proportional fair scheduling algorithm in the base station;

(c) Send the scheduling mode of the selected proportional fair scheduling algorithm to the base station through the remote maintenance station or the near-end maintenance station;

(d) After receiving the scheduling mode of the modified proportional fair scheduling algorithm, the base station immediately or periodically makes the specified mode effective.

Further, the above method can also have the following characteristics: the proportional public scheduling algorithm determined in the step (a) calculates the user scheduling weight according to the channel conditions, historical traffic and the configuration parameters; in the step (d) After the base station calculates scheduling parameters, namely user scheduling weights, according to the configuration parameters, user channel conditions, and historical traffic, it performs scheduling in the specified mode according to the user scheduling weights.

Further, the above method may also have the following characteristics: the greater the weight of the channel condition, the worse the fairness; the greater the weight of the historical traffic, the better the fairness.

Further, the above method may also have the following characteristics: in the step (a), the proportional fair scheduling algorithm is: user scheduling weight Priority=f ₁ (Chcond, m1)/f ₂ (Th, m2), where Priority Represents the user scheduling weight, Chcond represents the current channel condition of the user, Th represents the current historical data flow of the user, the functions _f1 and _f2 adopt basic elementary functions, and the m1 and m2 are configuration parameters.

Further, the above method may also have the following characteristics: the function f ₁ =(Chcond) ^ml , or f1=(m1) ^Chcond , or $f_1={\log }_{m1}^{\mathrm{Chcond}};$ said function f ₂ =(Th) ^m2 , or f ₂ =(m2) ^Th , or $f_2={\log }_{m2}^{\mathrm{Th}}.$

Further, the above method may also have the following characteristics: in the step (a), the determination of the configuration parameters of various modes of the proportional fair scheduling algorithm further includes the following steps: (i) selecting a plurality of different configuration parameters, according to The calculation method of user scheduling weights calculates the results of user scheduling weights; (ii) applies the scheduling weight results to simulation or testing, and determines several scheduling modes and their corresponding configuration parameters according to different scheduling effects obtained.

Further, the above method may also have the following characteristics: in the step (c), the operator determines to adopt a proportional fair scheduling algorithm, and selects and determines the scheduling mode.

Further, the above method may also have the following features: in the step (d), the scheduling of users according to the calculated weights refers to sorting the users according to the calculated user scheduling weights, and for the scheduling weight users are prioritized for scheduling.

Based on the proportional fair scheduling algorithm, the invention can configure the scheduling mode and provide operators with multiple choices between sector throughput and service fairness. After the base station receives the scheduling mode of the modified proportional fair scheduling algorithm, the specified mode will take effect immediately or periodically, so as to achieve the purpose of configurable fairness of the proportional fair scheduling algorithm, and meet different demand scenarios, such as focusing on scheduling fairness and taking into account scheduling fairness performance and cell throughput, focusing on cell throughput, etc.

Description of drawings

Figure 1 is a schematic diagram of a proportional fair scheduling algorithm scheduling mode;

FIG. 2 is a flowchart of a multi-mode scheduling method for a proportional fair scheduling algorithm according to an embodiment of the present invention.

Detailed ways

When the operator decides to use the proportional fair scheduling algorithm as the sector MAC-hs scheduling algorithm for scheduling, the embodiments of the present invention can provide multiple scheduling modes of the proportional fair scheduling algorithm, such as a mode that focuses on scheduling fairness and takes into account scheduling Modes of fairness and cell throughput, modes focusing on cell throughput, etc., as shown in Figure 1, there are N types of proportional fairness scheduling algorithms in the figure, and the fairness of mode 1 is closest to the scheduling algorithm of maximum carrier-to-interference ratio, which is Among all modes, the fairness is the worst, and the sector throughput is the largest. The fairness of mode N is the closest to the fair service time, and the fairness is the best among all parameters, and the sector throughput is the smallest.

This embodiment provides a new weight calculation method. By setting different scheduling weight parameters, different user scheduling weight calculation results are obtained, so that the ranking of users by the base station according to the scheduling weight changes, thereby generating different scheduling effects. This achieves the purpose of meeting different fairness and cell throughput requirements, and enables operators to have multiple scheduling modes to choose from when using the proportional fair scheduling algorithm.

This embodiment adopts the multi-mode configuration and scheduling method of proportional fair scheduling algorithm, including the following steps:

Step 110, determining the configuration parameters of various modes of the proportional fair scheduling algorithm;

In this embodiment, the scheduling mode parameter configuration method may adopt the following steps:

(a) Select a plurality of different configuration parameters, and calculate the result of user scheduling weight according to the calculation method of user scheduling weight;

The calculation method of user scheduling weight can use the following calculation formula:

Priority=f ₁ (Chcond, m1)/f ₂ (Th, m2)

Among them, Priority represents the scheduling weight, Chcond represents the current channel condition of the user, and Th represents the current historical data traffic (service traffic or historical traffic) of the user. Channel conditions and historical traffic are the two most important parameters for calculating scheduling weights. Channel conditions can be represented by channel quality or the air interface transmission capability corresponding to channel quality.

The function f can be a power function f(x)=x ^m , an exponential function f(x)=m ^x or a logarithmic function $f\left(x\right)={\log }_m^x$ and other basic elementary functions, m>0, for the function f ₁ , by adjusting m1, the weight of the channel condition can be changed, the greater the weight of the channel condition, the worse the fairness; for the function f ₂ , by adjusting m2, the available To change the weight of historical traffic, the greater the weight of historical traffic, the better the fairness; by adjusting the values of m1 and m2 at the same time or separately, different scheduling modes of the proportional fairness algorithm can be obtained to achieve fairness and throughput between adjustments.

(b) applying the scheduling weight result to simulation or testing, and determining several scheduling modes and their corresponding user scheduling weight parameters according to the obtained different scheduling effects;

Different scheduling effects can be obtained by choosing different parameters m, and then different scheduling modes can be produced. The corresponding relationship between parameters and different modes needs to be determined through simulation experiments on the basis of estimated values. The range of parameters, taking f ₁ as an example, assumes that Chcond is greater than 1, and the preliminary estimation is as follows:

For power function: m1>0, when m1=m2=1 $\mathrm{priority}=\frac{\mathrm{Chcond}}{\mathrm{Th}},$ This is the traditional proportional fair scheduling algorithm. When m2 is constant, the larger m1 is, the more biased it is to the maximum load-to-interference ratio scheduling algorithm, that is, the scheduling mode shows that the greater the throughput, the lower the fairness.

For the exponential function: m1>1, when m2 is constant, the larger m1 is, the more biased it is to the maximum load-to-interference ratio scheduling algorithm;

Logarithmic function: m1>1, when m2 is constant, the smaller m1 is, the more biased it is towards the maximum load-to-interference ratio scheduling algorithm;

The selection of the above parameter ranges is not intended to limit the present invention, but only proposes an implementation situation. In practice, it is necessary to determine the parameters corresponding to the N scheduling modes based on empirical values and simulation results, and then determine the scheduling mode.

Step 120, saving the several scheduling modes and their corresponding user scheduling weight parameters in the MAC-hs scheduler of the base station.

The MAC-hs scheduler can store n scheduling modes and their corresponding scheduling weight parameters in the form of a table.

Step 130, after the operator determines that the MAC-hs scheduling algorithm of the cell adopts the proportional fair scheduling algorithm, select the scheduling mode of determining the proportional fair algorithm as required, and pass the scheduling mode of the proportional fair algorithm selected by the operator through the near-end maintenance station or the remote maintenance station. The station transmits to the MAC-hs scheduler of the base station;

Operators can set the scheduling mode to take effect immediately or periodically.

Step 140: After receiving the scheduling mode of the modified proportional fair scheduling algorithm, the base station immediately or periodically enables the specified mode to take effect, calculates scheduling parameters according to the configuration parameters, and performs scheduling under the specified mode.

The base station MAC-hs scheduler searches for the parameters corresponding to the scheduling mode set by the operator. For this embodiment, that is, the m1 and m2 values corresponding to the mode, according to the parameters m1 and m2, use the user scheduling weight calculation formula to calculate According to the user scheduling weight calculated, the users are sorted, and the user with a higher scheduling weight is prioritized for scheduling, such as sending the user's data first.

The user scheduling weight calculation formula may be: Priority=f ₁ (Chcond, m1)/f ₂ (Th, m2), where the parameters have been described above.

The base station can calculate the scheduling weight of users with data to be sent in the cell in real time at each transmission time interval TTI (HSDPA system is 2ms) in units of the cell, and perform scheduling, and can also be set to calculate at other times.

Claims (8)

1. A multi-mode configuration and scheduling method of a proportional fair scheduling algorithm, comprising the following steps: (a) Determine the configuration parameters of various modes of the proportional fair scheduling algorithm; (b) saving the configuration parameters corresponding to various modes of the proportional fair scheduling algorithm in the base station; (c) Send the scheduling mode of the selected proportional fair scheduling algorithm to the base station through the remote maintenance station or the near-end maintenance station; (d) After receiving the scheduling mode of the modified proportional fair scheduling algorithm, the base station immediately or periodically makes the specified mode effective. 2. The method according to claim 1, characterized in that the proportional public scheduling algorithm determined in the step (a) calculates user scheduling weights according to channel conditions, historical traffic and the configuration parameters; in the step In (d), the base station performs scheduling in the specified mode according to the user scheduling weights after calculating the scheduling parameters, that is, user scheduling weights, according to the configuration parameters, user channel conditions, and historical traffic. 3. The method according to claim 2, wherein the greater the weight of the channel condition, the worse the fairness; the greater the weight of the historical traffic, the better the fairness. 4. The method according to claim 2, characterized in that, in the step (a), the proportional fair scheduling algorithm is: User scheduling weight Priority=f ₁ (Chcond, m1)/f ₂ (Th, m2) Among them, Priority represents the user scheduling weight, Chcond represents the current channel condition of the user, Th represents the current historical data flow of the user, the functions _f1 and _f2 adopt basic elementary functions, and the m1 and m2 are configuration parameters. 5. The method of claim 4, wherein the function f ₁ =(Chcond) ^m1 , or f ₁ =(m1) ^Chcond , or $f_1={\log }_{m1}^{\mathrm{Chcond}};$ said function f ₂ =(Th) ^m2 , or f ₂ =(m2) ^Th , or $f_1={\log }_{m1}^{\mathrm{Chcond}};$ said function f ₂ =(Th) ^m2 , or f ₂ =(m2) ^Th , or $f_2={\log }_{m2}^{\mathrm{Th}}.$ 6. The method according to claim 4, wherein in said step (a), said determining the configuration parameters of various modes of the proportional fair scheduling algorithm further comprises the following steps: (i) Select a plurality of different configuration parameters, and calculate the result of user scheduling weight according to the calculation method of user scheduling weight; (ii) Apply the scheduling weight results to simulation or testing, and determine several scheduling modes and corresponding configuration parameters according to different scheduling effects obtained. 7. The method according to claim 1, characterized in that, in the step (c), the operator determines to adopt a proportional fair scheduling algorithm, and selects and determines the scheduling mode. 8. The method according to claim 2, wherein in the step (d), the scheduling of users according to the calculated weights refers to sorting the users according to the calculated user scheduling weights, Scheduling is given priority to users with high scheduling weights.

Images