KR101203867B1 - Method of performing scheduling in wired or wireless communications system and scheduler therefor - Google Patents

Abstract

The present invention relates to a method for performing scheduling in a communication system and a scheduler thereof. In the scheduling method in the communication system according to the present invention, in the scheduling method for providing a service to at least two terminals in the communication system, a utility function in which the height of the derivative is varied according to the priority of each terminal Calculating a utility value for each terminal using a) and determining a terminal to which data is transmitted at a specific time point among the at least two terminals using the utility value for each terminal. It is done.

Scheduling, Utility Functions, Priorities, CBR, EMG

Description

Method of performing scheduling in wired or wireless communications system and scheduler therefor}

1 is a schematic structural diagram of a mobile communication system to which the present invention can be applied.

2 is a block diagram of a scheduler according to an exemplary embodiment of the present invention.

FIG. 3 is a view for explaining how the derivative U _i '(R _i ) of the utility function defined by Equation 1 is changed according to a value of b _i , c _i in an exemplary embodiment of the present invention.

4A and 4B are diagrams for describing a scheduling method according to a determination measurement.

5A to 5E illustrate simulation results in a homogeneous channel according to an exemplary embodiment of the present invention.

The present invention relates to a wired and wireless communication system. More specifically, the present invention relates to a method of performing scheduling in a communication system and a scheduler thereof.

Scheduling in a communication system can be used in a variety of senses, and is primarily related to the problem of efficiently distributing limited resources. For example, in the case of uplink, scheduling is a problem of determining when a UE transmits data using which channel resources (code, frequency, time, power, data rate, etc.) and downlink ( In the case of downlink), scheduling is a problem in which the network side determines which service quality (QoS) is provided to which terminal at a specific time among a plurality of terminals receiving a specific service. The present invention relates to downlink scheduling.

In the downlink scheduling according to the prior art, most scheduling algorithms can satisfy the requirements on the assumption that the network capacity is sufficient so that the QoS requirements of all users are satisfied once. However, because network capacity is not always sufficient, it may not always satisfy the QoS of all users, and therefore, consideration should be given to which criteria can satisfy the QoS of the user in a situation where network capacity is insufficient.

In addition, to ensure that the algorithms are actually implemented and work as desired, determine whether the quality requirements of all users can be met when the user is accepted for a new user entering the network, and then accept if possible. There must be an action that does not accept the user's entry into the network. This is called Call Admission Control (CAC). During that time, there was no Rockac Control algorithm that works in conjunction with the scheduling algorithm.

Due to the time-varying nature of the wireless channel's capacity, the existence of a rocking control algorithm does not completely solve the problem. In other words, even if the new rock control algorithm determines that a user arriving at the network is acceptable, it may change to a situation that is not possible over time. However, existing scheduling algorithms do not provide any countermeasure against this problem.

Finally, since the conventional scheduling algorithms do not deal with guaranteeing the QoS of the user and how to use the remaining capacity, this problem can lead to serious radio resource waste in the situation of guaranteeing various types of QoS.

The present invention has been made to solve the problems of the prior art as described above, an object of the present invention is to provide a scheduling method in a communication system that can maximize the satisfaction of users receiving the service.

Another object of the present invention is to provide a scheduling method that can guarantee QoS of all users and efficiently distribute the remaining network capacity.

In order to achieve the above technical problem, the present invention proposes a scheduling execution method based on a utility utilization maximization problem and a newly defined average throughput utility function. Utility functions are defined for each user, and are defined in different forms for users who require different QoS according to the type of service, and users with different priorities have different utility functions. .

It is desirable that the utility function for each user be strictly concave. Using the utility function for each user, the network side (e.g., the scheduler of the base station) selects the user whose service is to be provided for a certain period of time (e.g., each time slot). In order to maximize the overall utility function, the user to which the service is to be provided may be selected according to the utility value of each user calculated by the utility function for each user. An example of the utility value is a product of a data rate that is currently achievable and the derivative of the utility function (hereinafter referred to as a 'decision metric'), and the determination is made to maximize the overall utility function. The user with the largest weighing can be selected.

The utility function of each user should be determined to enable efficient scheduling. One feature of the utility function according to the present invention is that the priority of each user is reflected in the height of the derivative of the utility function. In other words, the utility function is set to have a high derivative for a high priority user, and the utility function with a low derivative is applied to a low priority user. According to this method, the network may allow a user having a higher priority to be selected first when selecting a user to be provided with a service. The priority for each user can be freely determined by the billing system, type of service, and other policy considerations of the network operator.

Another characteristic of the utility function according to the present invention is that the derivative value takes different forms depending on the type of service provided to each user. For example, not only must the utility function and minimum average data rate be guaranteed for users whose minimum average data rate is guaranteed (CBR: Constant Bit Rate), but if possible receive data above the minimum average data rate. It is desirable that the derivative of the utility function for the desired user with elasticity (EMG) has a different form.

More specifically, the derivative value of the utility function of the CBR user drops sharply to 0 when the average transfer rate for the CBR user exceeds the minimum requirement, and the derivative value of the utility function of the EMG user is determined to be the average transfer rate for the EMG user. It is preferable to take a form that drops rapidly when the minimum requirement is exceeded but gradually decreases when the derivative value falls below a certain value. If the utility function is selected according to the above method, in case of CBR users, if the average data rate exceeds the minimum requirement, the decision quantity will drop to zero, so that the CBR user will not receive any more choices when performing scheduling. On the other hand, even if the EMG user achieves the minimum requirement, the derivative value does not become zero, so that the scheduler has a choice. Considered together with the above-mentioned priorities, the minimum requirement is satisfied from the highest priority user, that is, the user with the highest height of the derivative of the utility function, and the minimum requirement is subsequently satisfied in order of priority.

In one aspect of the present invention, in the method of performing scheduling in a communication system according to the present invention, in the method of performing scheduling for providing a service to at least two terminals in a communication system, a height of a derivative is determined according to the priority of each terminal. Calculating a utility value for each terminal using a variable utility function, and determining a terminal to which data is transmitted at a specific time point among the at least two terminals using the utility value for each terminal Characterized in that comprises a step comprising.

In another aspect of the present invention, a network scheduler according to the present invention is a network scheduler for determining a terminal to be provided with a service in a communication system, using a utility function in which the height of the derivative is varied according to the priority of each terminal. And means for calculating a utility value for each terminal and means for determining a terminal to which data is transmitted at a specific time point among at least two or more terminals by using the utility value for each terminal.

The construction, operation, and other features of the present invention will be readily understood by the preferred embodiments of the present invention described below with reference to the accompanying drawings.

1 is a schematic structural diagram of a mobile communication system to which the present invention can be applied. The mobile communication system is composed of a network and a plurality of terminals. In FIG. 1, a base station (BS) is a part corresponding to an end of the network and communicates with the plurality of terminals through an air interface.

The plurality of terminals may be roughly classified into a CBR terminal and an EMG terminal, as described above, mainly by the type of service provided. An example of a CBR user is a voice user. Since voice is encoded at a constant average rate, it only needs to guarantee the same data rate as that encoding rate, and allocating a higher rate than that is a waste of resources. An example of an EMG user may be an MPEG-4 moving picture experts group-4 fine granularity scability (EGS) user. Multimedia encoded with MPEG-4 FGS requires a minimum bit rate to be decoded, and a higher bit rate can be assigned to receive better image quality. Therefore, the EMG user must meet the minimum data rate and allocate more if there is remaining capacity. Of course, the EMG user may be an ordinary data user who wants to receive a premium service. In addition, an EMG user with a minimum requirement of 0 becomes an elastic user, and the elastic user can allocate the remaining capacity without any minimum guarantee. Unless specifically mentioned below, EMG users include resilient users.

The following embodiments presuppose a cellular mobile communication system capable of serving only one user at a predetermined period, for example, one time slot, at a base station. The cellular mobile communication system may include an HSDPA, a high data rate (HDR) system, and the like. However, the technical features of the present invention are not applicable only to the cellular mobile communication system as described above, and various modifications are possible within the scope of the present invention according to the intention of the designer.

In FIG. 1, a scheduler for determining a terminal to provide a service at a specific time point among a plurality of terminals is generally located at a base station, but is not limited thereto. That is, the scheduler may be located not only in the base station but also in a specific part of the network connected to the base station.

2 is a block diagram of a scheduler according to an exemplary embodiment of the present invention. Referring to FIG. 2, the scheduler 20 stores a memory module 21 storing parameters such as a utility function, priority, minimum transfer rate, and the like for each user, and the memory module 21. Determinant calculation module 22 for calculating a decision metric using utility functions and parameters stored in each user), and the determinant for each user calculated by the determinant calculation module 22. The scheduling module 23 selects a user to which data is to be transmitted at every time slot.

If S is the set of all users, and C and E are the set of CBR users and the set of EMG users, respectively, then S = C∪E. If R _i is the average transfer rate received by user i, the utility function U _{i for} each user stored in the memory module 20 is represented by Equation 1 and Equation 1 according to whether the corresponding user is a CBR user or an EMG user. It is defined as Equation 2.

는 m_i+δ_i로 정의되는데, 여기서 δ_i는

를 0이라 봐도 무방할 정도로 아주 작은 양의 상수이거나 또는 탄력 사용자에 대해서는 0이다. a_i와 c_i는 사용자 i의 우선순위에 따라 결정되는 상수인데, 특별히 a_i는 모든 EMG 사용자에 대해 동일한 값을 갖는다. b_i도 모든 사용자에 대해 동일한 값을 갖는다. 수학식 1 및 수학식 2에 의해 정의되는 효용함수는 연속성과 엄밀한 위로 볼록(strict concavity)의 성질을 갖는다. Where a _i , b _i , c _i , ㅿ _i are positive constants, and m _i is the minimum transfer rate requirement of user i.

Is defined as m _i + δ _i , where δ _i is

Is a very small positive constant or 0 for elastic users. a _i and c _i are constants determined according to the priority of user i, in particular a _i has the same value for all EMG users. b _i also has the same value for all users. The utility function defined by Equations 1 and 2 has the property of continuity and strict concavity.

FIG. 3 is a diagram for describing how the derivative U _i '(R _i ) of the utility function defined by Equation 1 varies according to values of b _i and c _i . For all cases fixed m _i = 4. First, when b _i = 1 and c _i = 8, even if the minimum requirement of 4 is exceeded, it can be seen that U _i '(R _i ) falls slowly rather than rapidly falling to zero. When b _i is fixed and c _i = 16, it can be seen that only the height is higher and almost the same characteristics. On the other hand, when b _i = 50, the value of U _i '(R _i ) drops sharply above the minimum requirement of 4; when c _i increases, U _i ' (R _i ) increases in height, and c _i decreases in value. U _i '(R _i ) is lowered.

인 부분을 제외하고는 똑같은 특성을 갖는다. 이런 특성에 따라 수학식 2에 정의된 효용함수에서 R_i≥

인 부분을 탄력 부분(elastic part), R_i＞

인 부분을 비탄력 부분(inelastic part)이라 칭하기로 한다.Therefore, when the value of b _i is set sufficiently large, the height of the derivative of U _i '(R _i ), that is, the utility function U _i ' (R _i ) can be adjusted while changing only the value of c _i . Here, even if the value of c _i is changed, the characteristic that U _i '(R _i ) falls to 0 does not change when the minimum requirement is exceeded. Thus, if U _i (R _i ) is used in the utility function maximization problem, it is easy to see that R _i will not exceed m _i . This property of the utility function can be used when implementing the scheduling algorithm. The function defined by Equation 2 is also R _i ≥

It has the same characteristics except for the part. According to this characteristic, R _i ≥ in the utility function defined in Equation 2

The phosphorus part is an elastic part, R _i ＞

The phosphorus part will be referred to as an inelastic part.

In FIG. 2, as described above, the memory module 21 stores not only a utility function for each user but also parameters necessary for scheduling. Each user can be classified into a CBR user and an EMG user, and in detail, can be classified according to the QoS class. The QoS class may be defined for each user according to whether it is CBR or EMG, priority, minimum average data rate requirement, billing system, or the like.

As an example, the QoS class and its corresponding utility function may be expressed as C _i (m _i , c _i , b _i ), E _i (m _i , c _i , b _i , a _i ). For example, C ₁ (10, 5, 50) means CBR class 1 and the utility function of users belonging to the class is m ₁ = 10, c ₁ = 5, b ₁ = 50 Will be the function defined in.

The determination weighing calculation module 22 calculates the determination weighing for each user receiving the current service stored in the memory module 21 using the utility function and parameters for each user. The decision metric is a utility value for each user for maximizing the overall utility function, and can be calculated by multiplication of the data transfer rate currently available for each user and the derivative of the utility function for each user.

The data rate that can be provided for each user may be determined according to a minimum average data rate requirement previously promised for each user. The minimum average data rate requirement is the minimum requirement of the average data rate to be provided for each user. If the average data rate for a user at a particular time point is less than the minimum average data rate requirement, the data rate that can be provided to the user at that time point becomes large.

The method of calculating the determination quantity by the product of the data rate currently available for each user and the derivative of the utility function for each user is merely an example, and can be calculated by various methods to achieve the maximization of the overall utility function. will be.

The scheduling module 23 determines a user to which data is to be transmitted at a corresponding time point (time slot) based on each user's decision measurement calculated by the decision measurement calculation module 22. For example, the user who has the largest product of the data rate currently available for each user and the derivative of the utility function for each user can be determined as the user to which data will be transmitted at that time. When the above process is expressed by Equation 3, Equation 3 below.

Here, R _i (t) is the average transfer rate up to t time for user i, and r _{i, t + 1} is the data transfer rate available to i user at (t + 1). If the base station can transmit data to two or more users at a specific time point, the base station may determine the user to which data is transmitted at the corresponding time point in order of increasing determination.

4A and 4B are diagrams for explaining a scheduling method according to the determination measurement, C ₁ (6.4, 204.8, 50), C ₂ (1.6, 25.6, 50), E ₁ (5, 40, 50, 1 ), E ₂ (10, 40, 50, 1), E ₃ (0, 0, 0, 1) shows the utility function and its derivative. In FIG. 4B, it can be seen that the height of the derivative of CBR class 1 (C ₁ ) is the highest. If the utility functions of FIG. 4B are used for the utility function maximization problem as it is, resources (data transfer rate) will be allocated to CBR class 1 first. This is because allocating resources to CBR class 1 contributes most to maximizing the utility function. If the average data rate assigned to CBR class 1 exceeds the minimum average data rate requirement, 64 Kbps, the derivative will drop sharply to zero, so no more resources will be allocated to CBR class 1, with the next highest derivative. Resources will be allocated to CBR class 2 (C ₂ ).

If the system has sufficient capacity, allocate resources to CBR class 1 with an average data rate of 64 Kbps, and then allocate the remaining resources to CBR class 2, which is the next priority, and continue to have the lowest derivative in the manner described above. If the minimum requirement is met up to EMG Class 2 (E ₂ ) and there is still capacity available, the remaining capacity will be divided among the EMG classes except CBR. This is because the derivative of the utility function of the EMG classes drops slowly, rather than dropping to zero even if the minimum average data rate requirement is exceeded.

In summary, it can be seen that the priority is set to C ₁ > C ₂ > E ₁ > E ₂ > E ₃ depending on the height of the derivative value. As described above, since the height of the derivative can be adjusted with only the value of c _i, it can be seen that it is very easy to prioritize classes.

The data rate that can be provided to user i at (t + 1) can be determined by whether a minimum average data rate requirement is being provided for the user i, the priority, the capacity of the system, and the like. For example, if the i user is a CBR user and the minimum average data rate requirement for the i user is currently satisfied (time t), then the data rate available to the i user in the next time slot t + 1 is 0. . If the minimum average data rate requirement for the i user is not currently met (time t), then the data rate available to the i user must be determined so that it can be met. Preferably, it is desirable to determine the channel state for each user in scheduling. Channel status for each user can be identified by each user checks the CINR, BER, FER and the like with respect to the reference signal transmitted from the base station.

5A to 5E illustrate simulation results in a homogeneous channel according to an exemplary embodiment of the present invention. The cellular radio communication system subjected to the simulation includes five classes C ₁ (6.4, 204.8, 50), C ₂ (1.6, 25.6, 50), E ₁ (5, 40, 50, 1), E ₂ (10, Suppose there are 40, 50, 1), E ₃ (0, 0, 0, 1) and there are 20 users in each class. The utility function and derivative corresponding to the above classes are as shown in FIGS. 4A and 4B, respectively.

It is assumed that the attainable data rate r _{i, τ} known for each time slot is given by Shannon bound, as shown in Equation 4 below.

는 채널 대역폭이고,

는 수신된 신호의 세기이며,

는 잡음의 세기이다. here

Is the channel bandwidth,

Is the strength of the received signal,

Is the strength of the noise.

=1MHz로 했을 때 사용자의 평균 전송율 분포도와 시간에 따른 평균 전송율의 추이를 도시한 것이다. 도 5a에서, CBR 사용자들의 집합인 C₁과 C₂에 속하는 모든 사용자들의 평균 전송율이 거의 정확하게 최소 평균 데이터 전송률 요구량인 64kbps와 16kbps를 달성한다는 것을 볼 수 있다. 반면에, EMG 클래스인 E₁과 E₂에 속하는 사용자들의 최소 요구량은 기본적으로 다 만족이 되고 거의 비슷한 양만큼 더 받는다는 것을 알 수 있다. 또한, 추가적으로 받은 양은 EMG 클래스 E₃에 속하는 탄력 사용자들이 받는 전송율과 거의 비슷하다는 것을 알 수 있다. 즉, 모든 사용자들의 최소 요구량을 만족시킨 다음에 남는 용량은 EMG 사용자들이 거의 비슷한 양으로 나눠 갖는다는 것을 알 수 있다. 도 5b는 각 클래스로부터 한 명의 사용자만을 뽑아 시간에 따른 평균 전송율을 도시한 도면으로서, CBR 사용자는 정확하게 최소 평균 데이터 전송률 요구량으로 수렴하는 것을 볼 수 있다. 반면에, EMG 사용자의 전송율은 일단 최소 평균 데이터 전송률 요구량으로 수렴했다가 모든 QoS 사용자의 최소 요구량이 다 만족되고 남은 용량을 받는다는 것을 알 수 있다.5A and 5B

When = 1MHz, it shows the distribution of the average rate of the user and the trend of the average rate over time. In Figure 5a, it can be seen that almost exactly that the average rate of all users belonging to the set of C ₁ and C ₂ of the CBR users achieve the minimum average data rate requirement of 64kbps and 16kbps. On the other hand, it can be seen that the minimum requirements of users belonging to the EMG classes E ₁ and E ₂ are basically satisfied and receive almost the same amount. In addition, it can be seen that the amount received is almost the same as the transmission rate received by elastic users belonging to EMG class E ₃ . In other words, it can be seen that the capacity remaining after satisfying the minimum requirements of all users is divided by almost the same amount of EMG users. FIG. 5B is a diagram illustrating the average transfer rate over time by extracting only one user from each class, and the CBR user can see that it converges exactly to the minimum average data rate requirement. On the other hand, it can be seen that the transmission rate of the EMG user once converged to the minimum average data rate requirement, and then the minimum requirements of all QoS users are satisfied and the remaining capacity is received.

를 0.8MHz로 줄여서 시스템 용량을 줄어들게 하여 시뮬레이션을 다시 해본 결과를 도시한 것이다. 도 5c에 도시된 바와 같이, QoS 사용자의 최소 요구량이 정확하게 보장이 되고 E₃에 속하는 탄력 사용자는 모두 0의 전송율을 얻는다는 것을 알 수 있다. 따라서, E₃가 가장 낮은 우선권을 받는다고 할 수 있다.

를 0.6MHz로 줄이면 도 5d와 같은 결과를 얻을 수 있는데, C₁, C₂, E₁의 모든 최소 요구량은 만족되고 E₂의 요구량은 만족되지 않는다는 것을 볼 수 있다. 따라서, E₂ 가 그 다음으로 낮은 우선권을 갖는다고 할 수 있다. 물론, E₃는 이 경우에도 0의 전송율을 달성한다는 것을 볼 수 있다. 도 5e는

를 0.3MHz로 정했을 때의 결과이고, 결국

를 계속 줄이면서 앞에서 기대했던 우선순위 C₁＞C₂＞E₁＞E₂＞E₃을 실험적으로 확인할 수 있다.5C

Figure 2 shows the results of the simulation again by reducing the system capacity to 0.8 MHz. As shown in FIG. 5C, it can be seen that the minimum requirement of the QoS user is guaranteed correctly and the elastic users belonging to E ₃ all get a transmission rate of zero. Therefore, it can be said that E ₃ receives the lowest priority.

Reducing to 0.6 MHz yields the same result as in FIG. 5D. It can be seen that all minimum requirements of C ₁ , C ₂ , and E ₁ are satisfied, and the requirements of E ₂ are not satisfied. Thus, it can be said that E ₂ has the next lower priority. Of course, it can be seen that E ₃ achieves a transmission rate of zero even in this case. 5E

Is the result of setting 0.3 MHz to

We continue to reduce the number and experimentally confirm the priorities C ₁ > C ₂ > E ₁ > E ₂ > E ₃ .

It will be apparent to those skilled in the art that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. Accordingly, the above detailed description should not be construed as limiting in all aspects and should be considered as illustrative. The scope of the invention should be determined by reasonable interpretation of the appended claims, and all changes within the equivalent scope of the invention are included in the scope of the invention.

The present invention has the following effects.

First, it is possible to provide differential quality of service according to QoS class having priority and minimum average data rate requirement for each user.

Second, various profit models can be created by enabling various applications requiring QoS guarantees to be serviced by a wireless Internet system such as WiBro.

Third, even if the channel is changed and the network capacity is reduced to satisfy the user's demands, but suddenly fails to satisfy the user's demands, the service is guaranteed in the order of priorities that are already set, thus providing users with various quality of service levels. The user can select a service that suits his or her taste from various quality of service.

Fourth, the present invention can be easily applied to the wired Internet as well as the wireless Internet can provide a variety of revenue models to wired Internet providers.

Claims (21)

In the scheduling method for providing a service to at least two terminals in a communication system, Calculating a utility value for each terminal by using a utility function whose derivative value is changed according to the priority of each terminal; And And determining a terminal to which data is transmitted at a specific time point among the at least two terminals by using the utility value of each terminal. The method of claim 1, The utility value of each terminal is calculated by a product of a data rate provided for each terminal and the derivative of the utility function for each terminal at the specific time point. The method according to claim 1 or 2, The higher priority terminal, the higher the derivative value of the utility function is scheduling method in a communication system, characterized in that the higher. The method of claim 3, The utility function for the terminal with the minimum average data rate guaranteed (CBR) and the minimum average data rate with the guaranteed function, and the utility function for the terminal receiving data above the minimum average data rate (EMG) have different forms. A scheduling method in a communication system characterized by the above-mentioned. 5. The method of claim 4, The utility function for the CBR terminal is a scheduling method in a communication system, characterized in that the derivative value of the utility function is set to zero when the average transmission rate for the terminal exceeds the minimum average data rate. 5. The method of claim 4, The utility function for the EMG terminal is set such that the instantaneous value of the utility function drops sharply to a specific threshold and gradually drops to a specific threshold when the average transmission rate for the terminal exceeds the minimum average data rate. A scheduling method in a communication system. 3. The method of claim 2, A method of performing scheduling in a communication system, characterized in that the terminal having the largest utility value calculated among the at least two terminals is determined as the terminal to which data is to be transmitted at the specific time point. The method of claim 5, The utility function is

R _i is the average transfer rate received by user i, b is a positive constant, c _i is a constant determined according to the priority of the user i, and m _i is the minimum transfer rate requirement of the user i A method of performing scheduling in a communication system. The method of claim 6, The utility function is

R _i is the average transfer rate received by user i, b is a positive constant, a _i and c _i are constants determined according to the priority of the user i, and m _i is the minimum transfer rate requirement of user i A scheduling method in a communication system characterized by the above-mentioned. The method of claim 1, And determining a channel state of each terminal when determining terminals to which data is to be transmitted at the specific time point. A network scheduler for determining a terminal to be provided with a service in a communication system, Means for calculating a utility value for each terminal by using a utility function whose derivative value is varied according to the priority of each terminal; And And a means for determining a terminal to which data is transmitted at a specific time point among at least two or more terminals by using the utility value for each terminal. 12. The method of claim 11, The utility value of each terminal is calculated by a product of a data rate that can be provided for each terminal at the specific time point and the derivative of the utility function for each terminal. 13. The method according to claim 11 or 12, The higher priority terminal is a scheduler, characterized in that the value of the derivative of the utility function is set higher. 14. The method of claim 13, The utility function for a terminal having a minimum average data rate (CBR) guaranteed and the utility function for a terminal having a minimum average data rate guaranteed and receiving data above the minimum average data rate (EMG) have different forms. Scheduler. The method of claim 14, The utility function for the CBR terminal is a scheduler, characterized in that the derivative value of the utility function is set to zero when the average transmission rate for the terminal exceeds the minimum average data rate. The method of claim 14, The utility function for the EMG terminal is set such that the instantaneous value of the utility function drops sharply to a specific threshold and gradually drops to a specific threshold when the average transmission rate for the terminal exceeds the minimum average data rate. Scheduler. The method of claim 12, The scheduler, characterized in that the terminal having the largest utility value calculated among the at least two or more terminals is determined as the terminal to which data is to be transmitted at the specific time point. 16. The method of claim 15, The utility function is

R _i is the average transfer rate received by user i, b is a positive constant, c _i is a constant determined according to the priority of the user i, and m _i is the minimum transfer rate requirement of the user i Scheduler. 17. The method of claim 16, The utility function is

R _i is the average transfer rate received by user i, b is a positive constant, c _i is a constant determined according to the priority of the user i, and m _i is the minimum transfer rate requirement of the user i Scheduler. 12. The method of claim 11, The scheduler, characterized in that further considering the channel state of each terminal when determining the terminal to be transmitted data at the specific time. delete

Images