A resource allocation method based on user demand in non-orthogonal multiple access

Abstract

The present invention belongs to the field of information processing and communication technology, and is particularly a resource allocation method based on user needs under non-orthogonal multiple access. The basic idea of this method using reinforcement learning to solve NOMA resource allocation is to use spectrum utilization or energy efficiency as the optimization goal and user service quality as the constraint to obtain a resource allocation optimization function; then define the multi-agent user state space, reward, and action space, obtain state space information through low communication overhead, obtain one-dimensional state space data, and then obtain spectrum and power allocation strategies based on the state space information; finally, find the optimal resource allocation strategy by training a deep neural network. Based on the traditional reinforcement learning algorithm, the present invention completes service classification according to user needs, adds service type information to limit the influence of channel state information, and proposes to change the constraint conditions of the target optimization problem according to user needs, which is more in line with the actual situation and makes the system more flexible.

Description

Technical Field

[0001] The present invention relates to the field of information processing and communication technology, and in particular to a resource allocation method based on user demand under non-orthogonal multiple access. Background Art

[0002] Non-orthogonal Multiple Access (NOMA) has been proposed as a key enabling technology for the fifth-generation (5G) wireless networks, and it is currently expected to be used in 5G and beyond and 6G. The basic idea of NOMA is to provide services to multiple users in the same resource (time / frequency / code) block (RB). When the signals of different users are superimposed in the power domain, the receiver uses continuous serial interference cancellation (SIC) to distinguish each other. As a result, the number of users and the spectrum efficiency can be increased many times. Usually, due to the limitations of the hardware settings of the receiver and in order to meet the requirements of interference cancellation, NOMA with two-user multiplexing is a typical case.

[0003] NOMA is based on an orthogonal system, with power reuse at the transmitter and serial interference cancellation (SIC) at the receiver. NOMA can be used in both downlink and uplink. Currently, downlink NOMA has received more attention in wireless communication research. Therefore, this paper focuses on downlink NOMA.

[0004] In the downlink of the NOMA system, the base station encodes each user individually, and then allocates power according to a certain algorithm based on the user's own channel conditions. The signals of each user in a group are multiplexed on the same time and frequency domain resources, and then sent through OFDM modulation.

[0005] At the receiving end, the superimposed user signals are first obtained through OFDM demodulation, and then multi-user detection is performed using the serial interference cancellation technology based on the signal-to-interference-and-noise ratio (SINR) or power of each superimposed user to correctly receive the corresponding information.

[0006] The serial interference cancellation is achieved by first decoding high-power users, reconstructing the signal, and then eliminating the user signal from the superimposed signal. The low-power users are then decoded and reconstructed.

[0007] Currently, there are many resource algorithms in NOMA, which mainly focus on intelligent algorithms, game theory and deep reinforcement learning to solve the optimal objective function.

[0008] Traditional NOMA resource allocation methods, using reinforcement learning algorithms, can achieve theoretically optimal performance. However, their optimal solutions are inflexible, relying solely on channel state information to determine optimal spectrum utilization or energy efficiency. This leads to significant resource waste and neglects the impact of user needs. For example, some users may not require high transmission rates but receive more resources due to better channel conditions. While this improves system throughput, it has no real impact.

[0009] Therefore, we propose a resource allocation method based on user demand under non-orthogonal multiple access to solve the above problems. Summary of the Invention

[0010] (1) Technical problems solved

[0011] In response to the shortcomings of the existing technology, the present invention provides a resource allocation method based on user needs under non-orthogonal multiple access, which solves the problem that the optimal solution of the reinforcement learning algorithm used in the traditional NOMA resource allocation method is not flexible enough, and only uses channel state information to determine constraints to solve the optimal spectrum utilization or optimal energy efficiency, which leads to a large amount of resource waste and ignores the impact of user needs.

[0012] (2) Technical solution

[0013] The present invention is mainly implemented using reinforcement learning. The basic idea of using reinforcement learning to solve NOMA resource allocation is to use spectrum utilization or energy efficiency as the optimization target and user service quality as the constraint to obtain the resource allocation optimization function. Then define the multi-agent user state space, reward and action space, obtain the state space information through a small communication overhead, obtain one-dimensional state space data, and then obtain the spectrum and power allocation strategy based on the state space information. Finally, find the best resource allocation strategy by training a deep neural network. Based on the traditional reinforcement learning algorithm, the present invention completes the service classification according to user needs, adds service type information to limit the influence of channel state information, and proposes to change the constraint conditions of the target optimization problem according to user needs, which is more in line with the actual situation and the system is more flexible.

[0014] In order to achieve the above-mentioned purpose, the present invention specifically adopts the following technical solutions:

[0015] A resource allocation method based on user demand under non-orthogonal multiple access comprises the following steps:

[0016] Step 1: Establish a cellular network model;

[0017] Step 2: User initialization, obtaining user channel state information, which includes channel gain, attenuation, interference, and noise, and then obtaining the user transmission rate based on this information;

[0018] Step 3: Use the traditional user grouping method to group users with the goal of maximizing system throughput;

[0019] Step 4: Mark users based on their needs and regroup them;

[0020] Step 5: Classify the marked users, select different intra-subchannel constraints, and provide inter-subchannel constraints based on user requirements;

[0021] Step 6: Based on the constraints determined in step 5, the optimal power allocation is solved with energy efficiency as the objective function;

[0022] Step 7: Train the deep neural network to find the optimal resource allocation strategy.

[0023] Furthermore, the user transmission rate in step 2 is as follows:

[0024] Weak user OMA rate:

[0025] Strong user OMA rate:

[0026] Weak user NOMA rate:

[0027] Strong user NOMA rate:

[0028] Where B is the bandwidth, p is the power, H is the channel gain, and N0 is the noise.

[0029] Furthermore, the marked users in step 4 are: the traditional user grouping scheme aims to maximize system throughput. After completing the traditional grouping, the users are regrouped according to the needs of the users in their user group to ensure that users with greater service demands are not reused on the same channel resources as much as possible. At the same time, high-demand users are marked, hereinafter referred to as marked users.

[0030] Furthermore, the constraints in step 5 are: there are three channel types of constraints: weak users are marked users, strong users are marked users, and both users in the user group are not marked users in the sub-channel, specifically:

[0031] When only strong users are marked users: the constraint is P1>P2. This constraint requires that the power of weak users be greater than the power of strong users in the same group. This constraint is to ensure the minimum standard for serial interference cancellation.

[0032] When only weak users are marked users: the constraint takes This constraint requires that the transmission rate of weak users is greater than that of strong users in the same group. This constraint is to ensure the maximum service quality for weak users;

[0033] When both users in the user group are not marked users: the constraint takes and This constraint is that the NOMA transmission rate of both users is greater than their OMA transmission rate. This constraint is to ensure the most basic requirements of the NOMA system;

[0034] The transmission rate of the channel between the sub-channels that satisfies the marked user is greater than the transmission rate of the unmarked user, that is, R 标 >R 非标 .

[0035] Furthermore, the objective function optimization problem in step 6 can be expressed as:

[0036]

[0037] C3:R m,n ≥R min

[0038] C4:(type1:P1>P2

[0039]

[0040] C5:R 标 >R 非标 .

[0041] (3) Beneficial effects

[0042] Compared with the prior art, the present invention provides a resource allocation method based on user demand under non-orthogonal multiple access, which has the following beneficial effects:

[0043] 1. The user grouping result of the present invention is reorganized based on the original result according to the power allocation strategy, so that the user grouping result is more suitable for actual communication systems.

[0044] 2. The present invention uses channel state information and user demand information as parameters for resource allocation, which has more reference value.

[0045] 3. The present invention utilizes a neural network to classify the service requirements of user groups and selects the constraints of the optimization function according to the classification results, which is more flexible in solving the optimal resource allocation strategy. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a flow chart of a resource allocation method based on user demand under non-orthogonal multiple access of the present invention;

[0047] Figure 2 This is a flow chart of the regrouping process of the present invention;

[0048] Figure 3This is a schematic diagram of the channel type classification network structure of the present invention. DETAILED DESCRIPTION

[0049] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0050] Example

[0051] like Figure 1-3 As shown, an embodiment of the present invention proposes a resource allocation method based on user demand under non-orthogonal multiple access, which includes the following steps:

[0052] Step 1: Establish a cellular network model;

[0053] Step 2: User initialization, obtaining user channel state information, which includes channel gain, attenuation, interference, and noise, and then using this information to obtain the user transmission rate;

[0054] Step 3: Use the traditional user grouping method to group users with the goal of maximizing system throughput;

[0055] Step 4: Mark users based on their needs and regroup them;

[0056] Step 5: Classify the marked users, select different intra-subchannel constraints, and provide inter-subchannel constraints based on user requirements;

[0057] Step 6: Based on the constraints determined in step 5, the optimal power allocation is solved with energy efficiency as the objective function;

[0058] Step 7: Train the deep neural network to find the optimal resource allocation strategy.

[0059] Specifically, the present invention proposes a resource allocation method based on user demand under non-orthogonal multiple access, which includes the following steps:

[0060] like Figure 1 As shown in FIG, a flow chart of a resource allocation method based on user demand under non-orthogonal multiple access is shown. The method specifically includes the following steps:

[0061] Step 1: Establish a cellular network model.

[0062] Step 2: User initialization, obtain user channel state information, mainly including channel gain H, attenuation, and noise N0, and then use this information to obtain the user transmission rate:

[0063] Weak user OMA rate:

[0064] Strong user OMA rate:

[0065] Weak user NOMA rate:

[0066] Strong user NOMA rate:

[0067] Step 3: Use the traditional user grouping method to group users with the goal of maximizing system throughput.

[0068] Step 4, use Figure 2 The user regrouping process is to mark users according to their needs and regroup them. In principle, users with higher business needs are divided into different groups.

[0069] Step 5, use Figure 3 The neural network classifies the labeled users, selects different inter-subchannel constraints, and gives inter-subchannel constraints based on user needs:

[0070] When only strong users are marked users: the constraint is P1>P2. This constraint means that the power of weak users is greater than the power of strong users in the same group. This constraint is to ensure the minimum standard for serial interference cancellation.

[0071] When only weak users are marked users: the constraint takes This constraint is that the transmission rate of weak users is greater than that of strong users in the same group. This constraint is to ensure the maximum service quality for weak users.

[0072] When both users are not marked users: the constraint takes and This constraint is that the NOMA transmission rate of both users is greater than their OMA transmission rate. This constraint is to ensure the most basic requirements of the NOMA system.

[0073] Step 6: Based on the constraints determined in step 5, the optimal power allocation is solved with energy efficiency as the objective function:

[0074]

[0075] C3:R m,n ≥R min

[0076] C4:(type1:P1>P2

[0077]

[0078] C5:R标 >R 非标 .

[0079] Step 7: Train the deep neural network to find the optimal resource allocation strategy.

[0080] The present invention can improve the service quality of sub-channels with higher user demands and marked users in sub-channels to a certain extent by constructing a resource allocation method based on user demands under non-orthogonal multiple access, so that the resources of the communication system can be effectively utilized.

[0081] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A resource allocation method based on user demand in non-orthogonal multiple access, characterized by: The following steps are involved: Step 1: Establish a cellular network model; Step 2: User initialization, obtaining user channel state information, which includes channel gain, attenuation, interference, and noise, and then obtaining the user transmission rate based on this information; The user transmission rate in step 2 is as follows: Weak user OMA rate: Strong user OMA rate: Weak user NOMA rate: Strong user NOMA rate: Where B is the bandwidth, p is the power, H is the channel gain, and N0 is the noise Step 3: Use the traditional user grouping method to group users with the goal of maximizing system throughput; Step 4: Mark users based on their needs and regroup them; Step 5: Classify the marked users, select different intra-subchannel constraints, and provide inter-subchannel constraints based on user requirements; The constraints in step 5 are: three types of channel constraints: weak users are marked users, strong users are marked users, and two users in the user group are not marked users. Specifically, When only strong users are marked users: the constraint is P1>P2. This constraint requires that the power of weak users be greater than the power of strong users in the same group. This constraint is to ensure the minimum standard for serial interference cancellation. When only weak users are marked users: the constraint takes This constraint requires that the transmission rate of weak users is greater than that of strong users in the same group. This constraint is to ensure the maximum service quality for weak users; When both users in the user group are not marked users: the constraint takes and This constraint is that the NOMA transmission rate of both users is greater than their OMA transmission rate. This constraint is to ensure the most basic requirements of the NOMA system; The transmission rate of the channel between the sub-channels that satisfies the marked user is greater than the transmission rate of the unmarked user, that is, R 标 >R 非标 ; Step 6: Based on the constraints determined in step 5, the optimal power allocation is solved with energy efficiency as the objective function; The objective function optimization problem in step 6 can be expressed as: C3:Rm ,n ≥Rmin C4:(type1:P1>P2 C5: R standard>R non-standard Step 7: Train the deep neural network to find the optimal resource allocation strategy.

2. The method for allocating resources based on user demand in non-orthogonal multiple access according to claim 1, characterized in that: The marked users in step 4 are: the traditional user grouping scheme aims to maximize system throughput. After completing the traditional grouping, the users in the user group are regrouped according to their needs to ensure that users with greater service demands are not reused on the same channel resources. At the same time, high-demand users are marked, hereinafter referred to as marked users.

Images