Joint resource allocation method for d2dm communication based on cellular network

Abstract

The application discloses a kind of joint resource allocation method based on cellular network D2DM communication, for the D2DM communication of cellular network, the subcarrier and power allocation scheme of D2D group and cellular user is in the most general "many-to-many" sharing model, from the perspective of entire cellular network, the joint resource management scheme that all D2D group multicast data rate maximization is decomposed into two sub-targets of subcarrier allocation and power allocation, for subcarrier allocation sub-target, a kind of heuristic subcarrier allocation algorithm is presented, with lower complexity, subcarrier resource is effectively allocated to cellular user and D2D group, for power allocation sub-target, promote the worst SINR condition D2D receiver of each D2D group on the allocated subcarrier, to maximize the total multicast data rate of all D2D group, so as to optimize the multicast data rate of all D2D group by subcarrier and power allocation, without affecting the premise of cellular user service quality, the multicast data rate of all D2D group.

Description

Technical Field

[0001] This invention belongs to the field of 5G communication technology and relates to D2DM communication technology, specifically a joint resource allocation method based on cellular network D2DM communication. Background Technology

[0002] In fifth-generation (5G) wireless communication networks, a major challenge is that the rapid growth of smart devices (such as mobile phones, laptops, and tablets) and the proliferation of mobile applications (such as Facebook, Instagram, and WeChat) will generate diverse types of content and drive a 1000-fold increase in data traffic. This content is likely to be disseminated and shared anytime, anywhere in our daily lives due to people's habit of sharing online with familiar groups and even strangers. Sharing this content simultaneously with a selected group of users via broadband wireless networks is also becoming increasingly common. For example, teachers can distribute course materials to students before class, and team members can conduct collaborative data analysis in the workplace. These requirements of 5G are driving the demand for packet services (i.e., multicast and broadcast). Multicast communication, which leverages the broadcast nature of wireless transmission, can both alleviate data traffic pressure and simultaneously meet the needs of multiple requesters, attracting significant interest from mobile network providers.

[0003] Another challenge for 5G systems is finding ways to offload network traffic. One key technology is Device-to-Device (D2D) communication, standardized by the 3rd Generation Partnership Project (3GPP). D2D communication allows direct communication between user devices that are physically close, reducing the reliance on cellular base stations for content sharing and effectively balancing the traffic load on the central base station. For 5G's key content sharing scenario—where a group of geographically proximate users requests the same content—leveraging the broadcast nature of wireless media to simultaneously deliver popular content to multiple requesters is highly convenient.

[0004] D2D multicast (D2DM) communication combines D2D communication with multicast technology. It enables large-scale parallel implementation of local content sharing in high-density environments, supporting massive connections. Furthermore, it can further reduce redundant transmission, signaling overhead, and bandwidth waste in cellular base stations. Therefore, it can be widely used in 5G cellular networks. For example, by enabling D2DM communication, scenarios such as online teaching, multiplayer games, and multimedia communication can be supported.

[0005] In D2DM communication, the achievable data rate for each D2D group is limited by the worst-case signal-to-interference-plus-noise-ratio (SINR) condition. Specifically, the D2D transmitter sends data packets to all D2D receivers at a common rate that allows resource allocation to be tailored to the D2D receiver with the weakest channel conditions. Firstly, due to the varying locations of D2D receivers and the complex radio environment, even geographically close receivers can experience significant differences in channel conditions. Therefore, we should optimize subcarrier and power allocation to favor the D2D receivers with the worst channel conditions within each D2D group, thereby increasing the multicast data rate for each group. Secondly, since D2DM communication reuses the spectrum resources of cellular users (CUs), it is important to ensure that the impact of D2D transmitters on cellular user performance is negligible or tolerable. Therefore, when introducing D2DM communication into a cellular network, interference from D2D groups should be properly managed, and an effective resource management scheme should be designed to promote the improvement of system performance.

[0006] Existing technologies have extensively researched resource management in D2DM communication, considering different resource sharing scenarios between D2D groups and cellular users, mainly divided into three scenarios. The first scenario allows each D2D group to reuse only one subcarrier, and each subcarrier is allocated to at most one D2D group (i.e., a "one-to-one" sharing model). Kim et al. designed a practical power and subcarrier management scheme for D2D groups, improving the aggregation multicast data rate of D2D groups while ensuring the Quality of Service (QoS) for cellular users. Meshgi et al. used Geometric Programming (GP) for power allocation to users, proposing two subcarrier allocation algorithms (one based on the Hungarian algorithm and the other based on a heuristic algorithm) to improve overall capacity. Bhardwaj et al. raised the problem of improving the total system throughput while ensuring the QoS for cellular users and D2D groups, and proposed a heuristic algorithm for the subcarrier allocation problem, deriving a closed-form solution for power allocation. Jiang et al., in their designed power allocation algorithm, applied the Hungarian algorithm to search for the optimal subcarrier allocation solution between cellular users and D2D groups, maximizing the total energy efficiency (EE) of all D2D groups. Zhao et al. proposed a joint power and subcarrier allocation algorithm to promote the number of successful accesses for D2D groups while minimizing the total energy consumption of both D2D groups and cellular users. However, the first scenario only studied a "one-to-one" sharing model; if multiple D2D groups share the same subcarrier, the spectrum utilization will be further increased. To improve system capacity and spectrum efficiency (SE), the second scenario extends the "one-to-one" sharing model to a more general "one-to-many" scenario, allowing a subcarrier to be shared by multiple D2D groups, while each D2D group can only reuse one subcarrier; the spectrum efficiency is still not high enough. Bhardwaj et al. designed a subcarrier sharing and power allocation algorithm from the perspective of maximizing system capacity. Gong et al. used particle swarm optimization for power control, improving overall throughput while ensuring fairness among D2D groups. However, the spectral efficiency of the second scenario is still not high enough, because it could be further improved if multiple subcarriers were allowed to be reused in each D2D group.

[0007] Compared to the first two scenarios, the third scenario investigates the most general and efficient "many-to-many" sharing model (i.e., a subcarrier can be shared by multiple D2D groups, and each D2D group can reuse multiple subcarriers). Meshgi et al. applied a variant of the Generalized Bender Decomposition (GBD) method to enhance system capacity under all user SINR constraints (see reference: "Optimal Resource Allocation in MulticastDevice-to-Device Communications Underlaying LTE Networks," H. Meshgi, D. Zhao, and R. Zheng, IEEE Transactions on Vehicular Technology, vol. 66, no. 9, pp. 8357–8371, Sep. 2017). However, this method only considers the high SNR region of users and does not take into account subcarrier allocation for cellular users. Furthermore, the GBD method theoretically cannot guarantee an optimal solution in polynomial time; in the worst case, its computational complexity is the same as the exhaustive search method. Bhardwaj et al. employed a power control scheme to achieve an optimal tradeoff between overall energy efficiency and spectral efficiency while ensuring the power outage thresholds for D2D groups and cellular users (see "Energy-and Spectral-Efficiency Tradeoff for D2D-Multicasts in Underlay Cellular Networks," A. Bhardwaj and S. Agnihotri, IEEE Wireless Communications Letters, vol. 7, no. 4, pp. 546–549, Aug. 2018). However, this method only considers resource allocation for D2D groups, which is still not optimal from the perspective of the entire system. Further in-depth research is needed on other performance indicators such as the data rate of D2D groups. Based on the above analysis, the performance improvements brought by D2DM communication in existing wireless cellular networks are not fully utilized. Summary of the Invention

[0008] To address the technical problems existing in the background art, this invention proposes a joint resource allocation method based on D2DM communication in cellular networks. Targeting the emerging application of D2DM communication, this method optimizes the multicast data rate of all D2D groups from the perspective of the entire cellular network by allocating subcarriers and power, without affecting the service quality of cellular users.

[0009] The objective of this invention can be achieved through the following technical solutions:

[0010] A joint resource allocation method based on D2DM communication in a cellular network, wherein the cellular network includes M cellular users and N D2D groups;

[0011] The joint resource allocation method specifically includes the following steps:

[0012] By adopting an equal power allocation scheme, given the initial transmit power of M cellular users and N D2D groups, and on the basis of prioritizing the data rate requirements of cellular users, subcarriers are allocated to cellular users and D2D groups to obtain a subcarrier allocation scheme that maximizes the total multicast data rate of cellular users and D2D groups.

[0013] By adopting the aforementioned subcarrier allocation scheme and adjusting the power allocation scheme for cellular users and D2D groups while prioritizing the data rate requirements of cellular users, a joint resource allocation scheme is obtained that maximizes the total multicast data rate of cellular users and D2D groups.

[0014] Furthermore, the objective of the joint resource allocation scheme is:

[0015]

[0016] An equal power allocation scheme is adopted in target P1. and The objective of obtaining the subcarrier allocation scheme is:

[0017]

[0018] After adopting the aforementioned subcarrier allocation scheme, the binary variables were determined. The objective of the adjusted power allocation scheme is:

[0019]

[0020] In this context, the number of subcarriers in a cellular network is represented by the set K = {1, 2, ..., k, ..., K}, cellular users are represented by the set C = {1, 2, ..., j, ..., M}, and D2D groups are represented by the set D = {1, 2, ..., i, ..., N}; p = {p j ,p i Let} be the set of power vectors for cellular users and D2D groups, and D i D represents the set of D2D receivers belonging to D2D group i. i ={1,2,…,l,…,L},|D i | indicates the total number of receivers in the multicast group. C1 represents the maximum transmit power for cellular users and D2D groups, respectively, and the transmit rate required for cellular user j; C2, C3, and C4 represent the quality of service protection for cellular users, and the power constraints for cellular users and D2D groups, respectively; for binary variables... If user x occupies subcarrier k, then otherwise That is, C5 and C6 indicate that each subcarrier can be occupied by at most one cellular user, and C7 indicates whether group i in D2D occupies subcarrier k.

[0021] Furthermore, the method for allocating subcarriers to cellular users and D2D groups specifically includes:

[0022] Using a power allocation scheme for cellular users and D2D groups, given the initial transmit power of cellular users and D2D groups to satisfy the constraints of C2, C3, and C4 in P1, P1 is transformed into P2.

[0023] Subcarriers are preferentially allocated to cellular users to protect their quality of service requirements, in order to satisfy C1 in P2;

[0024] If there are idle subcarriers that are not allocated, the idle subcarriers are allocated to the D2D group with higher channel gain to improve the data rate of the D2D group.

[0025] For each subcarrier, the D2D group that can achieve a higher data rate is selected to reuse the subcarrier, thereby increasing the capacity gain of the D2D group on the subcarrier while meeting the data rate requirements of cellular users.

[0026] Furthermore, for each idle subcarrier, select The largest D2D group is used for allocation. This indicates that the channel condition of each group so far is limited by the receiver with the worst channel gain.

[0027] Furthermore, the capacity gain ΔR k The following can be calculated:

[0028]

[0029] Where, the right side of the formula represents the data rates of the D2D group added to subcarrier k, the original D2D group interfered with by the newly added D2D group, and the original D2D group not interfered with by the added D2D group, respectively. If ΔR k >0 indicates that multiplexing subcarrier k in D2D group i can bring data rate increment to D2D group.

[0030] Furthermore, the specific methods for adjusting the power allocation scheme for cellular users and D2D groups include:

[0031] After adopting the aforementioned subcarrier allocation scheme, the binary variables were determined. The objective of adjusting the power allocation scheme is transformed into P3;

[0032] Transform the non-concave objective function and constraint C1 in P3 into a concave form;

[0033] Reveal the hidden convexity of target P3 and solve it effectively to achieve fast convergence, thereby improving the multicast data rate of all D2D groups.

[0034] Furthermore, transforming the non-concave objective function and constraint C1 in P3 into a concave form specifically includes:

[0035] By introducing the "two-concave function difference" programming, P3 is transformed into DC programming P4.

[0036]

[0037] The concave function h is handled using a first-order Taylor approximation. i,l (p) and h j (p), approximating the function h at each step i,l (p) and h j The linear form of (p) transforms P4 into P5.

[0038]

[0039] Further, iterative power allocation schemes are generated, and in each iteration, the power allocation solution p is used. t Get h i,l (p) and h j The gradient of (p) is used to calculate the power allocation solution p. t+1 The objective function of P5 is improved through iteration and will eventually converge.

[0040] Furthermore, an auxiliary variable was introduced. Indicates each D2D receiver in The achievable data rate at that time, converting P5 to P6,

[0041]

[0042] P6 is a smooth standard convex optimization objective, which is solved using the interior point method.

[0043] The beneficial effects of this invention are as follows: The joint resource allocation method provided by this invention, for D2DM communication in cellular networks, uses the most common "many-to-many" sharing model for subcarrier and power allocation schemes for D2D groups and cellular users. From the perspective of the entire cellular network, the joint resource management scheme that maximizes the multicast data rate of all D2D groups is decomposed into two sub-objectives: subcarrier allocation and power allocation. For the subcarrier allocation sub-objective, a heuristic subcarrier allocation algorithm is proposed to effectively allocate subcarrier resources to cellular users and D2D groups with low complexity. For the power allocation sub-objective, a method is proposed to promote the worst-case SINR condition D2D receiver for each D2D group on the allocated subcarriers to maximize the total multicast data rate of all D2D groups. Thus, through subcarrier and power allocation, the multicast data rate of all D2D groups is optimized without affecting the service quality of cellular users. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the cellular network architecture of the present invention.

[0045] Figure 2 A performance comparison chart of the HSAPP, RSAPP, HSAEP, RSAEP, and CAFIL solutions.

[0046] Figure 3 A performance comparison chart of the HSAPP, RSAPP, HSAEP, RSAEP, and CAFIL solutions.

[0047] Figure 4 A performance comparison chart of the HSAPP, RSAPP, HSAEP, RSAEP, and CAFIL solutions.

[0048] Figure 5 A performance comparison chart of the HSAPP, RSAPP, HSAEP, RSAEP, and CAFIL solutions.

[0049] Figure 6 A performance comparison chart of the HSAPP, RSAPP, HSAEP, RSAEP, and CAFIL solutions.

[0050] Figure 7 A performance comparison chart of the HSAPP, RSAPP, HSAEP, RSAEP, and CAFIL solutions. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] This invention provides a joint resource allocation method based on cellular network D2DM communication, such as... Figure 1 As shown, the cellular network consists of a very large-scale antenna array equipped with Q antennas, M cellular users, and N D2D groups. Uplink subcarrier resources are shared only with the D2D groups because i) co-channel interference from multiplexed uplink is handled better, and ii) uplink subcarrier resources are typically underutilized compared to downlink. When cellular users can only use orthogonal subcarriers, the number of subcarriers in the cellular network is represented by the set K = {1,2,…,k,…,K}, the cellular users by the set C = {1,2,…,j,…,M}, and the D2D groups by the set D = {1,2,…,i,…,N}. Using D… i D represents the set of D2D receivers belonging to D2D group i. i ={1,2,…,l,…,L}. Let |D i | represents the total number of receivers in the multicast group, when |D i When |=1, because there is only one D2D receiver in the D2D group, the D2DM communication scenario becomes D2D unicast communication. In D2DM communication, each D2D transmitter shares content with multiple D2D receivers in the same group, and each D2D receiver belongs to only one D2D group. For all i ≠ i′.

[0053] Furthermore, from cellular user j to cellular base station The near-field channel can be written as:

[0054]

[0055] In the formula, L and a(r) j,0,l θ j,0,l ) represent the path from cellular user j to the cellular base station and the total number of near-field turning vectors, respectively. j,0,l and θ j,0,l Let w and h represent the distance and angle from cellular user j to the cellular base station, respectively. For XL-MIMO cellular networks, based on the law of large numbers, the maximum ratio combining method can be used, employing optimal receive precoding, which can be expressed as w = h. H Thus, the channel gain from cellular user j to the cellular base station can be expressed as:

[0056]

[0057] Furthermore, based on path loss, fast fading with exponential distribution, slow fading with log-normal distribution, the distance between cellular user j and cellular base station, D2D transmitter, D2D receiver, and the path loss exponent of a single antenna in a D2D group,... and Let be the channel gains from the D2D transmitter of D2D group i to the cellular base station, from the D2D transmitter of D2D group i to the D2D receiver l of D2D group i, from cellular user j to the D2D receiver l of D2D group i, and from the D2D transmitter of D2D group i′ to the D2D receiver l of D2D group i, respectively. Then, the received signal-to-noise ratios (SNRs) for cellular user j at the cellular base station on subcarrier k and for the D2D receiver l of D2D group i are respectively:

[0058]

[0059] in, and Let N0 represent the transmit power of cellular user j on subcarrier k and the transmit power of the D2D transmitter in D2D group i on subcarrier k, respectively. N0 represents additive white Gaussian noise (AWGN). For binary variables... If user x occupies subcarrier k, then otherwise

[0060] For each D2D group, the most typical approach to achieving the data rate is determined by the D2D receiver under worst-case SINR conditions. The worst-case SINR condition for D2D group i on subcarrier k is:

[0061]

[0062] Therefore, the achievable data rates (bit / s / Hz) for cellular user j and D2D group i on subcarrier k are expressed as follows:

[0063]

[0064] Therefore, the achievable data rates for cellular user j and D2D group i are respectively:

[0065]

[0066] The objective of this application is to maximize the multicast data rate of all D2D groups while meeting the basic data rate requirements of cellular users through joint subcarrier and power allocation for D2D groups and cellular users. Therefore, the optimized objective of this application is stated as follows:

[0067]

[0068] In target P1, p = {p j ,p i} represents the set of power vectors for cellular users and D2D groups. Let C1 be the maximum transmit power for cellular users and C2 be the transmit rate required for D2D groups. C2, C3, and C4 are the quality of service protection for cellular users and power constraints for cellular users and D2D groups, respectively. C5 and C6 indicate that each subcarrier can be occupied by at most one cellular user, and C7 indicates whether D2D group i occupies subcarrier k. Objective P1 contains not only continuous variables but also discrete variables, making it difficult to solve directly.

[0069] Proposition 1: When the optimal solution is obtained, constraint C1 in P1 must satisfy the equation.

[0070] Proof: Given the optimal subcarrier allocation solution ρ * Suppose that P1 has a condition that satisfies The optimal power solution p * ; However, cellular users can always reduce their transmit power. j* ,make This means that cellular users always reduce their transmit power within the required transmission rate range to reduce interference with D2D groups, thereby increasing the data rate of D2D groups. Therefore, there exists a solution p that satisfies P1(ρ * ,p)>P 1(ρ * ,p * This violates the assumption.

[0071] Note 1: If not properly configured and Considering relevant parameters, P1 may be infeasible. In this invention, we adopt the reasonable assumption that P1 is always feasible.

[0072] Given that P1 requires joint power and subcarrier allocation for all D2D groups and cellular users, this leads to a Mixed Integer Non-Linear Programming (MINLP) problem. To solve this problem efficiently, this application divides the complex objective into two sub-objectives: subcarrier allocation and power allocation.

[0073] For subcarrier allocation, although some algorithms (such as Benders decomposition and branch-and-bound methods) can find the optimal allocation result, they cannot theoretically guarantee that they will obtain the optimal solution in polynomial time. In the worst case, their computational complexity is comparable to that of exhaustive search. Therefore, this application proposes a heuristic subcarrier allocation scheme from the perspective of the entire cellular network to allocate subcarriers to cellular users and D2D groups.

[0074] To allocate subcarriers to cellular users and D2D groups, an equal power allocation scheme for cellular users and D2D groups is first adopted, which can be represented as follows: and Given the initial transmit power for cellular users and D2D groups, and to automatically satisfy constraints C2, C3, and C4 in P1, the original objective is transformed into:

[0075]

[0076] It can be seen that due to the existence of binary variables P2 still struggles to obtain the optimal solution. To make the objective solution more efficient, this application proposes a heuristic subcarrier assignment (HSA) algorithm (Algorithm 1).

[0077] The basic idea of ​​Algorithm 1 is to allocate subcarriers to cellular users and D2D groups while meeting the data rate requirements of cellular users, thereby maximizing the total multicast data rate for both cellular users and D2D groups. First, subcarriers are allocated to cellular users to protect service quality requirements. Then, to improve the data rate of D2D groups, idle subcarriers are allocated to D2D groups with higher channel gain. Furthermore, the reuse of cellular user subcarriers by D2D groups (“many-to-many” sharing) can further improve spectral efficiency and explore the benefits of underlying D2DM communication.

[0078] A heuristic subcarrier allocation algorithm:

[0079] Step 1: Initialization.

[0080] Let R j =0 and

[0081] Let R i =0 and

[0082] set up and

[0083] Step 2: Subcarrier allocation that satisfies constraint C1.

[0084] 1. Find j = argmaxΔRj ,

[0085] 2. Search

[0086] 3. Assign subcarrier k to cellular user j and update R. j A j =A j ∪{k},S k =S k ∪{j},V k =V k ∪{j}, K=K-{k}.

[0087] 4. Repetition.

[0088] 5. Until For all 1≤j≤M.

[0089] Step 3: Assign subcarriers to D2D groups.

[0090] 6. If

[0091] 7. Perform the following for each k∈K

[0092] 8. Search

[0093] 9. Assign subcarrier k to D2D group i and update R. i F i =F i ∪{k},T k =T k ∪{i},V k =V k ∪{i}, K=K-{k}.

[0094] 10. For k * =1:K

[0095] 11. Calculate according to equation (6)

[0096] 12. Searching

[0097] 13. [The following appears to be a separate, unrelated sentence:] Put k * Assign the value to group i in D2D * And recalculate R j (if any)

[0098] 14. If And ΔR k >0

[0099] 15. Subcarrier k* Assign to group i in D2D * ,renew R j , D = D - {i *}

[0100] 16. Repeat

[0101] 17. Until

[0102] 18. End

[0103] Symbol A j F i S k T k and V k Let represent the set of subcarriers allocated to cellular user j. The subcarrier set allocated to D2D group i includes the set of cellular users using subcarrier k, the set of D2D groups reusing subcarrier k, and the sets of cellular users and D2D groups sharing subcarrier k respectively. k is the set of unallocated subcarriers. The data rate demand gap for cellular user j (j∈C) is:

[0104]

[0105] Specifically, Algorithm 1 consists of two phases. In the first phase (lines 1-5), the goal is to protect the needs of cellular users. In each iteration, the cellular user has the maximum ΔR. j It is necessary to select a favorable subcarrier to meet its quality of service requirements. Thus, after this period, all cellular users... It can be guaranteed that constraint C1 will be satisfied. In the second phase, if there are still some idle subcarriers not allocated to users after the first phase, they are allocated to the D2D group (lines 6-9). For each idle subcarrier, select... The largest D2D group is used for allocation. This indicates that the channel conditions for each group so far are limited by the receiver with the worst channel gain. Next, the benefits of D2D communication are further utilized (lines 10-17), where the "many-to-many" sharing mode can significantly improve spectral efficiency. For each subcarrier, the D2D group i that can achieve a higher data rate is selected. * This allows for the reuse of subcarriers without violating the data rate requirements of cellular users, and can improve the capacity gain of D2D groups on subcarrier k.

[0106] Capacity gain ΔR k (Line 17) is defined as the improvement in data rate on subcarrier k for D2D groups due to spectrum sharing. If ΔR k>0 indicates that multiplexing subcarrier k in D2D group i can bring an increase in data rate to the D2D group. Capacity gain ΔR k The following can be calculated:

[0107]

[0108] Note that the right side of equation (13) represents the data rates of the D2D group added to subcarrier k, the original D2D group interfered with by the newly added D2D group, and the original D2D group not interfered with by the added D2D group, respectively.

[0109] After effective subcarrier allocation design, the binary variables were determined. Next, the power allocation sub-objective can be expressed as:

[0110]

[0111] The properties and difficulties in solving P3 are analyzed from two main aspects. First, the objective function is non-smooth. Second, it should be noted that the objective function and constraint C1 are non-concave with respect to p. Therefore, P3 is neither smooth nor concave, making it difficult to solve.

[0112] To address this issue, this application first introduces the "Difference of Two Concave Functions" (DC) programming, which transforms the non-concave objective function and constraint C1 in objective P3 into a concave form. Then, it reveals the latent convexity of objective P3 and solves it effectively, resulting in fast convergence and significantly improving the multicast data rate of all D2D groups.

[0113] Specifically, first, the non-concave objective function and constraint C1 are rearranged into a concave objective function, and the objective function of P3 is rewritten in DC form:

[0114]

[0115] For any D2D receiver l belonging to D2D group i, the following equation can be obtained:

[0116]

[0117] Among them, f i,l (p) and h i,l (p) are two concave functions, which can be calculated as:

[0118]

[0119] Therefore, the above formula can be rewritten as:

[0120]

[0121] In the formula, z is the user index of the cellular user and the D2D group, i.e., z∈Z={1,2,…M,M+1,…,M+N}. Similarly, the non-concave constraint C1 can be rewritten in the following DC form:

[0122]

[0123] Furthermore, f j (p) and h j (p) can be given separately:

[0124]

[0125] Then rewrite equation (21):

[0126]

[0127] It can be verified that f i,l (p), h i,l (p), f j (p), h j (p) is a concave function of p. Through the above transformation, the original power allocation sub-objective P3 is converted into a DC planning function:

[0128]

[0129] constraints of P4 C4 and C4 are linear, therefore concave. Due to the objective function and constraints... For non-concave functions, P4 still cannot be solved directly. To find a tractable solution, this application uses the first-order Taylor approximation to handle concave functions h. i,l (p) and h j (p). h i,l (p) and h j The gradients of (p) at any p are as follows:

[0130]

[0131] In the formula, It is a K(M+N) dimensional column vector. otherwise Similarly, otherwise,

[0132] Next, using a first-order Taylor expansion, the function h is approximated at each step. i,l (p) and h j The linear form of (p) is used to generate iterative power allocation. Equations (20) and (23) are transformed into:

[0133]

[0134] Under the above approximation, the non-concave type can be transformed into a concave type. Therefore, the power allocation sub-objective can be rewritten as:

[0135]

[0136] All constraints in P5 are concave functions. However, the objective function of P5 remains non-smooth. Transforming the non-smooth optimization objective into a smooth one reveals the hidden convexity of P5. In particular, an auxiliary variable is introduced. Indicates each D2D receiver in The achievable data rate. Target P5 can be converted to:

[0137]

[0138] Note the objective function on page 6. C4 is a linear function. C5 is a concave function. Therefore, objective P6 is a smooth, standard convex optimization objective, which can be solved using the interior point method.

[0139] In summary, the power allocation method can be summarized as a proposed power (PP) allocation algorithm (Algorithm 2).

[0140] A proposed power allocation algorithm:

[0141] 1. Input: Subcarrier allocation solution of Algorithm 1.

[0142] 2. Output: Power allocation for cellular users and D2D groups.

[0143] 3. Initialization

[0144] Let t = 0, and let p = 0.

[0145] ·calculate

[0146] 4. Calculation and

[0147] 5. t = t + 1.

[0148] 6. In p t-1 Solution P 5 gives p t .

[0149] 7. Update separately and

[0150] 8. Calculate

[0151] 9. Repeat.

[0152] 10. Until |I t -I t-1 |≤ε

[0153] In each iteration, h is obtained using the power allocation solution pt. i,l (p) and h j The gradient of (p) is used to calculate the power allocation solution p. t+1 Therefore, we can obtain a lower bound for the objective function proved by the following proposition 2.

[0154] Proposition 2: Algorithm 2 generates an improved sequence of feasible solutions for P6 through iteration.

[0155] Proof: For constraints Because h j (p) is concave, so h j The gradient of (p) is indeed its hypergradient, that is:

[0156]

[0157] Therefore, if the inequality If it is established, then This holds true. Therefore, the solution generated in each iteration p is valid for the constraints. It's always possible.

[0158] Similarly, based on P5 and Algorithm 2, we can obtain:

[0159]

[0160] The first inequality holds because the local optimal solution for P4 is p. t+1 Therefore, p t+1 Substituting P5 into the equation yields a value greater than or equal to the value of pt. Furthermore, due to h... i,l The concavity of (p) also means that the second inequality holds, where the gradient... It is its hypergradient. Specifically, for any given p...

[0161]

[0162] Therefore, the objective function of P5 is improved through iteration under these constraints and will eventually converge.

[0163] Specific examples

[0164] The cellular network parameters are shown in Table 1. Cellular users and D2D groups are uniformly and randomly distributed within a cell with a radius of 500m, and the maximum distance between the D2D transmitter and the D2D receiver is 30m. First, the total data rate of the D2D groups is compared with some parameters, including the number of cellular users, the number of D2D groups, the number of subcarriers, and the data rate requirements of the cellular users. Then, the convergence of the proposed method is presented.

[0165] Table 1 Relevant Simulation Parameter Values

[0166]

[0167] In the simulation, the performance of the following five resource allocation schemes will be compared first:

[0168] ·HSAPP: The heuristic subcarrier allocation scheme proposed in Algorithm 1 and the power allocation scheme proposed in Algorithm 2;

[0169] • RSAPP: A random subcarrier allocation and power allocation scheme based on Algorithm 2;

[0170] • HSAEP: Heuristic subcarrier allocation and equal power allocation scheme in Algorithm 1;

[0171] • RSAEP: Random Subcarrier Allocation and Equal Power Allocation Scheme;

[0172] • CAFIL: The CAFIL scheme addresses a problem similar to that of this invention, namely, maximizing the aggregated multicast data rate of all D2D receivers. This algorithm assumes that each D2D receiver can reuse at most one cellular base station subcarrier, considering only the subcarrier allocation and power allocation for D2D receivers, while the cellular base station subcarriers are predetermined. For fairness, we apply the subcarrier allocation scheme to the cellular base station according to Algorithm 1.

[0173] Wireless networks typically employ a random strategy. The random subcarrier allocation method first allocates the subcarriers with the best channel gain to the cellular base station to meet its quality of service (QoS) requirements. Then, any unused subcarriers (if any) are randomly allocated to D2D receivers. These subcarriers are then randomly selected for reuse by the D2D receivers without violating the QoS of the cellular base station or improving capacity gain.

[0174] The HSAPP scheme proposed in this invention outperforms the other four benchmark schemes. In particular, the aggregation rate of the D2D receiver monotonically decreases as the number of cellular base stations increases. This is because the more cellular base stations in the network, the more spectrum resources need to be allocated to these base stations to ensure their data rate requirements. Furthermore, the power allocation algorithm can significantly improve the aggregation data rate of the D2D receiver, as the scheme proposed in this invention and RSAEP outperform power allocation schemes such as HSAEP and RSAEP. Moreover, compared to the "one-to-one" sharing scenario of the CAFIL scheme, the "many-to-many" sharing methods of HSAPP, RSAPP, HSAEP, and RSAEP have superior performance because the D2D receiver can reuse more subcarriers.

[0175] When there is only one D2D receiver, resources can only be allocated to that single receiver, resulting in HSAEP and RSAEP having the same performance under the same power control scheme, and thus the same for both schemes. However, as the number of D2D receivers increases, HSAPP and HSAEP, employing the heuristic subcarrier allocation (HSA) algorithm, respectively outperform the random subcarrier allocation strategies of RSAPP and RSAEP. This is because, using our proposed HSA algorithm, higher data rate D2D receivers have a greater chance of acquiring resources.

[0176] From the perspective of the entire system, as the number of subcarriers increases, the D2D receiver can use more subcarriers, which can significantly improve the data rate of the D2D receiver. Furthermore, the HSAPP scheme proposed in this invention is superior to other benchmark methods.

[0177] The HSAPP scheme proposed in this invention exhibits the best performance, with the total data rate of the D2D receiver decreasing as the data rate demand of the cellular base station increases. This is because as the data rate requirement increases, more spectrum resources and greater power need to be allocated to the cellular base station to meet the required data rate. Therefore, the aggregated data rate of the D2D receiver is affected by greater interference and reduced spectrum resources.

[0178] Algorithm 2 converges faster and increases the objective function, as proven in Proposition 2, where each cellular base station eventually converges to the basic data rate requirement described in Proposition 1. Furthermore, the convergent solution increases as the maximum distance between the D2D transmitter and the D2D receiver decreases. This is because the D2D receiver exhibits better channel conditions and higher data rates over shorter communication distances, better reflecting the characteristics of D2D communication.

[0179] The above description is merely an example and illustration of the structure of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the structure of the invention or exceed the scope defined in the claims, all of which should fall within the protection scope of the present invention.

Claims

1. A joint resource allocation method based on cellular network D2DM communication, characterized in that, The cellular network comprises M cellular users and N D2D groups; The joint resource allocation method specifically includes the following steps: By adopting an equal power allocation scheme, given the initial transmit power of M cellular users and N D2D groups, and on the basis of prioritizing the data rate requirements of cellular users, subcarriers are allocated to cellular users and D2D groups to obtain a subcarrier allocation scheme that maximizes the total multicast data rate of cellular users and D2D groups. By adopting the aforementioned subcarrier allocation scheme and adjusting the power allocation scheme for cellular users and D2D groups while prioritizing the data rate requirements of cellular users, a joint resource allocation scheme is obtained that maximizes the total multicast data rate of cellular users and D2D groups.

2. The joint resource allocation method according to claim 1, characterized in that, The objective of the joint resource allocation scheme is: An equal power allocation scheme is adopted in target P1. and The objective of obtaining the subcarrier allocation scheme is: After adopting the aforementioned subcarrier allocation scheme, the binary variables were determined. The objective of the adjusted power allocation scheme is: In this context, the number of subcarriers in a cellular network is represented by the set K = {1, 2, ..., k, ..., K}, cellular users are represented by the set C = {1, 2, ..., j, ..., M}, and D2D groups are represented by the set D = {1, 2, ..., i, ..., N}; p = {p j ,p i Let} be the set of power vectors for cellular users and D2D groups, and D i D represents the set of D2D receivers belonging to D2D group i. i ={1,2,…,l,…,L},|D i | indicates the total number of receivers in the multicast group. C1 represents the maximum transmit power for cellular users and D2D groups, respectively, and the transmit rate required for cellular user j; C2, C3, and C4 represent the quality of service protection for cellular users, and the power constraints for cellular users and D2D groups, respectively; for binary variables... If user x occupies subcarrier k, then otherwise That is, C5 and C6 indicate that each subcarrier can be occupied by at most one cellular user, and C7 indicates whether group i in D2D occupies subcarrier k.

3. The joint resource allocation method according to claim 2, characterized in that, The specific methods for allocating subcarriers to cellular users and D2D groups include: Using a power allocation scheme for cellular users and D2D groups, given the initial transmit power of cellular users and D2D groups to satisfy the constraints of C2, C3, and C4 in P1, P1 is transformed into P2. Subcarriers are preferentially allocated to cellular users to protect their quality of service requirements, in order to satisfy C1 in P2; If there are idle subcarriers that are not allocated, the idle subcarriers are allocated to the D2D group with higher channel gain to improve the data rate of the D2D group. For each subcarrier, the D2D group that can achieve a higher data rate is selected to reuse the subcarrier, thereby increasing the capacity gain of the D2D group on the subcarrier while meeting the data rate requirements of cellular users.

4. The joint resource allocation method according to claim 3, characterized in that, For each idle subcarrier, select The largest D2D group is used for allocation. This indicates that the channel condition of each group so far is limited by the receiver with the worst channel gain.

5. The joint resource allocation method according to claim 3, characterized in that, The capacity gain ΔR k The following can be calculated: Where, the right side of the formula represents the data rates of the D2D group added to subcarrier k, the original D2D group interfered with by the newly added D2D group, and the original D2D group not interfered with by the added D2D group, respectively. If ΔR k >0 indicates that multiplexing subcarrier k in D2D group i can bring an increase in data rate to D2D group.

6. The joint resource allocation method according to claim 2, characterized in that, The specific methods for adjusting the power allocation scheme for cellular users and D2D groups include: After adopting the aforementioned subcarrier allocation scheme, the binary variables were determined. The objective of adjusting the power allocation scheme is transformed into P3; Transform the non-concave objective function and constraint C1 in P3 into a concave form; Reveal the hidden convexity of target P3 and solve it effectively to achieve fast convergence, thereby improving the multicast data rate of all D2D groups.

7. The joint resource allocation method according to claim 6, characterized in that, Transforming the non-concave objective function and constraint C1 in P3 into a concave form specifically includes: By introducing the "two-concave function difference" programming, P3 is transformed into DC programming P4. The concave function h is handled using a first-order Taylor approximation. i,l (p) and h j (p), approximating the function h at each step i,l (p) and h j The linear form of (p) transforms P4 into P5. Further, an iterative power allocation scheme is generated, and in each iteration, the power allocation solution p is used. t Get h i,l (p) and h j The gradient of (p) is used to calculate the power allocation solution p. t+1 The objective function of P5 is improved through iteration and will eventually converge.

8. The joint resource allocation method according to claim 7, characterized in that, An auxiliary variable was introduced. This indicates that each D2D receiver is in The achievable data rate at that time, converting P5 to P6, P6 is a smooth standard convex optimization objective, which is solved using the interior point method.

Images