Method for resource allocation of uav-d2d multi-relay offloading mec system based on energy collection

Abstract

The application provides a resource allocation method of an energy collection-based UAV-D2D multi-relay offloading MEC system, comprising the following steps: S1: modeling a network structure of a multi-relay offloading system of UAV-D2D; S2: modeling a total task transmission time slot allocation structure; S3: modeling D2D link transmission time and relay forwarding link transmission time; S4: modeling relay forwarding data energy consumption and unmanned aerial vehicle wireless energy transmission time; S5: modeling edge server computing task time; S6: modeling a total sum of D2D link transmission, relay forwarding link transmission and edge computing time; S7: modeling preset constraints of user transmission power, energy collection, computing resources, channel resources and link rate; S8: modeling an optimization model of minimizing total task completion time of the system; S9: solving the optimization model of minimizing total task completion time of the system; and the application of the technical solution can realize minimization of total task completion time of the system.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of wireless communication technology and mobile edge computing technology, in particular to a resource allocation method of an energy collection-based UAV-D2D multi-relay offloading MEC system. BACKGROUND

[0002] With the rapid development of modern network communication technology, people's requirements for network performance are also improving accordingly. In view of the growing mobile service demand of the Internet of Things, the device-to-device communication network assisted by unmanned aerial vehicles (UAVs) has attracted much attention. In a resource-limited system, UAVs have become an important part of mobile communication networks due to their easy deployment and high mobility; D2D communication technology can effectively improve the spectrum efficiency and reduce the burden of base stations in dense communication, and is a key technology for realizing massive access in future mobile communication systems.

[0003] The introduction of MEC into the UAV-assisted D2D system can not only save local computing resources and reduce battery consumption, but also significantly improve the computing power of mobile devices. In addition, radio frequency energy collection is a technology that converts received radio frequency signals into electrical energy, providing continuous and stable energy for user devices with its new green power supply method. Through a reasonable resource allocation strategy, the device task computing time can be effectively reduced. Therefore, it is of great significance to study the MEC resource allocation strategy of UAV-assisted D2D communication.

[0004] However, most existing UAV-assisted D2D relay forwarding systems often consider the problem of a single relay node forwarding a single task or multiple tasks. However, the communication resource burden of the relay node itself is ignored, resulting in a significant reduction in the battery life of the relay node and rapid aging of the device hardware. SUMMARY

[0005] Therefore, the purpose of the present application is to provide a resource allocation method of an energy collection-based UAV-D2D multi-relay offloading MEC system to solve the problems of inability to achieve task offloading due to a long distance from the base station and heavy communication burden of the relay in the system. The transmission power, energy collection time, channel resources and computing resources are jointly optimized to minimize the total task completion time of the system.

[0006] To achieve the above purpose, the present application adopts the following technical solution: a resource allocation method of an energy collection-based UAV-D2D multi-relay offloading MEC system, comprising the following steps:

[0007] Step S1: modeling the network structure of the multi-relay offloading system of UAV-D2D;

[0008] Step S2: modeling the total task transmission time slot allocation structure;

[0009] Step S3: modeling the transmission time of the D2D link and the transmission time of the relay forwarding link;

[0010] Step S4: modeling the energy consumption of the relay forwarding data and the transmission time of the UAV wireless energy;

[0011] Step S5: modeling the time of the edge server computing task;

[0012] Step S6: modeling the sum of the D2D link transmission, the relay forwarding link transmission and the edge computing time;

[0013] Step S7: modeling the preset constraints of the user transmission power, energy collection, computing resources, channel resources and link rate;

[0014] Step S8: modeling the optimization model of minimizing the total task completion time of the system;

[0015] Step S9: using the alternating iterative optimization algorithm based on Padé approximation to solve the optimization model of minimizing the total task completion time of the system, obtaining the optimal allocation result of the transmission power, energy collection time, computing resources and channel resources, and substituting the optimal allocation result into the function of minimizing the total task completion time of the system to obtain the total task completion time of the system.

[0016] In a preferred embodiment, the step S1 is specifically:

[0017] Step S11: constructing a UAV-D2D multi-relay offloading network structure, including a base station carrying a MEC server, U UAVs deployed at the radiation edge of the BS, each UAV covering a GNs cluster, and each GNs cluster consisting of R u groups of D2D pairs, wherein The UAV provides wireless power for the DR in the cluster through wireless energy transmission (WPT), and performs D2D communication through centralized D2D control; the DT is located outside the radiation edge of the BS, and the DR is located in the area jointly covered by the UAV and the BS;

[0018] Step S12: the flight height of all UAVs is H1, and the position coordinates of the u-th UAV are q u = [x u ,y u ,H1]; it is assumed that each group of D2D pairs in the cluster covered by the u-th UAV consists of one DT and R u ,···,R U}, The DT and the k-th DR of the r-th group of D2D pairs in the u-th cluster are respectively and The UAV hovers and the user position is known, the coordinates of the ground base station are M=[x M ,y M ,H M ]; the distance between the UAV of the cluster u and the kth DR paired in the rth group of D2D pairs of the cluster is The distance between the kth DR paired in the rth group of D2D pairs of the cluster u and the BS is The distance between the DT in the rth group of D2D pairs of the cluster u and the kth DR is

[0019] In a preferred embodiment, the step S2 is specifically: constructing a total task transmission time slot allocation structure; the total task completion time is divided into U cluster task completion times, and the task of each cluster includes three parts, which are transmission part, calculation part and download part; wherein, it is assumed that the required transmission data bit result of the download part is very small and is ignored in time. In the transmission part, it is divided into three stages, which are energy transmission, D2D transmission and relay forwarding three stages.

[0020] In a preferred embodiment, the step S3 is specifically:

[0021] Step S31:

[0022] The transmission rate of the kth D2D link in the rth group of D2D pairs in the uth cluster is calculated according to the following formula

[0023]

[0024] Wherein, B u,r represents the subchannel resource allocated by the rth group of D2D pairs in the uth cluster; represents the transmission power of the DT of the rth group of D2D pairs in the uth cluster to the kth DR; σ 2 represents the system noise power;

[0025] The transmission time of the rth group of data tasks in the D2D stage in the uth cluster is calculated according to the following formula

[0026]

[0027] Wherein, represents the transmission time of each DR receiving data;

[0028] Since the FDMA transmission mode is adopted, the D2D stage transmission time of the uth cluster is calculated according to the following formula

[0029]

[0030] Step S32:

[0031] In the relay forwarding phase, the DRs receiving data as relay devices are forwarded one by one; according to the following formula, the forwarding transmission rate of each DR is calculated

[0032]

[0033] wherein, P u, r, k represents the transmission power of the kth DR of the rth group of D2D pairs in the uth cluster;

[0034] According to the following formula, the total transmission time of the relay forwarding phase in the uth cluster is calculated

[0035]

[0036] wherein, P u, r, k represents the transmission time of each DR forwarding data; since the performance of the DF relay system is limited to the channel condition of the two links of the source node to the relay and the relay to the target node; therefore, the transmission rate of the kth offloading transmission link of the rth DT in the uth cluster is

[0037] In a preferred embodiment, the step S4 is specifically: according to the following formula, the energy collected by each DR is calculated

[0038]

[0039] wherein, η represents the energy collection efficiency of the DR, P u, r, k represents the energy collection time of each DR, P u P u represents the transmission power of the uth UAV;

[0040] According to the following formula, the energy consumption of the kth DR of the rth group of D2D pairs in the uth cluster is calculated

[0041]

[0042] The D2D transmission phase and the energy collection phase are carried out at the same time, and the relay needs to complete the collection of the required energy before the end of the D2D phase.

[0043] In a preferred embodiment, the step S5 is specifically: according to the following formula, the task calculation time of each task is calculated

[0044]

[0045] wherein C u,r P u, r represents the number of CPU machine cycles required by the rth user task in the uth cluster to calculate its original data; the calculation task bit number and the required CPU machine cycle number satisfy the linear relationship C u,r = w u,r Iu,r where w u,r The table is to calculate the number of machine cycles required per bit;

[0046] The calculation completion time of all tasks in the u-th cluster is calculated according to the following formula

[0047]

[0048] In a preferred embodiment, the step S6 is specifically: the D2D stage transmission time of the u-th cluster is The total transmission time of the relay forwarding stage in the u-th cluster is represented as: The calculation completion time of all tasks in the u-th cluster is: The total task completion time is calculated according to the following formula

[0049]

[0050] In a preferred embodiment, the step S7 is specifically: modeling the constraints of transmission power, energy collection, computing resources, channel resources and link rate

[0051] The power allocation constraint is:

[0052] The energy collection constraint is:

[0053] The computing resource allocation constraint is:

[0054] The channel resource constraint is:

[0055] The link rate constraint is:

[0056] Where, P max represents the maximum transmission power of the GNs; F represents the total computing resources of the base station edge server; B represents the total channel resources of the system; R min represents the minimum link rate threshold.

[0057] In a preferred embodiment, the step S8 is specifically: under the constraint conditions of transmission power, energy collection, computing resources, channel resources and link rate, the optimization resource allocation strategy is determined by taking the minimum system total task completion time as the goal, that is

[0058] In a preferred embodiment, the step S9 is specifically: using an alternating iterative optimization algorithm based on Pad Approximation, solving the system total task completion time minimization optimization model to obtain the optimal allocation result of the transmission power, energy collection time, computing resource and channel resource, and substituting the optimal allocation result into the system total task completion time minimization function to obtain the system total task completion time, and the specific steps are:

[0059] S91: First, restate the maximum constraint problem by using the mirror image reconstruction technology to introduce auxiliary variables and respectively represent the upper bounds of and in the objective function, and the auxiliary variable represents the lower bound of ;

[0060] S92: Decouple part of the variables by using the alternating iterative method; divide the resource optimization problem into two sub-problems, the joint optimization of GNs transmission power and sub-channel bandwidth resource problem and the joint optimization of energy collection time and computing resource problem; and solve the new resource allocation problem by using the alternating optimization method

[0061] auxiliary variable restriction condition:

[0062] auxiliary variable restriction condition:

[0063] auxiliary variable X u,r,k restriction condition: X u,r,k ≥R min .

[0064] S93: Joint optimization of GNs transmission power and sub-channel bandwidth resource problem; due to the variable coupling relationship in the constraint term, the sub-problem is non-convex, and the auxiliary variable is introduced to restate it as P1 using the convexity of the perspective function:

[0065] auxiliary variable λ u,r,k restriction condition: 0≤λ u,r,k ≤B u,r P max ,

[0066]

[0067] Therefore, the sub-problem is a convex optimization problem;

[0068] S94: Joint optimization of energy collection time and computing resource; the energy collection time allocation and the computing resource allocation are optimized by the given transmission power control and sub-channel bandwidth resource allocation of GNs; this sub-problem is expressed as P2:

[0069] This sub-problem is a convex optimization problem;

[0070] S95: Alternating iterative optimization; since the objective functions of the two sub-problems and their related constraint conditions are convex, the two sub-problems are convex optimization problems, and the respective sub-problems are solved by the convex solvers CVX and CVXQUAD; firstly, in the first solving, suitable energy collection time and computing resource initial values and iteration times are given, and the optimal GNs transmission power and sub-channel bandwidth resource are obtained by solving the sub-problem 1; secondly, the optimal GNs transmission power and sub-channel bandwidth resource solved in the foregoing are used to solve the sub-problem 2, so that the optimal energy collection time and computing resource initial value are obtained; by continuously cyclically iterating the two solving steps, the minimum total task completion time can be obtained when the optimization target converges.

[0071] Compared with the prior art, the present application has the following beneficial effects:

[0072] 1. The present application adopts the D2D user cooperation mode to solve the task offloading problem of the user far from the base station.

[0073] 2. The present application splits the task data to multiple relays for offloading, and under the guarantee of the relay energy consumption and the demand of task computing service, the problem of large relay communication burden in the system is alleviated.

[0074] 3. The non-convex optimization problem of the optimization model of the present application is difficult to be directly solved. The present application introduces a series of auxiliary variables, converts some non-convex constraint conditions into convex constraint conditions by using the perspective function, adopts the alternating iterative optimization algorithm based on the Pad approximation, and realizes the minimization of the total task completion time of the system. BRIEF DESCRIPTION OF DRAWINGS

[0075] Figure 1 is the network structure diagram of the UAV-D2D multi-relay offloading MEC of the preferred embodiment of the present application;

[0076] Figure 2 is the total task completion time slot allocation diagram of the preferred embodiment of the present application;

[0077] Figure 3 is the resource allocation algorithm flowchart based on alternating iteration of the preferred embodiment of the present application;

[0078] Figure 4 is the method flowchart of the preferred embodiment of the present application. DETAILED DESCRIPTION

[0079] The application will be further described below in conjunction with the accompanying drawings and embodiments.

[0080] It should be noted that the following detailed description is exemplary in nature and is intended to provide further description of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0081] It should be noted that the terms used herein are merely for the purpose of describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application; as used herein, the singular form is intended to include the plural form unless the context clearly indicates otherwise, and it should also be understood that when the terms "comprise" and / or "include" are used in the specification, they indicate the presence of the features, steps, operations, devices, components and / or combinations thereof.

[0082] The present application aims at the problem that devices outside the edge of the BS cannot achieve task offloading and the problem that the relay communication burden is large, and proposes a resource allocation strategy for minimizing the total task completion time of a UAV-D2D multi-relay offloading MEC system based on energy collection. Remote users collect less energy from the base station, and cannot offload data to the MEC server of the base station with greater transmission power, thereby suffering from the influence of the double near-far problem. By deploying UAVs to centrally control each group of D2D pairs and configuring multiple DRs to realize serial communication, task data is split and offloaded, and each DR collects the radio frequency energy of the UAV for forwarding tasks that need to be offloaded for calculation from outside the edge. Thus, the problems of remote base station tasks that cannot be offloaded and large relay communication burden in the system are alleviated.

[0083] I. Network model of non-linear energy collection MEC system assisted by UAV

[0084] A resource allocation strategy for a UAV-D2D multi-relay offloading MEC system based on energy collection is proposed in this embodiment, and the network model of the UAV-D2D multi-relay offloading MEC is as shown in Figure 1 In this system, a base station carrying a MEC server, U UAVs are deployed at the radiation edge of the BS, and each UAV covers a cluster of GNs. The UAVs supply power wirelessly to the DRs in the cluster through wireless energy transmission, and perform D2D communication through centralized D2D control. The DTs are located outside the radiation edge of the BS, and the DRs are located in the area jointly covered by the UAVs and the BS. The completion of the tasks of each cluster needs to go through three parts, namely the transmission part, the calculation part and the download part. Among them, it is assumed that the required transmission data bit result of the download part is very small and is ignored in time. Figure 2A time slot allocation diagram for each cluster user to complete the computing task. In the transmission part, there are three stages, namely energy transmission stage, D2D stage and forwarding stage, and the energy collection stage is completed before the end of the D2D stage, and the next forwarding stage can be carried out. When the forwarding stage is over, the computing part can be carried out.

[0085] II. Establishing a resource allocation model for a UAV-assisted nonlinear energy harvesting MEC system

[0086] Please refer to Figure 4 The resource allocation strategy of the UAV-D2D multi-relay offloading MEC system based on energy harvesting proposed in the embodiment includes the following steps:

[0087] S1: Modeling the network structure of the UAV-D2D multi-relay offloading system

[0088] The system network structure mainly includes a base station, a UAV, and a D2D user equipment. In this embodiment, a three-dimensional Cartesian coordinate system is adopted, and the flight height of all UAVs is H1. The position coordinates of the u-th UAV are q u = [x u ,y u ,H1]. The DT of the r-th D2D pair in the u-th cluster and the position of the k-th DR are and The coordinates of the ground base station are M = [x M ,y M ,H M ]. The distance between the UAV of the u-th cluster and the k-th DR paired with the r-th D2D pair in the cluster is The distance between the k-th DR paired with the r-th D2D pair in the u-th cluster and the BS is The distance between the DT in the r-th D2D pair in the u-th cluster and the k-th DR is

[0089] S2: Modeling the total task transmission time slot allocation structure

[0090] The total task transmission time slot allocation structure is constructed. The total task completion time is divided into U cluster task completion times, and each cluster task includes three parts, namely the transmission part, the computing part and the download part. Among them, it is assumed that the required transmission data bit result of the download part is very small and is ignored in time. In the transmission part, there are three stages, namely energy transmission, D2D transmission and relay forwarding.

[0091] S3: Modeling the D2D link transmission time and relay forwarding link transmission time

[0092] Step 1:

[0093] The transmission rate of the kth D2D link in the rth group of D2D pairs in the u th cluster is calculated according to the following formula

[0094]

[0095] where B u,r represents the allocated sub-channel resource of the rth group of D2D pairs in the u th cluster. represents the transmission power of the kth DR of the rth group of D2D pairs in the u th cluster to the kth DT. 2 represents the system noise power.

[0096] The transmission time of the rth group of data tasks in the D2D stage in the u th cluster is calculated according to the following formula

[0097]

[0098] where, represents the transmission time of each DR receiving data.

[0099] Since the FDMA transmission mode is adopted, the transmission time of the D2D stage of the u th cluster is calculated according to the following formula

[0100]

[0101] Step 2: In the relay forwarding stage, the DR receiving data forwards one by one as a relay device.

[0102] The forwarding transmission rate of each DR is calculated according to the following formula

[0103]

[0104] where, represents the transmission power of the kth DR of the rth group of D2D pairs in the u th cluster.

[0105] The total transmission time of the relay forwarding stage in the u th cluster is calculated according to the following formula

[0106]

[0107] where, represents the transmission time of each DR forwarding data. Since the performance of the DF relay system is limited to the channel with poor channel conditions in the two links of the source node to the relay and the relay to the target node. Therefore, the transmission rate of the kth unloading transmission link of the rth DT in the u th cluster is

[0108] S4: Modeling relay forwarding data energy consumption and energy collection time

[0109] The energy collected by each DR is calculated according to the following formula

[0110]

[0111] where η represents the energy collection efficiency of the DR, P represents the energy collection time of each DR, u Puis the transmission power of the u-th UAV.

[0112] The energy consumption of the k-th DR of the r-th group of D2D pairs in the u-th cluster is calculated according to the following formula

[0113]

[0114] The D2D transmission phase is carried out simultaneously with the energy collection phase, and the relay needs to complete the collection of the required energy before the end of the D2D phase.

[0115] S5: Modeling the time of the edge server computing task

[0116] The computing time of each task is calculated according to the following formula

[0117]

[0118] where C u,r represents the number of CPU machine cycles required by the r-th user task in the u-th cluster to compute its original data. The number of computing task bits and the number of required CPU machine cycles satisfy a linear relationship C u,r = w u,r I u,r , where w u,r is the number of machine cycles required to compute each bit.

[0119] The computing completion time of all tasks within the u-th cluster is calculated according to the following formula

[0120]

[0121] S6: Modeling the sum of D2D link transmission, relay forwarding link transmission and computing time

[0122] The D2D phase transmission time of the u-th cluster is: The total transmission time of the relay forwarding phase in the u-th cluster is represented as: The computing completion time of all tasks within the u-th cluster is: The total task completion time is calculated according to the following formula

[0123]

[0124] S7: Modeling the preset constraints of transmission power, energy collection, computing resources, channel resources and link rate

[0125] The power allocation constraint condition is:

[0126] The energy collection constraint condition is:

[0127] The computing resource allocation constraint condition is:

[0128] The channel resource constraint condition is:

[0129] The link rate constraint condition is: Wherein, P max represents the maximum transmission power of the GNs; F represents the total computing resource of the base station edge server; B represents the total channel resource of the system; R min represents the minimum link rate threshold.

[0130] S8: Modeling the optimization model of minimizing the total task completion time of the system

[0131] Under the constraint conditions of transmission power, energy collection, computing resource, channel resource and link rate, the optimization resource allocation strategy is determined with the objective of minimizing the total task completion time of the system, that is, III. Solving the resource allocation optimization model based on an alternating iteration algorithm

[0132] Referring to Figure 3 , in this embodiment, an alternating iteration optimization algorithm based on Pad Approximation is used to solve the optimization model of minimizing the total task completion time of the system, to obtain the optimal allocation result of user transmission power, energy collection time, computing resource and channel resource, and to substitute the optimal allocation result into the function of minimizing the total task completion time of the system to obtain the total task completion time of the system. Specifically,

[0133] Step 1: First, the maximum constraint problem is re-expressed by the mirror image reconstruction technique to introduce auxiliary variables and represent the upper bounds of and in the objective function, and auxiliary variables represent the lower bounds of .

[0134] Step 2: Decouple part of the variables by the alternating iteration method. The resource optimization problem is divided into two sub-problems, the joint optimization of GNs transmission power and sub-channel bandwidth resource problem and the joint optimization of energy collection time and computing resource problem. The new resource allocation problem is solved by alternating and iteration through the alternating optimization method

[0135] Auxiliary variable Tu DD Constraints:

[0136] Auxiliary variable T u C Constraints:

[0137] Auxiliary variable X u,r,k Constraints: X u,r,k ≥ R min .

[0138] Step 3: Joint optimization of GNs transmit power and subchannel bandwidth resource problem. Due to the existence of variable coupling relationship in the constraint term, resulting in this subproblem is non-convex, by introducing auxiliary variable Using the convexity preserving of perspective function to restate as P1:

[0139] Auxiliary variable λ u,r,k Constraints: 0 ≤ λ u,r,k ≤ B u,r P max ,

[0140]

[0141] Therefore, this subproblem is a convex optimization problem.

[0142] Step 4: Joint optimization of energy harvesting time and computing resource. Through the given GNs transmit power control and subchannel bandwidth resource allocation, the energy harvesting time allocation and computing resource allocation are optimized. This subproblem is stated as P2:

[0143] This subproblem is a convex optimization problem.

[0144] Step 5: Alternating iterative optimization. Since the objective function of the two subproblems and their related constraint conditions are convex, the two subproblems are convex optimization problems, and the respective subproblems are solved by convex solver CVX and CVXQUAD. First, in the first solving, the appropriate energy harvesting time and computing resource initial value and iteration number are given, which are used to solve the subproblem 1 to obtain the optimal GNs transmit power and subchannel bandwidth resource. Second, the optimal GNs transmit power and subchannel bandwidth resource solved in the previous step are used to solve the subproblem 2, so as to obtain the optimal energy harvesting time and computing resource initial value. By constantly cycling the two solving steps, the optimization target converges, that is, the minimum total task completion time can be obtained.

[0145] Finally, it is to be understood that the application can be carried out by specifically different embodiments, and it is expressly intended that the claimed application should not be limited to the particulardescribed embodiments set forth in the above description as such may include any variations and modifications falling within the scope of the application.

Claims

1.A method for resource allocation of an energy-harvesting based UAV-D2D multi-relay offloading MEC system, characterized in that, The method comprises the following steps: Step S1: modeling a multi-relay offloading system network structure of UAV-D2D; Step S2: modeling a total task transmission time slot allocation structure; Step S3: modeling D2D link transmission time and relay forwarding link transmission time; Step S4: modeling relay forwarding data energy consumption and UAV wireless energy transmission time; Step S5: modeling edge server computing task time; Step S6: modeling D2D link transmission, relay forwarding link transmission and edge computing time sum; Step S7: modeling preset constraints of user transmission power, energy collection, computing resource, channel resource and link rate; Step S8: modeling an optimization model of system total task completion time minimization; Step S9: adopting an alternating iteration optimization algorithm based on Padé approximation, solving the system total task completion time minimization optimization model, obtaining optimal allocation results of transmission power, energy collection time, computing resource and channel resource, and substituting the optimal allocation results into the system total task completion time minimization function to obtain the system total task completion time. 2.The method of Claim 1, wherein, The step S1 is specifically: Step S11: Constructing a UAV-D2D multi-relay offloading network structure, including a base station carrying a MEC server, U UAVs deployed at the radiation edge of the BS, each UAV covering a GNs cluster, and each GNs cluster consisting of R u D2D pairs are formed, where u∈U=={1,2,···,U}; the UAVs provide wireless power supply for the DRs in the cluster through wireless energy transmission (WPT), and the D2D communication is carried out through centralized D2D control; the DTs are located outside the radiation edge of the BS, while the DRs are located in the area jointly covered by the UAVs and the BS; Step S12: All UAVs are flying at altitude H1, and the position coordinates of the u-th UAV are q. u =[x u ,y u [H1]; Assume that each D2D pair in the cluster under the coverage of the u-th UAV consists of a DT and a K r u It consists of DRs, where R=={R1,···,R u ,···,R U }, r∈R u =={1,2,···,R u }, k∈K u,r =={1,2,···,K u,r The DT and DR positions of the r-th D2D pair in the u-th cluster are respectively and The UAV is hovering and the user's location is known. The coordinates of the ground base station are M = [x M ,y M H M The distance between the drone in cluster u and the k-th DR paired in the r-th D2D pair of the cluster is... The distance between the k-th DR paired in the r-th D2D pair of cluster u and the BS is The distance between the DT and the k-th DR in the r-th D2D pair of cluster u is 3.The resource allocation method for energy-harvesting based UAV-D2D multi-relay offloading MEC system according to claim 1, wherein, The step S2 is specifically: constructing a total task transmission time slot allocation structure; the total task completion time is divided into U cluster task completion times, and each cluster task includes three parts, namely a transmission part, a computing part and a download part; wherein, it is assumed that the required transmission data bit result of the download part is very small and is ignored in time; in the transmission part, there are three stages, namely energy transmission, D2D transmission and relay forwarding three stages. 4.The method of Claim 1, wherein, The step S3 is specifically: Step S31: According to the following formula, the transmission rate of the kth D2D link in the rth D2D pair in the uth cluster is calculated wherein B u,r denotes the sub-channel resource allocated to the rth group of D2D pairs in the u-th cluster; denotes the transmit power of the DT of the rth group of D2D pairs in the u-th cluster to the kth DR; σ 2 denotes the system noise power; According to the following formula, the transmission time of the rth data task in the D2D stage in the uth cluster is calculated wherein, denotes the transmission time of each DR receiving data; Since the FDMA transmission mode is adopted, according to the following formula, the D2D stage transmission time of the uth cluster is calculated Step S32: In the relay forwarding stage, the DR receiving data is forwarded one by one as a relay device; According to the following formula, the forwarding transmission rate of each DR is calculated wherein, Pur,kdenotes the transmit power of the kth DR of the rth group of D2D pairs in the u-th cluster. According to the following formula, the total transmission time of the relay forwarding stage in the uth cluster is calculated wherein, denotes the transmission time of the data forwarded by each DR; since the performance of the DF relay system is limited to the channel with worse channel condition among the two links from the source node to the relay and from the relay to the destination node; thus the transmission rate of the kth offloaded transmission link of the rth DT in the u th cluster is 5.The resource allocation method for energy-harvesting based UAV-D2D multi-relay offloading MEC system according to claim 1, wherein, The step S4 is specifically: according to the following formula, the energy collected by each DR is calculated wherein η represents the energy collection efficiency of the DR, represents the energy collection time of each DR, P u represents the transmission power of the u-th UAV; According to the following formula, the energy consumption of the kth DR of the rth D2D pair in the uth cluster is calculated The D2D transmission stage and the energy collection stage are carried out at the same time, and the relay needs to complete the collection of the required energy before the end of the D2D stage. 6.The resource allocation method of the energy-harvesting based UAV-D2D multi-relay offloading MEC system according to claim 1, wherein, The step S5 is specifically: according to the following formula, the computing time of each task is calculated where C u,r represents the number of CPU machine cycles required by the rth user task in the u th cluster to compute its raw data; the number of task bits and the number of CPU machine cycles required to compute satisfy a linear relationship C u,r = w u,r I u,r where w u,r is a table of the number of machine cycles required to compute each bit; According to the following formula, the computing completion time of all tasks in the uth cluster is calculated 7. The method of Claim 1, wherein, The step S6 is specifically: the D2D stage transmission time of the u-th cluster is: The total transmission time of the relay forwarding stage in the u-th cluster is expressed as: The calculation completion time of all tasks in the u-th cluster is: The total task completion time is calculated according to the following formula: 8.The method of Claim 1, wherein, The step S7 is specifically: modeling the limitation conditions of transmission power, energy collection, computing resource, channel resource and link rate The power allocation restriction condition is: Energy harvesting constraints are: The computing resource allocation restriction condition is: The channel resource restriction condition is: The link rate limit condition is: where P max denotes the maximum transmit power of GNs; F denotes the total computing resource of the base station edge server; B denotes the total channel resource of the system; R min denotes the minimum link rate threshold. 9.The resource allocation method of the energy-harvesting based UAV-D2D multi-relay offloading MEC system according to claim 1, wherein, The step S8 specifically includes: determining an optimized resource allocation strategy with the minimum system total task completion time as the target under the constraint conditions of the transmission power, the energy collection, the calculation resource, the channel resource and the link rate, that is 10.The method of claim 1, wherein, The step S9 is specifically: using an alternating iterative optimization algorithm based on Padé approximation, solving a system total task completion time minimization optimization model, obtaining optimal allocation results of the transmission power, the energy collection time, the computing resource and the channel resource, and substituting the optimal allocation results into a system total task completion time minimization function to obtain the system total task completion time, and the specific steps are: S91: First, reformulate the min-max constraint problem by introducing auxiliary variables through mirror image reformulation technique and represent the upper bounds of and in the objective function, respectively, and introduce auxiliary variables represent the lower bounds of in the objective function; S92: decoupling part of variables by using an alternating iterative method; The resource optimization problem is divided into two sub-problems, joint optimization of GNs transmit power and sub-channel bandwidth resource problem and joint optimization of energy collection time and computing resource problem; the new resource allocation problem is solved by alternating optimization method Auxiliary variable T u DD Restriction: Auxiliary variable T u C Restriction: auxiliary variable X u,r,k Restriction: X u,r,k ≥ R min ; S93: Joint optimization of GNs transmit power and subchannel bandwidth resource problem; due to the existence of variable coupling relationship in the constraint term, resulting in the non-convexity of this sub-problem, by introducing auxiliary variables By using the convexity preserving of perspective function to restate as P1: Auxiliary variable λ u,r,k Constraint: 0 ≤ λ u,r,k ≤ B u,r P max , Therefore, the sub-problem is a convex optimization problem; S94: Joint optimization of energy harvesting time and computing resource; the energy harvesting time allocation and computing resource allocation are optimized by the given transmission power control of GNs and sub-channel bandwidth resource allocation; this sub-problem is formulated as P2: The sub-problem is a convex optimization problem; S95: alternating iterative optimization; since the objective functions of the two sub-problems and the related constraint conditions are convex, the two sub-problems are convex optimization problems, and the respective sub-problems are solved by using a convex solver CVX and CVXQUAD; first, in the first solving, appropriate energy collection time and computing resource initial values and the number of iterations are assigned, the optimal GNs transmission power and sub-channel bandwidth resource are obtained by solving the sub-problem 1; second, the optimal GNs transmission power and sub-channel bandwidth resource solved in the foregoing are used to solve the sub-problem 2, so that the optimal energy collection time and computing resource initial value are obtained; by continuously and repeatedly iterating the two solving steps, the optimization target converges, and the minimum total task completion time can be obtained.

Images