Resource allocation modeling method for SWIPT-assisted NOMA-MEC system

By using genetic algorithms to optimize the transmission power and time switching coefficients in the SWIPT-assisted NOMA-MEC system, the problems of spectrum shortage and energy limitation are solved, and the efficient utilization of energy resources in the system is achieved, and efficient scenarios such as smart cities and smart homes are supported. run.

CN114521023BActive Publication Date: 2024-11-12BEIJING UNIV OF POSTS & TELECOMM
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210166464.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-02-23
Publication Date
2024-11-12
Estimated Expiration
2042-02-23

AI Technical Summary

Technical Problem

The existing SWIPT-assisted NOMA-MEC system cannot effectively optimize system performance and energy collection in the case of shortage of spectrum resources and limited energy of smart terminal equipment, and cannot be applied to new application scenarios such as smart cities and smart homes.

Method used

A joint energy harvesting and task offload optimization scheme based on genetic algorithm is proposed. By maximizing the total number of system offload tasks as the goal, combining continuous convex approximation, Lagrangian multiplication method and gradient descent method, it is decomposed into power allocation and time switching Coefficient optimization problem, build an iterative power allocation algorithm and resource allocation algorithm to optimize the transmission power and time switching coefficients.

Benefits of technology

It realizes efficient energy collection and data unloading of smart terminal equipment under the conditions of spectrum shortage and energy limitation, improves the system's spectrum utilization rate and efficient utilization of energy resources, and supports the efficient operation of new application scenarios such as smart homes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114521023B_ABST
    Figure CN114521023B_ABST
Patent Text Reader

Abstract

The present invention discloses a resource allocation modeling method for a SWIPT-assisted NOMA-MEC system. Aiming at new application scenarios such as smart homes and smart cities, the service quality requirements of smart terminal devices, the maximum transmission power of terminal devices and the limited available energy are considered, and a problem of jointly optimizing the transmission power and the time switching coefficient with the goal of maximizing the total number of unloaded tasks in the system is proposed. In order to solve the modeled mixed integer programming problem containing multiple variables, the problem is decomposed into two sub-problems: a power allocation problem when the time switching coefficient is given and an optimization problem of the time switching coefficient. An iterative power allocation algorithm is constructed by using continuous convex approximation, Lagrange multiplier method and gradient descent method. A resource allocation algorithm based on genetic algorithm is constructed to jointly optimize the transmission power and the time switching coefficient, so as to obtain a joint resource allocation model of energy collection and data unloading of smart terminal devices, and ensure efficient utilization of energy resources in the SWIPT-assisted NOMA-MEC system.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of mobile device computing, and in particular to a SWIPT-assisted NOMA-MEC system resource allocation modeling method. Background Art

[0002] In the context of smart cities, tens of billions of IoT devices are expected to be deployed to automate social operations and make the surrounding environment smarter. Due to the deep integration of mobile Internet and IoT, new application scenarios such as smart homes are emerging rapidly, accelerating the pace of urban intelligence. In such new applications, the computing power and available energy of smart terminals are limited, making it difficult to effectively support the new application scenarios emerging in smart cities.

[0003] In order to effectively solve the problem of limited computing power of mobile devices, scholars have proposed the concept of Mobile Edge Computing (MEC). MEC provides computing, storage, network and communication resources at the edge of the mobile network. Because it is close to the user, it can reduce the latency of network operation and service delivery and improve the user service experience. At the same time, mobile devices can offload computing tasks to edge servers, which will also extend the battery life of mobile devices, thereby ensuring the service quality of mobile devices. Several systems proposed based on the MEC system are as follows:

[0004] 1. SWIPT-assisted MEC system

[0005] In terms of powering devices, the traditional battery-powered method has the problem that the mobile application will terminate after the battery is exhausted. Although this problem can be overcome by increasing the battery or charging the battery regularly, it will increase the hardware cost and cause a lot of inconvenience. In response to the huge energy consumption in mobile communication systems, scholars have proposed the Simultaneous Wireless Information and Power Transfer (SWIPT) technology, which can make full use of the characteristics of radio frequency signals that carry both data information and electromagnetic energy, and perform data transmission and energy harvesting (Energy Harvesting, EH) at the same time, thereby extending the online operation life of communication equipment.

[0006] In order to overcome the low service quality caused by the energy limitation and low computing power of smart terminal devices, scholars integrated SWIPT and MEC technology and conducted research on resource allocation and computing offloading solutions in SWIPT-assisted MEC systems. The SWIPT-assisted MEC system consists of an access point (AP), multiple single-antenna low-power devices and a MEC server. The AP sends control signals and energy to the device at the same time through SWIPT. The device decodes the control information by collecting energy and performs local calculations or offloads tasks to the MEC server.

[0007] 2. MEC system based on NOMA

[0008] Considering the shortage of spectrum resources in wireless communications, scholars use non-orthogonal multiple access technology in MEC networks to simultaneously offload tasks, thereby achieving the goal of improving spectrum efficiency, reducing energy consumption, and avoiding task transmission delays. NOMA uses the transmitter's superposition coding technology and the receiver's serial interference elimination technology to ensure that multiple devices can communicate efficiently in the same frequency resource blocks at the same time, thereby increasing the user capacity of the communication system and improving the system's spectrum efficiency.

[0009] 3. SWIPT-assisted NOMA system

[0010] The SWIPT-assisted NOMA system combines SWIPT technology with NOMA technology. Multiple devices use the same time-frequency resource blocks to transmit information, and use SWIPT technology to collect energy from signals sent by base stations, thereby improving the system's spectrum efficiency, reducing energy consumption, and extending the service life of terminal equipment.

[0011] There are certain application defects in the above systems, specifically:

[0012] 1. SWIPT-assisted MEC system

[0013] Existing work mainly focuses on the MEC scenario of Orthogonal Multiple Access (OMA). Devices use orthogonal communication resources to offload tasks to MEC servers, and reduce system energy consumption or increase the success probability of system computing offload by optimizing offloading decisions and resource allocation. Therefore, it is only applicable to scenarios with sufficient spectrum resources, and cannot be applied to large-scale access scenarios with insufficient spectrum resources.

[0014] 2. MEC system based on NOMA

[0015] Using non-orthogonal multiple access technology to offload tasks simultaneously in MEC networks can achieve the goal of improving spectrum efficiency, reducing energy consumption, and avoiding task transmission delays. However, most of the existing resource allocation and computation offloading schemes in NOMA-MEC systems only consider the constraints of transmission rate and transmission power, ignoring the limited energy of smart terminals in new application scenarios such as smart cities and smart homes, and lacking joint optimization research on system performance and energy collection.

[0016] 3. SWIPT-assisted NOMA system

[0017] In the SWIPT-assisted NOMA system, most of them only consider the energy consumption and interference problems during communication, without considering the task calculation problems of terminal devices. It is not suitable for scenarios where the computing power of terminal devices is limited in new applications such as smart cities and smart homes.

[0018] Therefore, the present invention adopts NOMA technology to perform computational offloading in the SWIPT-assisted MEC system. Compared with the orthogonal multiple access method, the non-orthogonal multiple access technology can serve more users under the condition of the same communication resources. The SWIPT-assisted NOMA-MEC system can not only avoid serious delays, but also reduce energy consumption and improve spectrum utilization. However, due to the influence of different user service quality requirements, new spectrum access modes and available energy constraints, the existing resource allocation scheme cannot be directly applied to new scenarios. Therefore, it is urgent to design a new resource allocation scheme for the SWIPT-assisted NOMA-MEC system based on constraints such as service quality requirements and energy constraints. Summary of the invention

[0019] The purpose of the present invention is to provide a SWIPT-assisted NOMA-MEC system resource allocation modeling method, which jointly considers the system resource limitation and terminal device energy limitation problems, and proposes a genetic algorithm-based joint energy collection and task offloading optimization scheme in the SWIPT-assisted NOMA-MEC scenario with the goal of maximizing the total number of unloaded tasks in the system. In the case of spectrum shortage, limited available energy of equipment, and limited computing resources, this scheme can ensure that smart terminal devices use the energy collected from access points to efficiently offload computing tasks to MEC servers, thereby supporting the efficient operation of new application scenarios such as smart homes.

[0020] In order to achieve the above object, the present invention provides the following technical solutions:

[0021] The present invention provides a SWIPT-assisted NOMA-MEC system resource allocation modeling method, comprising the following steps:

[0022] S1. Considering the service quality requirements of intelligent terminal devices, the maximum transmission power of terminal devices and the limited available energy, the problem of jointly optimizing the transmission power and time switching coefficient is proposed with the goal of maximizing the total number of offloaded tasks in the system;

[0023] S2. In order to solve the modeled mixed integer programming problem with multiple variables, the problem is decomposed into two sub-problems: the power allocation problem when the time switching coefficient is given and the optimization problem of the time switching coefficient;

[0024] S3, construct an iterative power allocation algorithm using continuous convex approximation, Lagrange multiplier method, and gradient descent method;

[0025] S4. Construct a resource allocation algorithm based on genetic algorithm to jointly optimize the transmission power and time switching coefficient, so as to obtain a joint resource allocation model of energy collection and data offloading for smart terminal devices.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] The SWIPT-assisted NOMA-MEC system resource allocation modeling method of the present invention jointly considers the actual bottlenecks in communication in new application scenarios such as smart cities and smart homes, namely, the limited system resources and the limited energy and computing power of terminal devices, while the research of the prior art only considers part of the problems. Due to the influence of different user service quality requirements, new spectrum access modes and available energy constraints, the existing resource allocation scheme cannot be directly applied to new scenarios. Therefore, in order to solve the problems of many computing tasks, high energy consumption during communication and spectrum shortage in smart terminal devices, the present invention proposes to maximize the total number of offloaded tasks in the system, model the revenue optimization problem of the SWIPT-assisted NOMA-MEC system as a mixed integer programming problem with multiple variables, and decompose the problem into two sub-problems: the power allocation problem when the time switching coefficient is given and the optimization problem of the time switching coefficient. An iterative power allocation algorithm is proposed by using continuous convex approximation, Lagrange multiplier method and gradient descent method. On this basis, a resource allocation algorithm based on genetic algorithm for jointly optimizing the transmission power and the time switching coefficient is proposed, so as to ensure the efficient utilization of energy resources in the SWITP-assisted NOMA-MEC system, obtain an efficient energy collection and data offloading joint scheduling model for smart terminal devices, and ensure the efficient utilization of energy resources in the SWIPT-assisted NOMA-MEC system. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0029] Figure 1 A diagram of the SWITP-assisted NOMA-MEC system architecture provided in an embodiment of the present invention;

[0030] Figure 2 A SWIPT solution based on time switching is provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0031] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0032] The SWIPT-assisted NOMA-MEC system resource allocation modeling method proposed in the present invention, which is applicable to the SWIPT-assisted NOMA-MEC system, comprises the following steps:

[0033] First, for new application scenarios such as smart cities and smart homes, NOMA technology is used in SWIPT-assisted MEC systems by considering the limited spectrum resources, energy and computing power of terminal devices in actual communications. In order to study the benefit optimization problem of SWIPT-assisted NOMA-MEC systems, the service quality requirements of smart terminal devices, the maximum transmission power of terminal devices, and the limited available energy are jointly considered, and a mixed integer programming problem is modeled with the goal of maximizing the total number of offloaded tasks in the SWIPT-assisted NOMA-MEC system.

[0034] Secondly, in order to effectively solve the optimization problem of maximizing the total number of system offload tasks, the mixed integer programming problem of jointly optimizing the transmit power and the time switching coefficient is decomposed into two sub-problems: the power allocation problem when the time switching coefficient is given and the optimization problem of the time switching coefficient. The power allocation problem when the time switching coefficient is given is a non-convex problem.

[0035] Again, in the power allocation sub-problem, an iterative power allocation algorithm is proposed using continuous convex approximation, Lagrange multiplier method and gradient descent method.

[0036] Finally, an efficient genetic algorithm-based method for jointly optimizing resource allocation and computational offloading of transmit power and time switching coefficient is proposed to ensure efficient utilization of energy resources in SWIPT-assisted NOMA-MEC systems.

[0037] 1. System model

[0038] The present invention adopts a wireless energy transmission assisted non-orthogonal multiple access mobile edge computing (SWIPT assisted NOMA-MEC) system, such as Figure 1 As shown in the figure, the system consists of an AP, K smart terminal devices and an MEC server. Due to the limited computing power of smart terminal devices, it is assumed that the devices choose to offload computationally intensive, delay-sensitive and indivisible tasks to the MEC server. Each device adopts a decoding and forwarding mode and uses the SWIPT technology based on time slot switching (TS) to receive energy and control information from the AP. The goal of the system is to offload part of the local computing tasks of the terminal device to the MEC server. Assume that all devices are equipped with one antenna, and use k∈{1,2,…,K} to represent the smart terminal device. In addition, it is assumed that the system operates in a time slot manner and the channel is quasi-static, that is, the channel coefficient remains unchanged within each time slot. The channel is modeled as a Rayleigh fading channel, and the channel response obeys an independent complex Gaussian distribution. Therefore, the channel gain can be expressed as where ζ represents the signal attenuation coefficient related to distance, d i,j represents the distance between transmitter i and receiver j, and γ is the path loss exponent.

[0039] In the above system, T is used to represent the duration of each time slot, such as Figure 2 As shown in the figure. It is assumed that each device is able to use TS-SWIPT technology for information decoding and energy collection. Therefore, each time slot is divided into three phases. The first phase is the energy transmission from AP to device, the second phase is information decoding (ID), and the third phase is to offload the task to the MEC server. We assume that the duration of energy collection is (1-α k -β k )T, where each device can collect the energy of the RF signal from the AP and use it to complete the tasks in the next two stages: information decoding and computing task offloading. k The information is decoded within the duration of T, and k Unload tasks to the MEC server within T time, where 0<β k <1. When β k = 0, it means that there is no task offloaded to the MEC server. Without loss of generality, assume that the time slot T is normalized to 1.

[0040] 1) Energy harvesting model: In the energy harvesting phase, each device k -β k )T time to collect energy from AP. Let the AP’s transmit power be P s , then the energy collected by device k is expressed as:

[0041]

[0042] Where 0<η<1 represents the energy conversion efficiency of the device, h s,k Indicates the channel gain between device k and AP.

[0043] 2) Information decoding model: In the information decoding stage, the AP transmits the control message to each device and uses the Time Division Multiple Access (TDMA) protocol to handle the calculation and offloading tasks. The device uses part of the collected energy for information decoding. The data rate that device k can achieve is expressed as:

[0044]

[0045] Where N0 represents the additive white Gaussian noise at the device, h s,k Indicates the channel gain between device k and AP.

[0046] The power consumption of the decoding process can be approximated as a linear function of the data rate. Therefore, the energy consumption of each device during the information decoding phase can be expressed as:

[0047]

[0048] Where ε (Joule / bit) is a constant based on the energy harvesting decoder.

[0049] 3) Offloading model: Assume that the AP has global channel state information (CSI) and makes decisions on the offloading process. After the information decoding stage, the terminal device uses the collected energy to offload the task to the MEC server. Assume that the channel bandwidth is unit bandwidth and the channel gain of each device satisfies h1≥…≥h k ≥…≥h K , where h k Represents the channel gain between device k and the MEC server.

[0050] In the NOMA scheme, the device with higher channel gain is decoded first, and the signal sent by the device with lower channel gain is treated as interference. Therefore, the achievable transmission rate of device k to offload the computing task is:

[0051]

[0052] Among them, p k Indicates the transmit power of device k.

[0053] When the task is unloaded, the TDMA protocol is used, and each device uses the same unloading time, that is, each device The task is offloaded to the MEC server within a certain time. The energy consumption of each device during the offloading process is:

[0054]

[0055] In addition, we assume that the MEC server has high computing power, so we ignore the task calculation and return process of the MEC server. We assume that the size of the tasks is the same and normalized to 1. Therefore, at t k The number of bits that device k can offload in time is:

[0056] L k =t k R k (12)

[0057] 2. Problem Statement

[0058] Based on the above equations, our goal is to maximize the sum of the offload tasks of all terminal devices in the system (referred to as problem (7)). The maximization problem should take into account the quality of service requirements, the limited collected energy, and the maximum transmission power constraint. Without loss of generality, it is assumed that the minimum transmission rate of each device is equal. Therefore, the optimization problem modeled is as follows:

[0059]

[0060] Among them, constraint (7b) represents the quality of service requirement of the minimum transmission rate of each device. Constraint (7c) ensures that the energy consumed by each device in decoding information and unloading tasks does not exceed the energy it collects. Constraint (7d) ensures that the transmission power of each device when unloading tasks does not exceed the maximum transmission power. (7e) gives the feasibility constraint of the time switching coefficient.

[0061] Problem (7) is a joint optimization problem. k , β k and p k The mixed integer programming problem with equal parameters is an NP-hard problem and is difficult to solve. Since the optimization of the time switching coefficient of each device can be directly performed on the energy harvesting device, it is independent of the transmission power allocation problem. Therefore, for the above optimization problem, we use a resource allocation algorithm based on a genetic algorithm (GA) consisting of two sub-algorithms.

[0062] 3. Problem decomposition

[0063] In the invention, we propose a new resource allocation algorithm based on genetic algorithm to solve the joint optimization problem of time switching coefficient and transmission power. In order to solve problem (7), we decompose the problem into two sub-problems. First, when the time switching coefficient is fixed, we establish a mathematical model to solve the optimal transmission power of each device. On this basis, we use genetic algorithm to obtain the optimal time switching coefficient.

[0064] a) Optimization of transmission power

[0065] In this section, we discuss the transmission power p involved in (7) under the assumption that the time switching coefficient of device k is known. k Calculation:

[0066]

[0067] Due to the influence of the rate function, problem (8) is obviously a non-convex problem. Applying the continuous convex approximation method, the rate function based on the signal-to-noise ratio is approximated as:

[0068]

[0069]

[0070]

[0071]

[0072] The parameters The initial value of corresponds to the initialization of system parameters.

[0073] Therefore, the problem can be simplified to:

[0074]

[0075] Obviously, all transmission powers p k The linear constraints on p are all convex. By proving that the objective function k The Hessian matrix on is positive semidefinite, which determines the convexity of the objective function. Then, we use the classic Lagrange multiplier method to solve the optimization problem. Based on the above description, the Lagrange function is defined as follows:

[0076]

[0077] where λ k , μ k and θ k is a Lagrange multiplier that satisfies λ k≥0,μ k ≥0 and θ k ≥ 0. Based on the Lagrange dual method, problem (9) is transformed into:

[0078]

[0079] It can be decomposed into two optimization problems. The inner problem is to maximize the Lagrangian function value, and the outer problem is to minimize the dual gap. According to the KKT condition, the optimal transmission power can be obtained:

[0080]

[0081] where [x] + =max{0,x}. In the outer loop, the Lagrange multiplier is updated according to the gradient descent principle:

[0082]

[0083]

[0084] θ k (l+1)=[θ k (l)+s3(l)(p max -p k )] + (15)

[0085] Where l is the number of iterations, s1(l), s2(l) and s3(l) are positive iteration steps.

[0086] In order to ensure the convergence of the algorithm, this parameter needs to be carefully designed. Then an iterative power allocation algorithm is proposed, as shown in Algorithm 1.

[0087]

[0088] b) Optimization of time switching coefficient

[0089] First, repeat Algorithm 1 to obtain the optimal transmission power allocation scheme. Based on this, the present invention proposes a resource allocation algorithm based on genetic algorithm (Algorithm 2). The resource allocation algorithm based on genetic algorithm is divided into six stages: initialization, evaluation, selection, crossover, mutation and evolution. In genetic algorithm, the selection of fitness function and the design of genetic operator are very critical. Therefore, we define the fitness function of each individual i as the sum of unloaded tasks:

[0090]

[0091] Among them, the transmission power is obtained by Algorithm 1 We use the roulette strategy to determine the adaptive crossover and mutation process in natural selection. The specific process of Algorithm 2 is:

[0092] Initialization: Initialize the feasible population P0 with a population size of N p , the maximum number of generations of the population is N g We set the number of chromosomes N of an individual c =2, corresponding to the time switching coefficient α k , β k .

[0093] Evaluation: Based on the transmit power obtained by Algorithm 1 We calculate the fitness function value F(i) of each individual in the population using formula (16).

[0094] Selection: Select the individual with the highest fitness function value F among the parents. Then let the parents enter the next generation according to the roulette strategy.

[0095] Crossover: With a certain crossover probability, two parents are selected for single-point crossover to obtain two new offspring, which are then included in the next generation population.

[0096] Mutation: Select a certain number of individuals for mutation operation. Randomly select one chromosome of the individual for mutation. This will produce a new generation of population.

[0097] Evolution: Calculate the fitness function value F(i+1) of each individual in the new generation population again according to the fitness function (16). If |F(i+1)-F(i)|<Δ, it means that the function converges, and the maximum fitness function value is selected as the optimal fitness function value.

[0098] At this time, the individual corresponds to the optimal time switching coefficient α k and β k Otherwise, repeat the above steps until the algorithm converges or reaches the maximum number of iterations N. g .

[0099]

[0100]

[0101] The technology of the present invention aims to solve the problems of multiple computing tasks, high energy consumption during communication and spectrum shortage of intelligent terminal devices. With the goal of maximizing the total number of offload tasks of the system, the profit optimization problem of the SWIPT-assisted NOMA-MEC system is modeled as a mixed integer programming problem. On this basis, a resource allocation scheme based on genetic algorithm for jointly optimizing the transmission power and time switching coefficient is proposed, thereby ensuring the efficient utilization of energy resources in the SWITP-assisted NOMA-MEC system.

[0102] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0103] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, electronic device embodiment, computer-readable storage medium embodiment, and computer program product embodiment, since they are basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0104] The above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the aforementioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the aforementioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical specialties therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application. They should all be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A SWIPT-assisted NOMA-MEC system resource allocation modeling method, characterized in that: The SWIPT-assisted NOMA-MEC system consists of an AP, K smart terminal devices and an MEC server, wherein the smart terminal device chooses to offload computationally intensive, delay-sensitive and indivisible tasks to the MEC server. Each device adopts a decoding and forwarding mode and uses the SWIPT technology based on time slot switching to receive energy and control information from the AP. The system works in a time slot manner and the channel is quasi-static, that is, the channel coefficient remains unchanged in each time slot. The channel is modeled as a Rayleigh fading channel, and the channel response obeys an independent complex Gaussian distribution. Each time slot is divided into three stages: the first stage is energy transmission from AP to device, the second stage is information decoding, and the third stage is to offload tasks to the MEC server. The device is in α k The information is decoded within the duration of T, and k Unload tasks to the MEC server within T time, where 0<β k <1, T represents the duration of each time slot, and the time slot T is normalized to 1; in the energy collection phase, each device is in (1-α k -β k )T, assuming the AP’s transmit power is P s , then the energy collected by device k is expressed as: Where 0<η<1 represents the energy conversion efficiency of the device, h s,k represents the channel gain between device k and AP; In the information decoding stage, the AP transmits the control message to each device and uses the time division multiple access protocol to handle the calculation and offload tasks. The device uses part of the collected energy for information decoding. The data rate of device k is expressed as: Where N0 represents the additive white Gaussian noise at the device; The energy consumption of each device during the information decoding stage is expressed as: Where ε is a constant based on the energy harvesting decoder; After the information decoding stage, the terminal device uses the collected energy to offload the task to the MEC server, assuming that the channel bandwidth is unit bandwidth and the channel gain of each device satisfies h1≥…≥h k ≥…≥h K , where h k represents the channel gain between device k and MEC server; First, the device with higher channel gain is decoded, and the signal sent by the device with lower channel gain is used as interference. The achievable transmission rate of device k to offload the computing task is: Among them, p k represents the transmit power of device k; When unloading tasks, the TDMA protocol is used, and each device uses the same unloading time, that is, each device The task is offloaded to the MEC server within a certain time. The energy consumption of each device during the offloading process is: L k =t k R k (6) Among them, L k is at k The number of bits offloaded by device k in time; The method comprises the following steps: S1. Considering the service quality requirements of intelligent terminal devices, the maximum transmission power of terminal devices and the limited available energy, a problem of jointly optimizing the transmission power and time switching coefficient is proposed with the goal of maximizing the total number of offloaded tasks in the system. The optimization problem is expressed as follows: Among them, constraint (7b) represents the service quality requirement of the minimum transmission rate of each device, constraint (7c) ensures that the energy consumed by each device in decoding information and unloading tasks does not exceed the energy it collects, constraint (7d) ensures that the transmission power of each device when unloading tasks does not exceed the maximum transmission power, and (7e) gives the feasibility constraint of the time switching coefficient; S2. In order to solve the mixed integer programming problem with multiple variables, the problem is decomposed into two sub-problems: the power allocation problem when the time switching coefficient is given and the optimization problem of the time switching coefficient. For the power allocation problem when the time switching coefficient is given, a mathematical model is established to solve the optimal transmission power of each device. The specific method is as follows: Assuming that the time switching coefficient of device k is known, the transmission power p k The calculation formula is: Applying the continuous convex approximation method, the rate function based on the signal-to-noise ratio is expressed as: The parameters The initial value of corresponds to the initialization of system parameters; Formula (8) is simplified to: The Lagrange multiplier method is used to solve the optimization problem. The Lagrange function is defined as follows: where λ k , μ k and θ k is a Lagrange multiplier that satisfies λ k ≥0,μ k ≥0 and θ k ≥0; Based on the Lagrange dual method, formula (9) is transformed into: Formula (9) is decomposed into two layers of optimization problems. The inner layer problem is to maximize the Lagrangian function value, and the outer layer problem is to minimize the dual gap. According to the KKT condition, the optimal transmission power is obtained: where [x] + =max{0,x}; In the outer loop, the Lagrange multiplier is updated according to the gradient descent principle: θ k (l+1)=[θ k (l)+s3(l)(p max -p k )] + , (15) Where l is the number of iterations, s1(l), s2(l) and s3(l) are positive iteration steps; S3. An iterative power allocation algorithm is constructed using continuous convex approximation, Lagrange multiplier method, and gradient descent method. The iterative power allocation algorithm is as follows: Initialization time switching coefficient α k , β k , define the iteration accuracy Δ, initialize the Lagrangian factors λ0, μ0 and θ0, initialize the iteration steps s1(0), s2(0) and s3(0) of the gradient descent method, and define l=1, k=1; Loop 1(k): Calculate the transmit power of each terminal device; Loop 2 (l): Update Lagrangian factors; The time switching coefficient α k , β k , Lagrangian factor λ k , μ k and θ k Substitute into equation (12) to obtain the optimal transmission power Update the Lagrangian factor using equations (13)(14)(15); l=l+1; Until |F k (l+1)-F k (l)|<Δ, where F k (l) = [λ k (l),μ k (l),θ k (l)] T , then the Lagrange factor converges, and cycle 2 ends; When k=K, end loop 1; Use equation (6) to calculate the total number of unloaded tasks of the system when the transmission power of the device is optimal; S4. Construct a resource allocation algorithm based on genetic algorithm to jointly optimize the transmission power and time switching coefficient, so as to obtain a joint resource allocation model of energy collection and data offloading for smart terminal devices.

2. The SWIPT-assisted NOMA-MEC system resource allocation modeling method according to claim 1 is characterized in that: The resource allocation algorithm based on genetic algorithm is divided into six stages: initialization, evaluation, selection, crossover, mutation and evolution. The roulette strategy is used to determine the adaptive crossover and mutation process in natural selection. The fitness function of each individual i is defined as the sum of the unloaded tasks: in, is the transmit power obtained by the iterative power allocation algorithm.

3. The SWIPT-assisted NOMA-MEC system resource allocation modeling method according to claim 2 is characterized in that: The specific process of the resource allocation algorithm based on genetic algorithm described in step S4 is: Define the maximum number of iterations N g , population size N p , number of chromosomes N c =2, initialize the transmission power p k , initialize the number of iterations i = 1; Generate N p The initial population set S(i) of individuals; cycle; Calculate the time switching coefficient α for each individual in the population k , β k And the optimal transmission power under given conditions Calculate the individual fitness function F(i) according to formula (16); Select the individual with the largest fitness function F(i) as the parent, let the parent enter the next generation according to the roulette strategy, select two parents for crossover with a certain crossover probability, and select a certain number of individuals for chromosome mutation operation; i=i+1, generate a new population set S(i+1); When the fitness function converges or reaches the maximum number of iterations, the loop ends; F * =max(F(i)), output the optimal time switching coefficient α k , β k With the optimal transmission power p k * .