Internet of Things equipment task unloading and resource allocation method based on photon counting

Abstract

The invention belongs to the technical field of optical communication task processing, and particularly relates to an Internet of Things equipment task unloading and resource allocation method based on photon counting. The application scenarios of the invention comprise a base station, a mobile edge computing server and N Internet of Things devices. The base station is equipped with N LDs and N PDs, the server is arranged on the base station to facilitate data communication, each Internet of Things device comprises an LD and a PD as a data transceiving pair, and each device is internally provided with a computing power unit; the method comprises the following specific steps: constructing a data model of a mobile edge computing server and equipment; constructing a task unloading system throughput optimization problem model of the mobile edge computing network; and making a task unloading and resource allocation strategy by adopting a task unloading and resource allocation algorithm based on MDPP-SCA according to the data volume generated by the equipment at each moment and the backlog amount of all queues. The method can maximize the system throughput and reduce the equipment power consumption, and is suitable for the application scene of the Internet of Things with low power consumption and high real-time performance requirements.

Description

Technical Field

[0001] The present invention belongs to the technical field of optical communication task processing, and in particular relates to a method for task unloading and resource allocation of an Internet of Things device. Background Art

[0002] With the rapid development of the Internet of Things (IoT), a large number of smart devices work together through complex network systems to generate massive amounts of data to support various application scenarios. However, due to the limitations of the computing power, storage capacity, and energy consumption of IoT devices, the traditional mode of transmitting data to the cloud for processing has gradually become difficult to meet the requirements of real-time and low power consumption. In this context, Mobile Edge Computing (MEC), as a new computing architecture, has become an important technical solution to the above problems. By deploying computing and storage resources close to the data source (such as base stations), MEC can offload computing tasks from IoT devices to edge servers, thereby reducing device burden, improving computing efficiency, and reducing network latency.

[0003] Although MEC technology has made significant progress in task offloading and resource allocation, existing research generally faces two major bottlenecks: the high energy consumption of wireless communication and the delay optimization problem of task offloading. In the IoT scenario, traditional wireless communication technologies (such as RF communication) face problems such as severe channel interference, high power consumption and limited rate, which directly affect the overall performance of the system. In addition, the coordinated optimization of uplink communication, computing resource allocation and downlink communication needs to be considered simultaneously during the task offloading process. Especially when multiple devices are running in parallel, the constraints of task return time are often ignored, which greatly reduces the feasibility of the solution in practical applications.

[0004] In recent years, the rise of photon counting wireless optical communication technology has provided a new solution to the above problems. Photon counting technology uses photodetectors (PDs) to capture the number of arrivals of single photons, which can achieve wireless data transmission with high precision and low power consumption. Compared with traditional wireless communications, photon counting technology has the advantages of low energy consumption and strong anti-interference, and has been widely studied.

[0005] On this basis, applying photon counting wireless optical communication technology to the MEC task offloading system can not only improve communication efficiency, but also maximize system throughput and reduce overall power consumption by co-optimizing communication power and computing resources. However, there is still little research on task offloading and resource allocation based on photon counting. Especially considering the multi-queue structure and dynamic task generation of IoT devices, how to efficiently schedule resources to meet multiple constraints is still a difficult problem to be solved. Summary of the invention

[0006] The object of the present invention is to provide a method for task unloading and resource allocation of Internet of Things devices based on photon counting, which has high system throughput, low device power consumption and strong real-time performance.

[0007] The present invention provides a method for task unloading and resource allocation of IoT devices based on photon counting, and its application scenario includes an outdoor mobile edge computing network consisting of a base station, a mobile edge computing server and N IoT devices, wherein the base station is equipped with N laser emitters (Laser diode, LD) and N photodetectors (Photodetector, PD), and the mobile edge computing server is arranged on the base station to facilitate data communication between the base station and the server, and the IoT devices each include an LD and a PD as a data transceiver pair, and each IoT device has a built-in computing unit and has a certain computing power. The specific steps of the method are:

[0008] S1. Build data models for mobile edge computing servers and IoT devices, including task data models, queue models, and communication models;

[0009] S2. Construct a throughput optimization problem model for the task offloading system of the mobile edge computing network;

[0010] S3. According to the amount of data generated by IoT devices at each moment and the queue backlog of all queues, the task offloading and resource allocation algorithm based on MDPP-SCA [1,2] is used to make task offloading and resource allocation strategies.

[0011] Furthermore, the data model described in step S1 is used to represent the amount of task data generated at each moment, and each task data consists of three parts, namely: is the task data volume (unit: bit), which is a random variable with a 0-1 distribution and is expressed as v is the ratio of the amount of data returned by the task to the amount of data in the task; ρ is the amount of CPU computation required per unit of data (unit cycle / bit); μ represents the amount of data in the task. The amount of data (in bits), P task Represents the probability of generating a new task.

[0012] Furthermore, the queue model described in step S1 refers to modeling the data accumulated by the server and the device at each moment according to the data structure of the queue, which includes four queue structures:

[0013] (1) Local computing queue Q lc:The local computing queue is used to store task data that needs to be calculated locally. When a task is selected to be executed locally, the local queue will be filled with the task data generated by the frame. The dequeued data is controlled by the computing power scheduling of the local device. When performing local calculations, the data in the local queue will be obtained. So Q lc The queue data length at time t can be expressed as:

[0014]

[0015] in, represents the queue length of the local computing queue at time t, x t Represents the uninstallation strategy, represents the local CPU frequency, ρ represents the computing density of the task, represents the calculation time of the local device, and Δ represents the frame length.

[0016] (2) Local unloading queue Q up :The local offload queue is used to store data that needs to be offloaded to the edge server. The input data source is the data generated by the task. The out-of-queue data is the transmission data at the time of uplink communication. So Q up The queue data length at time t can be expressed as:

[0017]

[0018] in, is the queue length of the local unloading queue at time t, is the uplink communication rate, Indicates the uplink communication time.

[0019] (3) Edge computing queue Q ec :The edge computing queue is used to store the task data that IoT devices unload to the edge side. The queued data is the amount of data uploaded by the user side, so the queued data needs to be less than or equal to the dequeued data of the unloaded queue. The dequeued data is the amount of data calculated by the edge end for each frame. So Q ec The queue data length at time t can be expressed as:

[0020]

[0021] in, represents the queue length of the edge computing queue at time t, Indicates the CPU computing frequency of the edge server at the current moment.

[0022] (4) Task returns to queue Q down :The unloading queue is used to store the calculation results of the task, so the queued data is the result data returned by the edge device after completing the task calculation, and the dequeued data is the communication data of the downlink time slot.down The queue data length at time t can be expressed as:

[0023]

[0024] in, represents the queue length of the edge queue at time t, is the downlink communication rate, Represents the downlink communication time.

[0025] Furthermore, the communication model described in S1 includes two parts: construction of a communication scheme and derivation of the lower bound of the approximate achievable rate of photon counting wireless optical communication.

[0026] In order to improve the communication efficiency of the network and reduce the power consumption of communication, the communication technology of photon counting wireless optical communication is used. The communication scheme is as follows: a time division duplex (TDD) communication scheme is adopted, that is, the communication time is divided into uplink time slots and downlink time slot All devices can only send signals to the base station in the uplink time slot and receive signals in the downlink time slot. The communication between the device and the base station adopts the photon counting wireless optical communication method to improve the communication rate between devices. This is a receiver model in wireless optical communication. Through the interaction between photons and photosensitive elements, the photodetector at the receiving end can calculate the exact number of photons that reach the photodetector in one symbol time slot. In one symbol time slot τ, the number of signal photons received Y is a discrete Poisson random variable with a probability mass function:

[0027]

[0028] Where Y is the number of received photons, y=0, 1, ..., ∞ is a non-negative integer; S is the data symbol sent by the device, s∈[0, 1]; λ is the average photon count of the received signal, satisfying the expression:

[0029]

[0030] Where η represents the quantum efficiency of photon counting, h represents the channel gain, represents Planck's constant, v represents the frequency of light, and n b represents the average photon count generated by background radiation in each symbol time slot τ, U represents the conversion factor of LD, A represents the current modulation parameter, I d represents DC bias, placing the output signal produces a negative value; represents the expected number of emitted photons, Represents bias strength. Considering the interference between devices in the communication process, WDMA is used to send information, and channels with different wavelengths are allocated to each device. The present invention uses an infrared light source with a reference wavelength of 1550nm and a channel spacing of 100GHz.

[0031] The lower bound of the approximate achievable rate of photon counting wireless optical communication is derived by first analyzing the achievable rate of photon counting wireless optical communication, and obtaining the approximate achievable rate and its lower bound through Gaussian approximation and entropy power inequality. The specific derivation steps are as follows:

[0032] S11, deriving the achievable rate according to the mutual information definition;

[0033]

[0034] S12, using Gaussian distribution instead of Poisson distribution to convert the number of received photons into a standard normal distribution form that is easy to express;

[0035] When the mean of the Poisson distribution approaches infinity, the PMF of the Poisson distribution approaches the PDF of a Gaussian distribution with equal mean and variance, i.e.;

[0036]

[0037] The number of received photons Y can be approximated as a Gaussian variable with equal mean and variance We can use the standard normal random variable Convert to standard normal distribution form:

[0038]

[0039] The achievable rate R can be approximated as:

[0040] R≈R (GA) =h(Y (GA) )-h(Y (GA) |X)=h(X+T)-h(T), (10)

[0041] S13. Use the entropy power inequality to obtain the lower bound of the approximate achievable rate.

[0042] According to the entropy power inequality, for independent random variables X and Y, with e 2h(X+Y) ≥e 2h(x) +e 2h(Y) At this time, the approximate achievable rate R can be obtained (GA) The lower bound of is:

[0043]

[0044] Considering that X is uniformly distributed, then:

[0045] h(X)=-∫ x f X (x)lnf x (x)dx=ln(2|ηn s h|), (12)

[0046] Since the information entropy of the random variable that obeys the Gaussian distribution is the largest when the variance is the same, when the random variable T is a Gaussian variable, the differential entropy h(T) takes the upper bound, which is:

[0047]

[0048] The final rate lower bound is:

[0049]

[0050] Furthermore, the throughput optimization problem model described in step S2 refers to maintaining all queues Under stable conditions, maximize the queue Q of all IoT devices lc and Q up The long-term dequeue data volume, the problem can be expressed as follows:

[0051]

[0052] subject to C1:

[0053] C2:

[0054] C3:

[0055] C4:

[0056] C5:

[0057] C6:x nt ∈{0,1},

[0058] C7:

[0059] C8:

[0060] C9:

[0061] C10:

[0062] To simplify the expression, is the set of local computing, uplink communication and downlink communication time slots, is the set of local computing power, communication power and base station communication power of IoT devices. Constraint C1 indicates that the storage space of IoT devices and servers is limited, and the corresponding queue length is limited, and unlimited data cannot be stored; C2-C4 indicates the communication power p of IoT devices and base stations. up 、p down and local computing power p lc Limited, C5 represents the long-term average power consumption of IoT devices must be lower than P ave , extending the service life of the equipment; C6 represents whether the IoT device needs to unload data at each moment, 1 represents unloading data, and 0 represents that the task is calculated locally; C7-C9 represent the time slot allocation of communication and local computing time. The IoT device cannot perform data unloading and task calculation at the same time, but it can receive data from the base station during task calculation; the base station can perform uplink and downlink communication and task calculation at the same time; C10 represents that the computing power of the edge server is limited.

[0063] Furthermore, in order to obtain the optimal offloading strategy and allocation effect, the steps of the task offloading and resource allocation algorithm based on MDPP-SCA described in S3 are as follows:

[0064] S31, use constraint C5 as virtual queue Y n , get the queue backlog of the virtual queue, combine C5 and C1 conditions to get a new queue set

[0065] S32. Using Lyapunov optimization, the throughput optimization problem is decoupled into a deterministic planning problem at each moment. The decoupled optimization problem has the following form:

[0066]

[0067] S33, decomposing the decoupled optimization problem into two resource allocation sub-problems, task offloading and edge computing power optimization problem, and communication and local computing power resource allocation problem;

[0068] S34, solving the problem of task offloading and edge computing power optimization;

[0069]

[0070] This is a linear programming problem, and after differentiation, a closed-form solution can be obtained:

[0071]

[0072] S35, using continuous convex approximation (SCA) to solve the communication and local computing resource allocation problem;

[0073]

[0074] According to the time slot τ t The non-negativity of can be used to decompose the problem into three sub-problems: uplink power allocation, downlink power allocation, and local computing resource allocation. The local computing resource allocation problem has a solution:

[0075]

[0076] The uplink power allocation problem has the following form:

[0077]

[0078] Introduce auxiliary variable α instead in:

[0079]

[0080] Using the first-order Taylor expansion, the optimization problem can be transformed into a convex problem, and the optimal solution can be obtained using the CVX toolkit.

[0081] The downlink power allocation problem has the following form:

[0082]

[0083] in:

[0084]

[0085] Introduce auxiliary variable β to replace Using the first-order Taylor expansion, the optimization problem can be transformed into a convex problem, and the optimal solution can be obtained using the CVX toolkit.

[0086] At this point, the communication power and local computing resources have been allocated, and the allocation strategy for time slot τ can be discussed by category.

[0087] Omitted: If:

[0088]

[0089] otherwise:

[0090]

[0091] Among them, there are the following marks:

[0092]

[0093] And all λ less than 0 1n The sum is recorded as And all λ less than 0 2n The sum is recorded as The above steps can get all variables p t , τt , f t ec , x t The optimal solution is obtained to maximize the system throughput.

[0094] S36. Obtain the maximum throughput of the system and update the queue backlog at the next moment.

[0095] The beneficial effect of the above scheme is that according to the channel status at each moment, the task volume and queue backlog status are generated, and the optimal resource allocation strategy can be obtained, thereby maximizing the throughput of the device. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Figure 1 This is an application scenario system model composed of the Internet of Things device, base station and MEC in Example 1 of the present invention.

[0097] Figure 2 This is a flow chart of the task offloading resource allocation framework in Example 1 of the present invention.

[0098] Figure 3 This is an algorithm flow chart of the task offloading and resource allocation algorithm based on MDPP-SCA in Embodiment 1 of the present invention.

[0099] Figure 4 This is a simulation effect diagram in Example 1 of the present invention. DETAILED DESCRIPTION

[0100] The specific implementation of the present invention is described below to facilitate the understanding of the present invention by those skilled in the art. However, it should be clear that the present invention is not limited to the scope of the specific implementation. It is obvious to those skilled in the art that as long as various changes are within the spirit and scope of the present invention defined and determined by the attached claims, these changes are obvious, and all inventions and creations using the concept of the present invention are protected.

[0101] As mentioned in the background technology, there are currently few works that consider using photon counting communication solutions to apply to task offloading. Moreover, when considering task offloading, few works consider the return time of the task, that is, the downlink communication between the base station and the IoT device. The present invention takes this time into consideration and is more in line with application conditions in real scenarios.

[0102] Embodiment 1:

[0103] First, the parameters used in the example are given, see Table 1.

[0104] Table 1

[0105]

[0106]

[0107] See also Figure 1 , is the application scenario system model of the system. This embodiment is based on an outdoor mobile edge computing network including a base station, a mobile edge computing server and several IoT devices. The base station is equipped with a laser transmitter (LD) and a photodetector (PD) and is connected to the edge computing server. Each IoT device is also equipped with an LD and a PD for data transmission and reception, and has a built-in computing unit to support a certain local computing capability.

[0108] See also Figure 2 The present invention provides a task offloading solution based on photon counting. The method includes steps S1, S2, and S3. Each step is described in detail below in conjunction with a specific embodiment.

[0109] Step S1: construct data models of servers and IoT devices, including task data models, queue models, and communication models;

[0110] Task data model

[0111] Each task data It consists of three parts: is the task data volume (unit: bit), which is a random variable with a 0-1 distribution and is expressed as v is the ratio of the amount of data returned by the task to the amount of data in the task; ρ is the amount of CPU computation required per unit data volume (unit cycle / bit).

[0112] Queue Model

[0113] Local computing queue: stores task data executed locally;

[0114] Local offloading queue: stores task data to be offloaded to the edge computing server;

[0115] Edge computing queue: stores task data uploaded to the edge server;

[0116] Task return queue: stores the task results returned after the calculation is completed.

[0117] communication model

[0118] Photon counting wireless optical communication is adopted, and the time division duplex (TDD) scheme is used to divide the uplink and downlink time slots; wavelength division multiplexing (WDMA) technology is introduced to allocate different wavelength channels; the number of signal photons is modeled using Poisson distribution, and the approximate achievable rate lower bound is derived through Gaussian approximation and entropy power inequality. The specific Gaussian approximate achievable rate lower bound is

[0119]

[0120] Furthermore, the channel attenuation h = h n (l) h n (a) h n (pe) h n (aoa) ,h n (1) =exp(-Z n (Θ) is the atmospheric scattering loss, Z n represents the distance between the base station and the device, Θ represents the atmospheric attenuation coefficient; h n (a) is the signal attenuation caused by atmospheric turbulence, and h n (pe) and h n (aoa) This is the error caused by laser alignment. The alignment error is ignored in this example.

[0121] This embodiment uses Gmma-Gmma distribution to h n (a) To model, h n (a) The PDF is

[0122]

[0123] where α n and β n represent the large-scale and small-scale vortex coefficients, Γ(·) is the Gamma function, K n (·) is the modified Bessel function of the second kind, order n.

[0124] Step S2: constructing a throughput optimization problem model for the task offloading system of the mobile edge computing network;

[0125] The goal is to maximize the long-term task throughput of IoT devices while keeping the queue stable. The specific optimization problem is as follows:

[0126]

[0127] subject to C1:

[0128] C2:

[0129] C3:

[0130] C4:

[0131] C5:

[0132] C6:x nt ∈{0,1},

[0133] C7:

[0134] C8:

[0135] C9:

[0136] C10:

[0137] Step S3: According to the amount of data generated by IoT devices in each time frame τ and the queue backlog of all queues, the task offloading and resource allocation algorithm based on MDPP-SCA is used to make a task offloading and resource allocation strategy, which is specifically divided into the following steps:

[0138] S31, converting the long-term power consumption constraint into a virtual queue backlog;

[0139] S32. Combined with Lyapunov optimization method. Using the Lyapunov drift plus penalty framework, the optimization problem is decoupled into a deterministic planning problem.

[0140]

[0141] subject to C1~C4,C6~C10.

[0142] S33, decomposing the decoupled optimization problem into two resource allocation sub-problems, task offloading and edge computing power optimization problem, and communication and local computing power resource allocation problem;

[0143] S34, solving the problem of task offloading and edge computing power optimization;

[0144]

[0145] subject to C6,C10.

[0146] This is a linear programming problem. After differentiation, we can get a closed-form solution.

[0147]

[0148] S35. Use continuous convex approximation (SCA) to solve the communication and local computing resource allocation problems and obtain the maximum throughput of IoT devices.

[0149]

[0150] subject to C1~C4,C7~C9.

[0151] According to the time slot τ t The non-negativity of can be used to decompose the problem into three sub-problems: uplink power allocation, downlink power allocation, and local computing resource allocation. The local computing resource allocation problem has a solution.

[0152]

[0153] The uplink power allocation problem has the following form,

[0154]

[0155] st C3.

[0156] Introduce auxiliary variable α instead in Using the first-order Taylor expansion, the optimization problem can be transformed into a convex problem, and the optimal solution can be obtained using the CVX toolkit.

[0157] The downlink power allocation problem has the following form,

[0158]

[0159] subject to C4

[0160] in, Introduce auxiliary variable β to replace Using the first-order Taylor expansion, the optimization problem can be transformed into a convex problem, and the optimal solution can be obtained using the CVX toolkit.

[0161] So far, the communication power and local computing resources have been allocated, and the allocation strategy of time slot τ can be discussed by category, if:

[0162] otherwise,

[0163] Among them, we have the following tags:

[0164]

[0165]

[0166] And all λ less than 0 1n The sum is recorded as And all λ less than 0 2n The sum is recorded as All variables p can be obtained t , τ t , f t ec , x t The optimal solution is obtained to maximize the system throughput.

[0167] Through the above method, this embodiment can effectively solve the problems of task offloading and unbalanced resource allocation in the prior art, significantly improve the computing throughput of the Internet of Things system, and reduce the long-term power consumption of the device. Figure 4 The implementation effect of this method is shown in Figure 1. The throughput effect of 10 IoT devices after 8000 frames using the above MDPP-SCA-based task offloading and resource allocation algorithm is plotted. In order to verify the effectiveness of the proposed algorithm, a benchmark algorithm, the SSC (Simple Static Configuration) algorithm, is provided as an example: This algorithm does not make decisions based on the environment, and performs the following actions in each frame:

[0168]

[0169] from Figure 4 As can be seen in (a), by using the task offloading and resource allocation algorithm proposed in this patent, the throughput of the terminal device can meet the task generation volume. However, by using the SSC algorithm, the throughput of the terminal cannot achieve this effect. Figure 4 As can be seen in (b), with the MDPP-SCA-based task offloading and resource allocation algorithm, the queue length of IoT devices remains stable after 8000 frames, and there is no memory overload problem, but this problem will occur with the SSC algorithm. Figure 4 (c) reflects the energy consumption of IoT devices. In this example, we set the power consumption limit to 1 Watt. It can be seen that the task offloading and resource allocation algorithm based on MDPP-SCA can meet this requirement. This simulation shows that the present invention can effectively solve the problem of task offloading and unbalanced resource allocation in the prior art.

[0170] The above technical details and algorithm implementation steps are only used as demonstration examples to better illustrate the method proposed by the present invention and should not be construed as limitations of the present invention. Other researchers in the field may make deformations and reorganizations within the scope of the present invention, and these deformations and combinations are still within the scope of protection of the present invention.

[0171] References

[0172] [1]Neely M.Stochastic network optimization with application to communication and queuing systems[M].Springer Nature,2022.

[0173] [2]Marks B R,Wright G P.A general inner approximation algorithm fornonconvex mathematical programs[J].Operations research,1978,26(4):681-683。

Claims

1. A method for task offloading and resource allocation of IoT devices based on photon counting, characterized in that: The application scenario includes an outdoor mobile edge computing network consisting of a base station, a mobile edge computing server and N IoT devices, wherein the base station is equipped with N laser transmitters (LDs) and N photodetectors (PDs), the mobile edge computing server is arranged on the base station to facilitate data communication between the base station and the mobile edge computing server, the IoT devices each include an LD and a PD as a data transceiver pair, and each IoT device has a built-in computing unit; the specific steps of the method are: S1. Build data models for mobile edge computing servers and IoT devices, including task data models, queue models, and communication models; S2. Construct a throughput optimization problem model for the task offloading system of the mobile edge computing network; S3. According to the amount of data generated by IoT devices at each moment and the queue backlog of all queues, the task offloading and resource allocation algorithm based on MDPP-SCA is used to make task offloading and resource allocation strategies.

2. The method according to claim 1, characterized in that: The task data model described in step S1 is used to represent the amount of task data generated at each moment. Each task data consists of three parts, namely: is the task data, which is a random variable that follows a 0-1 distribution and is expressed as is the ratio of the amount of data returned by the task to the amount of data in the task; ρ is the amount of CPU calculation required per unit data (unit cycle / bit); μ represents the amount of data in the task The amount of data (in bits), P task Represents the probability of generating new tasks; The queue model described in step S1 refers to modeling the data accumulated at each moment by the mobile edge computing server and the IoT device according to the data structure of the queue, which includes four queue structures: (1) Local computing queue Q lc , used to store task data that needs to be calculated locally. When the task is selected to be executed locally, the local queue is filled with the task data generated by the frame; the dequeued data is controlled by the computing power scheduling of the local device. When performing local calculations, the data in the local queue is obtained; so Q lc The queue data length at time t is expressed as: in, represents the queue length of the local computing queue at time t, x t Represents the uninstallation strategy, represents the local CPU frequency, ρ represents the computing density of the task, represents the computation time of the local device, and Δ represents the frame length; (2) Local unloading queue Q up , used to store data that needs to be offloaded to the edge server; the input data source is the data generated by the task; the dequeued data is the transmission data at the time of uplink communication; so Q up The queue data length at time t is expressed as: in, represents the queue length of the local unloading queue at time t, is the uplink communication rate, Represents the uplink communication time; (3) Edge computing queue Q ec , used to store the task data offloaded by IoT devices to the edge side. The enqueued data is the amount of data uploaded by the user side. The enqueued data is less than or equal to the dequeued data of the offload queue. The dequeued data is the amount of data calculated by the edge end for each frame. Therefore, Q ec The queue data length at time t is expressed as: in, is the queue length of the edge computing queue at time t, Represents the CPU computing frequency of the edge server at the current moment; (4) Task returns to queue Q down , used to store the calculation results of the task, so the queued data is the result data returned by the edge device after completing the task calculation, and the dequeued data is the communication data of the downlink time slot; Q down The queue data length at time t is expressed as: in, represents the queue length of the edge queue at time t, is the downlink communication rate, Represents the downlink communication time.

3. The method according to claim 2, characterized in that The communication model described in step S1 includes two parts: construction of a communication scheme and derivation of the lower bound of the approximate achievable rate of photon counting wireless optical communication; wherein: (i) The communication scheme adopts a time division duplex (TDD) communication scheme, that is, the communication time is divided into uplink time slots and downlink time slot All devices can only send signals to the base station in the uplink time slot and receive signals in the downlink time slot. The communication between the device and the base station adopts the photon counting wireless optical communication to improve the communication rate between devices. This is a receiver model in wireless optical communication. Through the interaction between photons and photosensitive elements, the photodetector at the receiving end can calculate the exact number of photons reaching the photodetector in one symbol time slot. In one symbol time slot τ, the number of signal photons received Y is a discrete Poisson random variable with a probability mass function: Where Y is the number of received photons, y=0, 1, ..., ∞ is a non-negative integer; S is the data symbol sent by the device, s∈[0, 1]; λ is the average photon count of the received signal, satisfying the expression: Where η represents the quantum efficiency of photon counting, h represents the channel gain, represents Planck's constant, v represents the frequency of light, and n b represents the average photon count generated by background radiation in each symbol time slot τ, U represents the conversion factor of LD, A represents the current modulation parameter, I d represents DC bias, placing the output signal produces a negative value; represents the expected number of emitted photons, Represents bias strength; considering the interference between devices during the communication process, WDMA is used to send information and allocate channels of different wavelengths to each device; (ii) The derivation of the lower bound of the approximate achievable rate of the photon counting wireless optical communication first analyzes the achievable rate of the photon counting wireless optical communication, and obtains the approximate achievable rate and its lower bound through Gaussian approximation and entropy power inequality. The specific derivation steps are as follows: S11, deriving the achievable rate according to the mutual information definition; S12, using Gaussian distribution instead of Poisson distribution to convert the number of received photons into a standard normal distribution form that is easy to express; When the mean of the Poisson distribution approaches infinity, the PMF of the Poisson distribution approaches the PDF of a Gaussian distribution with equal mean and variance, i.e.; The number of received photons Y can be approximated as a Gaussian variable with equal mean and variance With standard normal random variable Convert to standard normal distribution form: The achievable rate R is approximately: R≈R (GA) =h(Y (GA) )-h(Y (GA) |X)=h(X+T)-h(T), (10) S13, using the entropy power inequality to obtain the lower bound of the approximate achievable rate; According to the entropy power inequality, for independent random variables X and Y, with e 2h(X+Y) ≥e 2h(x) +e 2h(Y) ; At this time, the approximate achievable rate R is obtained (GA) The lower bound of : Considering that X is uniformly distributed, then: h(X)=-∫ x f x (x)lnf x (x)dx=ln(2|ηn s h|), (12) Since the information entropy of the random variable that obeys the Gaussian distribution is the largest when the variance is the same, when the random variable T is a Gaussian variable, the differential entropy h(T) reaches the upper bound, which is: The final rate lower bound is:

4. The method according to claim 3, characterized in that The throughput optimization problem model described in step S2 refers to maintaining all queues Under stable conditions, maximize the queue Q of all IoT devices 1c and Q up The long-term dequeue data volume can be expressed as follows: C6:x nt ∈{0,1}, To simplify the expression, is the set of local computing, uplink communication and downlink communication time slots, A collection of local computing power, communication power, and base station communication power for IoT devices; Constraint C1 indicates that the storage space of IoT devices and servers is limited, and the corresponding queue length is limited, and unlimited data cannot be stored; C2-C4 represents the communication power p of IoT devices and base stations up 、p down and local computing power p lc Limited, C5 represents the long-term average power consumption of IoT devices must be lower than P ave , extending the service life of the equipment; C6 represents whether the IoT device needs to unload data at each moment, 1 represents unloading data, and 0 represents that the task is calculated locally; C7-C9 represent the time slot allocation of communication and local computing time. The IoT device cannot perform data unloading and task calculation at the same time, but it can receive data from the base station during task calculation; the base station can perform uplink and downlink communication and task calculation at the same time; C10 represents that the computing power of the edge server is limited.

5. The method according to claim 4, characterized in that The specific steps of the MDPP-SCA-based task offloading and resource allocation algorithm described in step S3 are as follows: S31, use constraint C5 as virtual queue Y n , get the queue backlog of the virtual queue, combine C5 and C1 conditions to get a new queue set S32. Using Lyapunov optimization, the throughput optimization problem is decoupled into a deterministic planning problem at each moment. The decoupled optimization problem has the following form: S33, decomposing the decoupled optimization problem into two resource allocation sub-problems, task offloading and edge computing power optimization problem, and communication and local computing power resource allocation problem; S34, solving the problem of task offloading and edge computing power optimization; This is a linear programming problem, and the closed-form solution is obtained after differentiation: S35, using continuous convex approximation (SCA) to solve the communication and local computing resource allocation problem; According to the time slot τ t The non-negativity of can decompose the problem into three sub-problems: uplink power allocation, downlink power allocation, and local computing resource allocation. The local computing resource allocation problem has a solution: The uplink power allocation problem has the following form: Introduce auxiliary variable α instead in, The optimization problem is transformed into a convex problem using the first-order Taylor expansion, and the optimal solution is obtained using the CVX toolkit; The downlink power allocation problem has the following form: in: Introduce auxiliary variable β to replace Using the first-order Taylor expansion, the optimization problem can be transformed into a convex problem, and the optimal solution can be obtained using the CVX toolkit; So far, the communication power and local computing resources have been allocated. Further classification and discussion will lead to the allocation strategy of time slot τ: If: otherwise: Among them, there are the following marks: All λ less than 0 1n The sum is recorded as All λ less than 0 2n The sum is recorded as The above steps can get all variables The optimal solution of , thereby maximizing the system throughput; S36. Obtain the maximum throughput of the system and update the queue backlog at the next moment.

Images