CN111585916A

CN111585916A - LTE electric power wireless private network task unloading and resource allocation method based on cloud edge cooperation

Info

Publication number: CN111585916A
Application number: CN201911365711.8A
Authority: CN
Inventors: 李欢; 孙峰; 刘扬; 王东东; 卢盛阳; 杨智斌; 任帅; 李桐; 佟昊松
Original assignee: State Grid Corp of China SGCC; Electric Power Research Institute of State Grid Liaoning Electric Power Co Ltd
Current assignee: State Grid Corp of China SGCC; Electric Power Research Institute of State Grid Liaoning Electric Power Co Ltd
Priority date: 2019-12-26
Filing date: 2019-12-26
Publication date: 2020-08-25
Anticipated expiration: 2039-12-26
Also published as: CN111585916B

Abstract

The invention belongs to the technical field of electric power, and particularly relates to an LTE electric power wireless private network task unloading and resource allocation method based on cloud edge cooperation. The invention is that four kinds of different time delays exist in the cloud edge cooperative system, calculate the whole time delay of the apparatus and establish as mathematical model for the problem of minimizing the total time delay of all mobile apparatuses; and considering a task unloading scheme in four different system scenes, and solving an optimal allocation value of computing resources in the cloud edge cooperative system. Under all conditions, the performance of the method is optimal, and by executing the cloud edge cooperation strategy and adopting the task unloading and resource allocation strategy provided by the invention, the computing capacity of the edge node and the cloud server can be optimally allocated, so that the minimization of the average system time delay is realized.

Description

LTE electric power wireless private network task unloading and resource allocation method based on cloud edge cooperation

Technical Field

The invention belongs to the technical field of electric power, and particularly relates to an LTE electric power wireless private network task unloading and resource allocation method based on cloud edge cooperation.

Background

With the rapid development of the LTE wireless private network, the number of access devices in the network is increasing, which brings great challenges to the traditional cloud computing network. In the traditional cloud computing, task delay is large due to multi-hop data transmission, and the load of a core network is large. Mobile edge computing is considered as a key technology of next generation wireless communication, and provides low-latency computing service and improves the computing capacity of a network by a mobile edge computing center deployed at the edge of the network so as to meet the continuously improved user requirements.

How to offload the assumed tasks to the edge server by the mobile device for using the services provided by the edge network to make efficient and reasonable offload decisions has become a main research direction of the edge computing problem at present. However, with the development of technologies such as the internet of things, the scale of data computing services increases explosively, and task data is flooded into a computing network, which poses a great challenge to the limited computing resources and communication resources of the mobile edge computing center.

The prior art is as follows:

the technical scheme 1: the patent publication number is CN109302709A, which is named as a mobile edge computing-oriented task unloading and resource allocation strategy of the Internet of vehicles, and relates to a (MEC) -based task unloading mode decision and resource allocation method in a vehicle heterogeneous network, and the method is mainly completed by two steps: firstly, clustering the request vehicles according to different QoS by adopting an improved K-means algorithm so as to determine a communication mode; second, channel and power allocation is performed using a distributed Q-Learning algorithm using a Contention Free Period (CFP) based LTE-U in combination with a Carrier Aggregation (CA) technique.

The technical scheme 2 is as follows: the patent publication number is CN109698861A, which is named as a calculation task unloading algorithm based on cost optimization, and relates to a calculation task unloading algorithm based on cost optimization, and the calculation task unloading algorithm is mainly completed through four steps: firstly, constructing a new edge cloud computing model; the new edge cloud computing model comprises three important costs: calculating the execution cost of tasks, the communication cost between the same-end calculation tasks and the asymmetric communication cost between cross-end calculation tasks; secondly, expanding the new model of edge cloud computing; thirdly, combining the calculation costs; fourth, the optimized offloading strategy is solved based on a greedy criterion.

Technical scheme 3: the patent publication number is CN110489176A, entitled a multi-access edge computing task unloading method based on a boxing problem, which is mainly completed by three steps: firstly, calculating the capacity of each edge server and the ratio of the size of input data of each terminal task to required calculation resources; then two queues are formed from large to small according to the capacity and the task ratio; and finally, sequentially taking out the tasks in the task queue, configuring the tasks on the edge server with the largest capacity and the residual computing resources in the container queue, and repeating the operation until the task queue is empty.

Technical scheme 1 provides advantages of high bandwidth and low time delay brought by an IT service environment and cloud computing capability for Mobile Edge Computing (MEC), and simultaneously researches task unloading mode decision and resource allocation problems based on (MEC) in a vehicle heterogeneous network by combining an LTE authorization-free spectrum (LTE-U) technology. Considering the link differentiation requirements, i.e. high capacity of the vehicle-to-roadside unit (V2I) link and high reliability of the vehicle-to-vehicle (V2V) link, we model the user quality of service (QoS) as a combination of capacity and latency. However, such methods have the disadvantage that they are difficult to expand in other contexts.

Technical scheme 2 relates to a cost optimization-based calculation task unloading algorithm, and solves the problem of unloading optimization of calculation tasks in a combined framework of edge calculation and cloud calculation. But this approach does not fully account for the different types of latency that exist in cloud-edge collaborative systems.

Technical scheme 3 provides a multi-access edge computing task unloading method based on a boxing problem, which takes a user terminal and an edge server as a task container and takes a task as an article, so that a task unloading decision problem in edge computing is converted into a boxing problem, the number of the edge servers started in a network is minimized through a heuristic method, and a task unloading decision is solved. However, such methods have the disadvantage that they do not take into account the differences between the different types of systems.

Therefore, how to make effective task offloading and resource allocation decisions in a limited resource context becomes a key issue.

Disclosure of Invention

Aiming at the defects in the prior art, the invention provides a task unloading and resource allocation method for an LTE electric power wireless private network based on cloud-side coordination, aiming at researching different time delays in a cloud-side coordination system and realizing the minimization of the total time delay of all mobile equipment by task unloading and resource allocation calculation under limited resources.

The technical scheme adopted by the invention for solving the technical problems is as follows:

the LTE electric wireless private network task unloading and resource allocation method based on cloud-edge cooperation is characterized in that all time delays of equipment are calculated and established as a mathematical model for the problem of minimizing the total time delay of all mobile equipment in four different time delays existing in a cloud-edge cooperation system; considering task unloading schemes in four different system scenes, and solving an optimal allocation value of computing resources in the cloud edge cooperative system; the method comprises the following steps:

step 1, establishing a cloud edge cooperative system;

step 2, establishing an MEC cache model;

step 3, analyzing time delay existing in the cloud edge cooperative system;

step 4, aiming at the time delay analysis result, constructing a problem model;

step 5, solving the constructed problem;

and 6, providing a solution to the constructed problem.

Step 1, establishing a cloud-edge coordination system, wherein the cloud-edge coordination system is composed of a central cloud server and M SCeNBs and is represented by a set M ═ {1, 2., J }; each SCeNB deploys an MEC server, and the MEC server uses limited resources for data processing, caching and storing; each combination of SCeNB and MEC server is called an edge node; within the coverage area of the jth SCeNB, there is a set χ_jX of the representation_jEach mobile device has a calculation task with different time delay requirements; assuming that each user is connected with a base station, and the specific connection relation is determined by a user communication strategy; each mobile device is connected with a corresponding base station through a wireless channel, and the edge nodes transmit data to the cloud server through different backhaul links; in the system, each computing task is supposed to be processed on an edge node and a cloud server; according to [1]The model in (1), assuming all tasks are of the same type and arrive at the same time, with q_i,j＝(s_i,j,c_i,j) Representing computational tasks generated by the ith mobile device connected to the jth edge node, where s_i,jRepresenting the size of the computational task, c_i,jIndicating the CPU computation cycle required to compute this task; representing the total number of tasks requested by the system in a certain time period by N; defining the computing capacity of the MEC server and the computing capacity of the cloud server in the jth edge node as F_j ^eAnd F^c(ii) a The computing resources of each MEC server and cloud server are allocated to the mobile device by virtual machine technology.

Step 2, establishing the MEC cache model, which comprises the following steps: for each task, defining the cache vector of the MEC server as Y_i,j＝[y_1,1,y_1,2,...,y_i,j](ii) a If y is_i,j1, it means that the MEC server has cached the calculation result of the task, and y_i,jIf the result is 0, the corresponding task result is not cached; when y is_i,jWhen the value is 1, the MEC server directly transmits the calculation result to the mobile equipment without calculation; since the transmission power of the SCeNB is greater than the transmission power of the mobile device, the data volume of the calculation result is less than the data volume of the task itself, and the transmission delay on the wireless downlink is ignored, and the transmission delay required by the MEC caching scheme is considered to be 0; and the MEC cache model directly returns the results of frequently requested tasks to the mobile equipment, so that the time delay is reduced and the energy consumption is reduced.

Step 3, analyzing the time delay existing in the cloud edge coordination system, including: in a cloud-edge-end task allocation model, assuming that each mobile device sends a task request due to limited computing capacity and electric quantity of the mobile device, and judging whether a result of the task exists in a cache or not; if yes, directly returning a task result; if not, the mobile device uploads the task to the edge node; and determining whether the task is processed by the edge node alone or the edge node and the cloud server cooperatively by the corresponding edge node.

If the task is not existed, the mobile device uploads the tasks to the edge node, and the edge node is also responsible for determining the unloading proportion of each task at the cloud edge.

Step 3, analyzing the time delay existing in the cloud edge coordination system, specifically comprising the following steps:

step (1) the mobile device sends a corresponding task request to the connected edge node, and the task request is used for judging whether the data is wanted by the edge node, and the edge node judges whether the result of the task exists in a cache; if yes, directly returning a task result, and finishing the task processing; otherwise, continuing the next step;

the mobile equipment directly uploads the whole time delay sensitive task to a connected edge node through a wireless channel;

step (3) the MEC server positioned at each edge node divides the received task into two parts, wherein one part is left on the MEC server, and the other part is unloaded to the cloud server;

step (4), the MEC server distributes the available computing resources to each computing task, and simultaneously uploads partial data to the cloud server through a backhaul link;

step (5), the cloud server allocates the computing resources to corresponding computing tasks, so that parallel computing is achieved;

and (6) collecting the calculation result by each edge node, caching the calculation result through a task caching strategy, and then returning the calculation result to each mobile device.

The delay existing in the calculation process comprises the following steps:

a. transmission delay of mobile device: according to the Shannon formula, the MT is calculated_iOf (2) an upload rate r_i,jExpressed as:

where B denotes the channel bandwidth, p_iRepresents MT_iOf the transmission power, σ²Representing the noise power of the mobile device, I_i,jRepresenting the interference power, g, within the cell_i,jRepresenting a channel gain of the communication;

sc is communicated with MEC through optical fiber, and the transmission rate c between the two is far greater than MT_iOf (2) an upload rate r_i,jAnd the distance between the two is very small, neglecting the communication delay of sc and MEC, use

To indicate that the mobile device will task q_iUpload to sc_jThe resulting time delay:

in the above formula: MT (multiple terminal)_iOf (2) an upload rate r_i,j，S_iRepresenting the size of the computing task;

b. computing time delay of edge node: after the edge node successfully receives the complete computing task sent by the mobile equipment, the MEC server immediately executes an unloading strategy, and divides each computing task into two parts, wherein one part is executed by the MEC server, and the other part is executed by the cloud server; it is assumed that each calculation task is arbitrarily split without considering the task content, corresponding to the scene of video compression, speech recognition, and the like](ii) a Definition of_i,j∈[0,1]In order to calculate the division ratio of the task,_i,_jrepresenting the proportion of the computing task data executed at the MEC server; by using

The computing resource allocated to the ith mobile device by the jth edge node is represented, and the time delay generated by the task executed at the edge node is represented as:

in the above formula: c_i,jIndicating the CPU computation cycle required to compute this task;

c. transmission delay of edge node: for each edge node, separating the communication module and the computation module (CPU/GPU); the communication module is a transceiver, and the computing module is a CPU/GPU;

in the edge node, the calculation of the task and the transmission of the task are executed in parallel; all the edge nodes are connected with the cloud server through different backhaul links; an optimal cooperative strategy of edge computing and cloud computing is provided, and a resource scheduling strategy and a routing algorithm are assumed to be determined; h is to be_jExpressed as backhaul communication capacity per device associated with the jth edge node; the average backhaul transmission delay is proportional to the size of the transmitted data, and is expressed as:

wherein

Representing the time, S, required to transmit data of 1 bit size over the backhaul link_iWhich represents the size of the computing task,_i,jrepresenting the proportion of the computing task data executed at the MEC server;

d. computing time delay of the cloud server: when the cloud service successfully receives the task data sent from the edge node, the cloud server allocates available computing resources to each task to achieve parallel processing; (1-_i,j)s_i,jc_i,jRepresenting the number of CPUs, i.e., computing resources, required to execute tasks offloaded to the cloud server; by using

Representing cloud computing resources allocated to an ith mobile device served by a jth edge node; the computing latency of the cloud server is expressed as:

in the above formula: c_i,jIndicating the CPU computation cycles required to compute this task.

The different time delays existing in the cloud edge coordination system in the step 1 include: the method comprises the steps of transmitting delay of the mobile equipment, calculating delay of edge nodes, transmitting delay of the edge nodes and calculating delay of a cloud server; the total latency generated by each device includes:

assume that 1: since the offloading of tasks depends on specific parameters of each task;

assume 2: the task calculation depends on a specific data structure and the correlation between adjacent data, and the cloud server starts to process the task until the transmission between the edge node and the cloud server is finished;

based on the above assumptions, the total latency incurred by the ith mobile device served by the jth edge node can be expressed as:

in the above formula:

representing a mobile device to task q_iUpload to sc_jThe time delay that is generated is,

representing the time delay incurred in executing the task at the edge node,

representing the latency incurred by the edge node to upload the task to the cloud server,

representing the time delay generated by the processing of the computing task at the cloud server;

the minimization of the total time delay of all mobile devices is realized by task unloading and computing resource allocation under limited resources, and the goal is expressed as:

in the above formula: y is_i,jRepresenting the cache state of the task;

wherein (7a) and (7b) representNeither edge computing resources nor cloud computing resources allocated to the mobile device are likely to exceed their own maximum number of computing resources; optimization variables include the unload ratio of a task_i,j}, allocation of computing resources

Step 5, decomposing the constructed problem, including:

q1 is broken down into two parts:

task offloading and computing resource allocation only affect the solution of the latter part, dividing the problem:

wherein

The proposed solution to the constructed problem described in step 6 comprises:

(1) and (3) task caching strategy:

before task unloading decision is carried out, the problem Q2 is further simplified, and a task caching strategy is used for solving a caching vector Y_i,j＝[y_1,1,y_1,2,...,y_i,j]The algorithm process is shown in fig. 4:

the request times of the mobile equipment for requesting the task Q are assumed to follow the Poisson distribution

C represents the request times of the tasks, and lambda represents the average occurrence rate of the tasks in unit time; setting the vector group Y to be null during initialization, determining the task type according to Poisson distribution for each task, judging whether the task exists in a task cache table, and if the task exists in the task cache table, directly returning an execution result to the MEC serverMobile device, y_i,jIf not, continuing to judge whether the task exists in the task history record table; if present in the table, will y_i,jSetting the value to be 1, directly returning an execution result, storing the task into a task cache table, and updating a task history record table; if neither table has recorded task q_iThen y will be_i,jSetting to 0, traversing the subtask space in such a way, and finally obtaining a vector group Y_i,j＝[y_1,1,y_1,2,...,y_i,j]；

Using vector set Y_i,j＝[y_1,1,y_1,2,...,y_i,j]The simplification of Q2 yields:

wherein

(2) Task unloading strategy:

determining an optimal task offload ratio_i,jMaintained during the analysis

The value of (d) is unchanged;

calculating delay by analysis

And task allocation ratio_i,jDetermine an optimal division ratio for monotonicity between

Firstly, according to the formula

Deduce

Is monotonically increasing, with_i,jIs increased with an increase in; when in use_i,j∈[0,1]To obtain

Secondly, according to the formula

Deduce

Is monotonically decreasing with_i,jIncrease and decrease of (a) and also obtain the value range of the function

Observation formula

In connection with the previous discussion, it was discovered

First following_i,jIs increased and decreased with_i,jIs increased with an increase in; when in use

When it is obtained

Minimum value of (d):

wherein

Each mobile device defines two important parameters, including:

(2.1) normalized backhaul communication capacity is defined as the ratio of backhaul communication capacity to marginal computing power, i.e.

(2.2) normalized cloud computing power is defined as the ratio of cloud computing power to edge computing power, i.e.

Based on the definition, obtaining an optimal task unloading strategy; optimal task unloading strategy in cloud edge cooperative system

Expressed as:

in the above formula: mu.s_i,jDenotes the ratio of backhaul communication capacity to edge computing power, v_i,jRepresenting a ratio of cloud computing power to edge computing power;

the different systems in which said formula (10) exists are as follows:

the 1 st: communication-limited systems: the edge node and the cloud server have sufficient computing resources, but the communication capacity is insufficient;

in this case, it is preferable that the air conditioner,

this occurs when an edge node connects a large number of mobile devices and the capacity of the backhaul link is insufficient, simplifying the optimal task offload ratio:

equation (11) indicates that in this case, the optimal task offload ratio is determined only by the normalized backhaul communication capacity, and is not affected by the normalized cloud computing capacity; in this case, the communication resources reduce the main bottleneck of end-to-end delay per device; when mu is_i,j→ 0, optimal task unload ratio

1, all computing tasks are executed at the edge nodes;

the 2 nd: a compute-constrained system: the edge nodes and the cloud server have sufficient communication capacity, but insufficient computing resources;

in this case, it is preferable that the air conditioner,

simplifying the optimal task unloading ratio:

the optimal task offloading rate is determined only by the normalized cloud computing capability, in this case, as the backhaul communication capacity is relatively sufficient, the time delay generated by communication becomes small, and the size of the computing time delay determines the size of the overall time delay of the mobile device; in this case, the edge node and the cloud server are considered as a whole, according to the ratio of their computing power, v_i,jSplitting the tasks according to the proportion; if the computing power of the edge node is larger than that of the cloud server, v_i,jIf the number of the data is less than 1, unloading more data to the edge node; otherwise, more data is offloaded to the cloud, at v_i,jIn the special case of 1, data is evenly distributed between the edge node and the cloud server;

and (3) type: edge-dominated systems: the computing resources allocated to the mobile device by the edge node are far more than those allocated by the cloud server, i.e. the edge node is capable of allocating to the mobile device

Obtain v_i,j→ 0, this case corresponds to a large scale small cell network, where the cloud server serves many edge nodes; in this system, the optimal task offload ratio is simplified as follows:

indicating that the entire task should be executed on the edge node, the entire task only needs to be offloaded to the edge node;

and 4, the method comprises the following steps: cloud-side dominant system: the cloud server has sufficient computing resources, while the edge nodes have limited computing resources, i.e.

Obtain v_i,j→ infinity, such a system corresponds to a scenario where the cloud server is powerful in computing power and the edge nodes are weak in computing power; v is to_i,jSubstituting → ∞ into the formula gives:

according to the formula, the optimal task unloading rate is determined only by the normalized backhaul communication capacity; if backhaul communication capacity H_jComputing power of large or edge nodes

The unloading rate of the optimal task is reduced along with the reduction of the unloading rate of the optimal task; because when v is_i,jTime delay generated by the cloud end to execute the task can be ignored compared with time delay generated by the edge node to execute the task; so when mu is_i,jIf the delay is less than 1, the overall delay of the mobile device will be dominated by the delay generated by the backhaul transmission, so that the proportion of task data unloaded to the edge side for processing is larger, namely

Otherwise, when μ_i,jWhen the data is more than 1, more data should be unloaded to the cloud server for execution;

(3) computing resource allocation policy: unloading the optimal task rate

Substituting into formula to calculate time delay

The simplification is as follows:

question Q3 is rewritten as:

theorem 1: q4 is a convex optimization problem;

and (3) proving that: only by proving that the target function and the limiting condition of the target function are both convex functions, the theorem proves that the target function is a convex function; looking at the constraints (7a) and (7b), it can be seen that the constraints (7c) and (7d) are affine, reflecting their convexity; the following demonstrates that the objective function is also a convex function:

the hese matrix of (a) is represented as:

through mathematical calculations, all the former principals of W are obtained:

according to the linear algebra theory, when the front principal and the minor of a matrix are all positive definite, the matrix is a positive definite matrix; so as to obtain the compound with the characteristics of,

based on variables

And

a convex function of (d); and the objective function

Is the sum of a series of convex functions, and the obtained target function is also a convex function; theorem 1 obtaining the certificate;

according to theorem 1, Q4 meeting the KKT condition is obtained, and the optimal allocation value of the computing resources in the cloud edge cooperative system is obtained by using a Lagrange operator:

wherein

And θ is the optimum variable for lagrange.

The invention has the following advantages and beneficial technical effects:

the invention provides a task unloading and resource allocation (Cloud-Edge-TORD) scheme based on Cloud-Edge cooperation, a Cloud-Edge cooperation system and an MEC cache model are established, and time delay existing in the Cloud-Edge cooperation system is analyzed. The aim of minimizing the total time delay of all mobile equipment is achieved through task unloading and resource allocation calculation under limited resources, a problem model is established, and the problem is decomposed to obtain a task caching, unloading and resource allocation strategy.

The method is based on the established cloud-edge cooperative system and the MEC cache model, analyzes the time delay in the cloud-edge cooperative system, and aims to minimize the time delay. Simulation shows that the Cloud-Edge-TORD scheme provided by the invention has optimal performance under all conditions, and the task unloading and resource allocation strategy provided by the invention can be adopted to optimally allocate the computing capacity of the Edge node and the Cloud server by executing the Cloud Edge cooperation strategy, so that the minimization of the average system delay is realized.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings of the embodiments of the present invention will be briefly described below. Wherein the drawings are only for purposes of illustrating some embodiments of the invention and are not to be construed as limiting the invention to all embodiments thereof.

FIG. 1 is a schematic diagram of a cloud-edge collaboration system of the present invention;

FIG. 2 is a schematic diagram illustrating a task execution flow of the cloud-edge collaboration system according to the present invention;

FIG. 3 is a schematic diagram of the MEC cache model of the present invention;

FIG. 4 is a process diagram of a cache vector algorithm of the present invention;

FIG. 5 is a comparison graph of time delays for 4 scenarios when the computing power of the cloud server increases in accordance with the present invention;

fig. 6 is a time delay comparison diagram of 4 schemes when the computing power of the cloud server becomes smaller according to the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

Example 1

The invention relates to a LTE electric power wireless private network task unloading and resource allocation method based on cloud-edge cooperation. The invention considers task unloading schemes in four different system scenes and solves the optimal allocation value of computing resources in the cloud edge cooperative system.

The invention relates to a LTE electric power wireless private network task unloading and resource allocation method based on cloud edge cooperation, which comprises the following steps:

step 1, establishing a cloud edge cooperative system;

step 2, establishing an MEC cache model;

step 3, analyzing time delay existing in the cloud edge cooperative system;

step 4, aiming at the time delay analysis result, constructing a problem model;

step 5, decomposing the constructed problems;

and 6, providing a solution to the constructed problem.

In the invention, step 1 describes establishing a cloud-edge collaboration system, as shown in fig. 1, fig. 1 is a schematic diagram of the cloud-edge collaboration system of the invention. The cloud edge coordination system is composed of a central cloud server and M SCeNBs, and is represented by a set M ═ {1, 2. The SCeNB represents a small base station. Each SCeNB deploys an MEC server that uses limited resources for data processing, caching, and storage. Each combination of SCeNB and MEC server is called an edge node. Within the coverage area of the jth SCeNB, there is a set χ_jX of the representation_jEach mobile device has a computational task with different latency requirements. Assuming that each user is already connected to a base station, the specific connection relationship can be determined by some user connection policies. Further, each mobile device is connected to a respective base station through a wireless channel, and the edge nodes transmit data to the cloud server through different backhaul links. In this system, it is assumed that each computing task can be processed on both edge nodes and cloud servers according to [1 ]]Model of (1)]Y.Mao,C.You,J.Zhang,K.Huang,andK.B.Letaief,“Asurvey onmobile edge computing:The communication perspective,”IEEECommun.Surv.Tut.,vol.19,no.4,pp.2322–2358,Aug.2017。

We assume that all tasks are of the same type and arrive at the same time, so q can be used_i,j＝(s_i,j,c_i,j) Representing computational tasks generated by the ith mobile device connected to the jth edge node, where s_i,jRepresenting the size of the computational task, c_i,jIndicating the CPU computation cycles required to compute this task. N represents the total number of tasks requested by the system in a certain time period, and the computing capacity of the MEC server in the jth edge node and the computing capacity of the cloud server are respectively defined as

And F^c. The computing resources of each MEC server and cloud server may be allocated to the mobile device through virtual machine technology. As shown in fig. 2, fig. 2 is a schematic diagram of a task execution flow of the cloud edge collaboration system of the present invention.

Step 2, establishing an MEC cache model, as shown in fig. 3, fig. 3 is a schematic diagram of the MEC cache model of the present invention. For each task, the cache vector of the MEC server may be defined as Y_i,j＝[y_1,1,y_1,2,...,y_i,j]. If y is_i,j1, it means that the MEC server has cached the calculation result of the task, and y_i,jAnd 0 indicates that the corresponding task result is not cached. Furthermore, when y_i,jWhen 1, the MEC server directly transmits the calculation result to the mobile device without performing calculation. Since the transmission power of the SCeNB is much larger than that of the mobile device, and the data volume of the calculation result is much smaller than that of the task itself, the transmission delay on the wireless downlink can be ignored, and the transmission delay required by the MEC caching scheme is considered to be 0. The MEC cache model can directly return results of frequently requested tasks to the mobile equipment, so that time delay can be effectively reduced, and energy consumption can be effectively reduced.

And 3, analyzing time delay existing in the cloud edge coordination system, and in the cloud-edge-end task allocation model, assuming that each mobile device cannot directly process tasks due to limited computing capacity and electric quantity of each mobile device, and no local computing exists. The mobile device first sends a task request to determine whether the results of the task are in the cache. If the task exists, the task result is directly returned, and if the task does not exist, the mobile equipment uploads the task to the edge node. Then, whether the task is processed by the edge node alone or the edge node and the cloud server are cooperatively processed is determined by the corresponding edge node. In the latter case, the edge node is also responsible for determining the unloading proportion of each task at the cloud edge.

step (1), the mobile device firstly sends a corresponding task request to the connected edge node, which is equivalent to sending only the head part of the data at the beginning of the https protocol, and is used for judging whether the data is wanted by the edge node, and the edge node judges whether the result of the task exists in the cache. If yes, directly returning a task result, and finishing the task processing; otherwise, the next step is followed.

And (3) the mobile equipment directly uploads the whole time delay sensitive task to the connected edge node through a wireless channel without local calculation.

And (3) the MEC server positioned at each edge node divides the received task into two parts, wherein one part is left on the MEC server, and the other part is unloaded to the cloud server.

And (4) the MEC server allocates the available computing resources to each computing task and uploads part of data to the cloud server through the backhaul link.

And (5) the cloud server also allocates the computing resources to corresponding computing tasks, so that parallel computing is realized.

And (6) finally, collecting the calculation result by each edge node, caching the calculation result through a task caching strategy, and then returning the calculation result to each mobile device.

It should be noted that the time to divide the task is very short compared to the corresponding computation and communication delays and can therefore be neglected. Furthermore, the data volume of both the task request and the computation result is very small, so the send and return delays can be neglected, which corresponds to many practical computational scenarios, such as face recognition, virus detection and video analysis, etc.

Wherein, four delays existing in the calculation process comprise the following:

a. transmission delay of mobile device: as mentioned above, each computational task is uploaded to the corresponding edge node via a wireless channel. According to the Shannon formula, MT can be obtained_iOf (2) an upload rate r_i,jExpressed as:

where B denotes the channel bandwidth, p_iRepresents MT_iOf the transmission power, σ²Representing the noise power of the mobile device, I_i,jRepresenting the interference power, g, within the cell_i,jRepresenting the channel gain of the communication.

Note that sc is communicated with MEC via optical fiber, and the transmission rate c between the two is far greater than MT_iOf (2) an upload rate r_i,jAnd the distance between the two is very small, so the communication delay between the sc and the MEC can be ignored and used

in the above formula: MT (multiple terminal)_iOf (2) an upload rate r_i,j，S_iRepresenting the size of the computational task.

b. Computing time delay of edge node: after the edge node successfully receives the complete computing task sent from the mobile device, the MEC server immediately executes an unloading strategy, and each computing task is divided into two parts, wherein one part is executed by the MEC server, and the other part is executed by the cloud server. It is assumed herein that each calculation task can be arbitrarily split without considering task contents, which corresponds to a scenario of video compression and speech recognition, etc. [ 2 ]]. Definition of_i,j∈[0,1]In order to calculate the division ratio of the task,_i,jrepresenting the proportion of the computational task data performed at the MEC server. By using

The computing resources allocated to the ith mobile device by the jth edge node are represented, and therefore, the time delay caused by the task executed at the edge node can be represented as:

c. Transmission delay of edge node: the communication module and the computation module (CPU/GPU) are typically separated for each edge node. The communication module is a transceiver, and the computing module is a CPU/GPU.

Thus, in the edge node, the computation of the task and the transmission of the task may be performed in parallel. Furthermore, all edge nodes are connected to the cloud server via different backhaul links, which are typically equipped with very high bandwidth. In fact, the backhaul link is user-shared, so its delay is difficult to model due to the randomness of packet arrival, multi-user scheduling, complex routing algorithms, and other factors. The optimal cooperation strategy of edge computing and cloud computing is proposed, and the resource scheduling strategy and the routing algorithm are determined. Thus, H_jExpressed as the backhaul communication capacity of each device associated with the jth edge node. Therefore, similar to the average transmission delay in (2), the average backhaul transmission delay is also proportional to the size of the transmitted data, which can be expressed as:

wherein

Representing the time, S, required to transmit data of 1 bit size over the backhaul link_iWhich represents the size of the computing task,_i,jrepresenting the proportion of the computational task data performed at the MEC server.

d. Computing time delay of the cloud server: when the cloud service successfully receives the task data sent from the edge node, the cloud server allocates available computing resources to each task to achieve parallel processing. (1-_i,j)s_i,jc_i,jRepresenting the number of CPUs, i.e., computing resources, required to perform the tasks offloaded to the cloud server. By using

Representing cloud computing resources allocated to an ith mobile device served by a jth edge node. Thus, the computing latency of the cloud server can be expressed as:

And 4, aiming at the time delay analysis result, constructing a problem model:

the invention analyzes four different time delays existing in a cloud edge cooperative system, comprising the following steps: transmission delay of mobile device

Computing delay of edge nodes

Transmission delay of edge node

Computing latency of cloud server

To determine the overall delay incurred by each device, several reasonable assumptions are made as follows:

assume that 1: since the offloading of tasks depends on specific parameters of each task, such as its data size and the amount of work required for computation, the MEC server in the edge node can only offload tasks via the task offloading policy after receiving the task data.

Assume 2: in a real system, task computation may depend on a specific data structure and correlation between neighboring data, e.g. video analysis in a multimedia system. In order to ensure the reliability of the calculation result, the cloud server cannot start processing tasks until the transmission between the edge node and the cloud server is finished.

in the above formula:

representing the time delay incurred in executing the task at the edge node,

representing the time delay generated by the processing of the computing task at the cloud server.

The goal of the present invention is to achieve minimization of the total delay of all mobile devices by task offloading and computational resource allocation under limited resources, and finally, the goal can be expressed as:

in the above formula: y is_i,jIndicating the buffer status of the task.

Where (7a) and (7b) represent the impossibility of both the edge computing resources and the cloud computing resources allocated to the mobile device exceeding their own maximum number of computing resources. Optimization variables include the unload ratio of a task_i,j}, allocation of computing resources

Step 5, decomposing the constructed problem, including:

to solve the problem Q1, the structural characteristics of the function are first analyzed specifically. According to the formula (2), the transmission delay of the mobile device is only related to the nature of the transmitted task, and no additional optimization variable exists. Meanwhile, the transmission delay from the edge node to the cloud server, the calculation delay of the edge node and the calculation delay of the cloud server are irrelevant to the transmission delay of the mobile equipment. Thus, Q1 can be broken down into two parts:

the task offloading and computational resource allocation discussed in this invention only affects the solution of the latter part, so the problem can also be divided:

wherein

Step 6, the proposed solution to the constructed problem comprises:

(1) and (5) task caching strategy.

Before task offload decision making, the problem Q2 may be further simplified, and the cache vector Y may be solved using the task cache strategy presented in the previous section_i,j＝[y_1,1,y_1,2,...,y_i,j]The algorithm process is shown in fig. 4:

C denotes the number of requests of the task, and λ denotes the average occurrence rate of the task per unit time. Setting a vector group Y to be null during initialization, determining the task type according to Poisson distribution for each task, judging whether the task exists in a task cache table, if so, directly returning an execution result to the mobile equipment by the MEC server, and if so, Y_i,jIf the task exists in the table, the task does not need to be processed, and y is set to be 1_i,jSetting to 1, directly returning an execution result, storing the task into a task cache table, updating a task history record table, and if neither table records the task q_iThen y will be_i,jSetting to 0, traversing the subtask space in such a way, and finally obtaining a vector group Y_i,j＝[y_1,1,y_1,2,...,y_i,j]。

wherein

(2) And (4) task unloading strategy.

Considering that the problem Q3 is still too complex, it is still not directly solvable. Therefore, an optimal task unload ratio is determined first_i,jMaintained during the analysis

The value of (a) is not changed.

The delay can be calculated analytically

And task allocation ratio_i,jMonotonicity between them to determine the optimal division ratio

Firstly, according to the formula

Can deduce

Is monotonically increasing, with_i,jIs increased, therefore, when_i,j∈[0,1]Can obtain

Secondly, according to the formula

Can deduce

Is monotonically decreasing with_i,jIncrease and decrease of (2) can also obtain the value range of the function

Carefully observe the formula

In conjunction with the foregoing discussion, it can be seen that,

will follow first_i,jIs increased and decreased, and then follows_i,jIs increased. Therefore, when

When it is obtained

Minimum value of (d):

wherein

For ease of illustration, two important parameters are defined for each mobile device.

Based on the above definition, we can get the optimal task unloading strategy. Optimal task unloading strategy in cloud edge cooperative system

Can be expressed as:

in the above formula: mu.s_i,jDenotes the ratio of backhaul communication capacity to edge computing power, v_i,jRepresenting the ratio of cloud computing power to edge computing power.

Equation (10) shows that the optimal task offloading strategy depends only on two ratios: a normalized backhaul communication capacity and a normalized cloud computing capacity. Furthermore, the optimal task split strategy is determined by the harmonic mean of these two ratios. Can be easily verified at the edgeThe proportion of task data processed by a node will vary with μ_i,jOr v_i,jIs increased and decreased. Thus, when a mobile device is allocated few edge computing resources while at the same time having sufficient cloud computing, it should offload more task data to the cloud server. Conversely, if cloud computing resources are scarce, the edge node should be allowed to process more task data, for which it is easy to conclude that: offloading more tasks to a more powerful server is an effective solution to reduce the overall latency of the mobile device.

The 4 different systems that exist for the equation (10) are as follows:

the 1 st: communication-limited systems: the edge node and the cloud server have sufficient computing resources, but insufficient communication capacity.

In this case, it is preferable that the air conditioner,

this occurs when an edge node connects a large number of mobile devices and the capacity of the backhaul link is insufficient. The optimal task unload ratio can be simplified as follows:

equation (11) indicates that in this case, the optimal task offload ratio is determined only by the normalized backhaul communication capacity, and is not affected by the normalized cloud computing capacity. In this case, the communication resources are the main bottleneck to reduce the end-to-end delay of each device, so the edge computing power and cloud computing power size are relatively unimportant. Consider the special case when mu_i,j→ 0, optimal task unload ratio

The number is 1, which indicates that all computing tasks are executed at the edge node, and no task data is uploaded to the cloud server for execution. This is in line with our expectation that when the communication capacity of the backhaul link is insufficientIn extreme cases, tasks are not offloaded to cloud servers because offloading tasks to cloud servers exacerbates network congestion and results in longer transmission delays.

The 2 nd: a compute-constrained system: the edge nodes and cloud servers have sufficient communication capacity but insufficient computing resources.

In this case, it is preferable that the air conditioner,

the optimal task unload ratio can be simplified as follows:

from the formula, it can be seen that the optimal task offload ratio is determined only by the normalized cloud computing power. In this case, since the backhaul communication capacity is relatively sufficient, the time delay caused by the communication becomes very small, and the size of the calculation time delay determines the size of the overall time delay of the mobile device. Note that in this case, the edge node and the cloud server can be considered as a whole, and therefore should be based on the ratio of their computing power, v_i,jAnd splitting the tasks according to the proportion. More specifically, if the computing power of the edge node is greater than the computing power of the cloud server, v_i,jIf the number of the data is less than 1, unloading more data to the edge node; otherwise, more data should be offloaded to the cloud, at v_i,jIn the special case of 1, the data should be evenly distributed between the edge nodes and the cloud servers.

V can be obtained_i,j→ 0, this situation corresponds to a large scale small cell network where the cloud server serves many edge nodes. In this system, the optimal task offload ratio can be simplified as follows:

this indicates that the entire task should be performed on the edge node without having to be offloaded to the cloud server. Because offloading task data to the cloud server may create additional transmission delays and result in longer computational delays when the cloud computing power is much less than that of the edge nodes. Even if the normalized backhaul communication capacity is large enough, cloud computing latency can still have a major impact on overall latency. Thus, the entire task need only be offloaded to the edge node.

V can be obtained_i,j→ infinity, such a system corresponds to a scenario where the cloud server is powerful in computing power and the edge nodes are weak in computing power. V is to_i,jSubstituting → ∞ into the formula gives:

from the formula, it can be seen that the optimal task offload ratio is determined only by the normalized backhaul communication capacity. If backhaul communication capacity H_jComputing power of large or edge nodes

As the size becomes smaller, the optimal task unload rate will also decrease. This is because when v is_i,jAnd → infinity, the time delay generated by the cloud end to execute the task is negligible compared with the time delay generated by the edge node to execute the task. So when mu is_i,jIf the delay is less than 1, the overall delay of the mobile device will be dominated by the delay generated by the backhaul transmission, so that the proportion of task data unloaded to the edge side for processing is larger, namely

Otherwise, when μ_i,jAnd when the data is more than 1, more data should be unloaded to the cloud server for execution.

(3) And calculating a resource allocation strategy.

Firstly, the optimal task unloading ratio obtained in the previous chapter is obtained

Substituting into formula to calculate time delay

The simplification is as follows:

question Q3 may also be rewritten as:

theorem 1: q4 is a convex optimization problem

And (3) proving that: only by proving that the target function and the limiting condition of the target function are both convex functions, the theorem proves. Looking at the constraints (7a) and (7b), it can be seen that the constraints (7c) and (7d) are affine, which reflects their convexity. It is only necessary to prove that the objective function is also a convex function.

The hese matrix of (a) may be represented as:

all the former principals of W can be obtained by simple mathematical calculations:

according to linear algebra theory, when the front principal and minor components of a matrix are all positive definite, the matrix is a positive definite matrix. Thereby obtaining the result that,

is based on variables

And

a convex function of (a). And the objective function

Is the sum of a series of convex functions, from which it can be concluded that the objective function is also a convex function. Theorem 1 proves the syndrome.

According to theorem 1, it can be seen that Q4 satisfies the KKT condition, and then the lagrangian is used to find the optimal allocation value of the computing resources in the cloud edge coordination system:

wherein

And θ is the optimum variable for lagrange.

Example 2

The invention utilizes MATLAB simulation tool to carry out simulation verification on the proposed task unloading and resource allocation (Cloud-Edge-TORD) scheme based on Cloud Edge cooperation under the architecture of the Cloud Edge cooperation. To verify the superiority of this solution, the proposed solution will be compared with three other solutions, scheme 1: only edge, all tasks are unloaded to MEC server, all are processed by MEC server; scheme 2: all tasks are unloaded to the cloud server and are processed by the cloud server; scheme 3, Simple fixed group and edge, there is no effective task allocation strategy, and the task is simply divided into two halves, one half is executed by the MEC server, and the other half is executed by the cloud server.

The parameters of the embodiment are set as follows:

in the simulation experiment, assuming that each MEC server fixedly serves 25 mobile devices, the coverage radius of each MEC server is 500 meters. The size of the data volume of the task and the number of calculation cycles needed for execution are uniformly distributed, and the distribution range s_i,j∈[0.2,1]Mbits，c_i,j∈[500,2000]Hz. Computing power of MEC server connected with sc

Computing power F of cloud server^c∈[200,600]GHz, the transmission bandwidth B of the mobile device is 20MHz, and the channel gain and noise power of the communication are g_i,j＝10^-5W，σ²＝10^-9Transmission rate c of optical fiber is 1Gbps, capacity H of backhaul link_j∈[5,50]Mbps。

The performance of this example was analyzed as follows:

as shown in fig. 5, fig. 5 is a time delay comparison diagram of 4 schemes when the computing power of the cloud server is increased according to the present invention. The figure shows the time delay comparison of 4 schemes when the number of mobile devices changes. As can be seen from the figure, the delay of the Only Edge scheme is always stable in a smaller range, because each MEC server fixedly serves 25 mobile devices, and as the number of the mobile devices increases, the number of the MEC servers increases proportionally, so that the computing resources allocated to each mobile device are approximately constant, and the delay is always stable in a range; in addition, when the number of mobile devices is small, the performance of the Only closed scheme is better than that of the Only edge and Simple fixed and edge schemes, because when the number of mobile devices is small, the number of required edge nodes is small, and therefore, the cloud computing resources allocated to each mobile device are always more than that of the edge computing resources, but as the number of edge nodes is increased, the performance of the Only closed scheme is better than that of the Only closed scheme due to the limited cloud computing resources. Especially when the number of mobile devices is extremely large, the Only edge scheme has better performance than the Simple fixed group and edge scheme, which indicates that more computing tasks should be offloaded to the edge side rather than the cloud side; under all conditions, the Cloud-Edge-TORD scheme provided by the invention has optimal performance, and the task unloading and resource allocation strategy provided by the invention can be adopted to optimally allocate the computing capacity of the Edge node and the Cloud server by executing the Cloud Edge cooperation strategy, so that the minimization of the average system delay is realized.

Average system delay comparison of different schemes, as shown in fig. 6, fig. 6 is a delay comparison graph of 4 schemes when the computing power of the cloud server of the present invention is reduced. And time delay comparison of 4 schemes is given when the computing capacity of the cloud server is changed. When the computing capacity of the cloud server is increased, the time delay generated by the system adopting the Only closed scheme is greatly reduced. In this case, offloading the task to the cloud server is a better choice than offloading the task to the MEC server, which can effectively reduce a large amount of latency. In contrast, as the computing power of the cloud server becomes smaller, the performance of the Only edge scheme will be better than the Only closed scheme. Similar to fig. 5, the Cloud-Edge-TORD scheme proposed by the present invention is always optimal in all schemes.

Claims

1. The LTE electric power wireless private network task unloading and resource allocation method based on cloud edge cooperation is characterized by comprising the following steps: calculating all time delays of equipment and establishing a mathematical model for the problem of minimizing the total time delay of all mobile equipment by four different time delays in a cloud-edge cooperative system; considering task unloading schemes in four different system scenes, and solving an optimal allocation value of computing resources in the cloud edge cooperative system; the method comprises the following steps:

step 1, establishing a cloud edge cooperative system;

step 2, establishing an MEC cache model;

step 3, analyzing time delay existing in the cloud edge cooperative system;

step 4, aiming at the time delay analysis result, constructing a problem model;

step 5, solving the constructed problem;

and 6, providing a solution to the constructed problem.

2. The LTE electric power wireless private network task unloading and resource allocation method based on cloud edge coordination according to claim 1, which is characterized in that: step 1, establishing a cloud-edge coordination system, wherein the cloud-edge coordination system is composed of a central cloud server and M SCeNBs and is represented by a set M ═ {1, 2., J }; each SCeNB deploys an MEC server, and the MEC server uses limited resources for data processing, caching and storing; each combination of SCeNB and MEC server is called an edge node; within the coverage area of the jth SCeNB, there is a set χ_jX of the representation_jEach mobile device has a calculation task with different time delay requirements; assuming that each user is connected with a base station, and the specific connection relation is determined by a user communication strategy; each mobile device is connected with a corresponding base station through a wireless channel, and the edge nodes transmit data to the cloud server through different backhaul links; in the system, each computing task is supposed to be processed on an edge node and a cloud server; according to [1]The model in (1), assuming all tasks are of the same type and arrive at the same time, with q_i,j＝(s_i,j,c_i,j) Representing computational tasks generated by the ith mobile device connected to the jth edge node, where s_i,jRepresenting the size of the computational task, c_i,jIndicating the CPU computation cycle required to compute this task; representing the total number of tasks requested by the system in a certain time period by N; defining the computing capacity of the MEC server and the computing capacity of the cloud server in the jth edge node as

And F^c(ii) a The computing resources of each MEC server and cloud server are allocated to the mobile device by virtual machine technology.

3. The method of claim 1The LTE electric power wireless private network task unloading and resource allocation method based on cloud edge cooperation is characterized by comprising the following steps: step 2, establishing the MEC cache model, which comprises the following steps: for each task, defining the cache vector of the MEC server as Y_i,j＝[y_1,1,y_1,2,...,y_i,j](ii) a If y is_i,j1, it means that the MEC server has cached the calculation result of the task, and y_i,jIf the result is 0, the corresponding task result is not cached; when y is_i,jWhen the value is 1, the MEC server directly transmits the calculation result to the mobile equipment without calculation; since the transmission power of the SCeNB is greater than the transmission power of the mobile device, the data volume of the calculation result is less than the data volume of the task itself, and the transmission delay on the wireless downlink is ignored, and the transmission delay required by the MEC caching scheme is considered to be 0; and the MEC cache model directly returns the results of frequently requested tasks to the mobile equipment, so that the time delay is reduced and the energy consumption is reduced.

4. The LTE electric power wireless private network task unloading and resource allocation method based on cloud edge coordination according to claim 1, which is characterized in that: step 3, analyzing the time delay existing in the cloud edge coordination system, including: in a cloud-edge-end task allocation model, assuming that each mobile device sends a task request due to limited computing capacity and electric quantity of the mobile device, and judging whether a result of the task exists in a cache or not; if yes, directly returning a task result; if not, the mobile device uploads the task to the edge node; and determining whether the task is processed by the edge node alone or the edge node and the cloud server cooperatively by the corresponding edge node.

5. The LTE electric power wireless private network task unloading and resource allocation method based on cloud edge coordination according to claim 4, characterized in that: if the task is not existed, the mobile device uploads the tasks to the edge node, and the edge node is also responsible for determining the unloading proportion of each task at the cloud edge.

6. The LTE electric power wireless private network task unloading and resource allocation method based on cloud edge coordination according to claim 1, which is characterized in that: step 3, analyzing the time delay existing in the cloud edge coordination system, specifically comprising the following steps:

7. The LTE electric power wireless private network task unloading and resource allocation method based on cloud edge coordination according to claim 6, characterized in that: the delay existing in the calculation process comprises the following steps:

b. computing time delay of edge node: after the edge node successfully receives the complete computing task sent by the mobile equipment, the MEC server immediately executes an unloading strategy, and divides each computing task into two parts, wherein one part is executed by the MEC server, and the other part is executed by the cloud server; it is assumed that each calculation task is arbitrarily split without considering the task content, corresponding to the scene of video compression, speech recognition, and the like](ii) a Definition of_i,j∈[0,1]In order to calculate the division ratio of the task,_i,jrepresenting the proportion of the computing task data executed at the MEC server; by using

wherein

8. The LTE electric power wireless private network task unloading and resource allocation method based on cloud edge coordination according to claim 1, which is characterized in that: the different time delays existing in the cloud edge coordination system in the step 1 include: the method comprises the steps of transmitting delay of the mobile equipment, calculating delay of edge nodes, transmitting delay of the edge nodes and calculating delay of a cloud server; the total latency generated by each device includes:

in the above formula:

representing the time delay incurred in executing the task at the edge node,

Q1：

in the above formula: y is_i,jRepresenting the cache state of the task;

wherein (7a) and (7b) represent the number of computing resources that neither the edge computing resources nor the cloud computing resources allocated to the mobile device are likely to exceed their own maximum; optimization variables include the unload ratio of a task_i,j}, allocation of computing resources

9. The LTE electric power wireless private network task unloading and resource allocation method based on cloud edge coordination according to claim 1, which is characterized in that: step 5, decomposing the constructed problem, including:

q1 is broken down into two parts:

Q2：

wherein

10. The LTE electric power wireless private network task unloading and resource allocation method based on cloud edge coordination according to claim 1, which is characterized in that: the proposed solution to the constructed problem described in step 6 comprises:

(1) and (3) task caching strategy: before task unloading decision is carried out, the problem Q2 is further simplified, and a task caching strategy is used for solving a caching vector Y_i,j＝[y_1,1,y_1,2,...,y_i,j]The algorithm process is shown in fig. 4:

C represents the request times of the tasks, and lambda represents the average occurrence rate of the tasks in unit time; setting a vector group Y to be null during initialization, determining the task type according to Poisson distribution for each task, judging whether the task exists in a task cache table, if so, directly returning an execution result to the mobile equipment by the MEC server, and if so, Y_i,jIf not, continuing to judge whether the task exists in the task history record table; if present in the table, will y_i,jSetting the value to be 1, directly returning an execution result, storing the task into a task cache table, and updating a task history record table; if neither table has recorded task q_iThen y will be_i,jSetting to 0, traversing the subtask space in such a way, and finally obtaining a vector group Y_i,j＝[y_1,1,y_1,2,...,y_i,j]；

Q3：