CN112416554A - Task migration method and device, electronic equipment and storage medium - Google Patents

Task migration method and device, electronic equipment and storage medium Download PDF

Info

Publication number
CN112416554A
CN112416554A CN202011312378.7A CN202011312378A CN112416554A CN 112416554 A CN112416554 A CN 112416554A CN 202011312378 A CN202011312378 A CN 202011312378A CN 112416554 A CN112416554 A CN 112416554A
Authority
CN
China
Prior art keywords
task
migrated
edge computing
user equipment
task migration
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202011312378.7A
Other languages
Chinese (zh)
Other versions
CN112416554B (en
Inventor
丰雷
高娴
王亚鹏
谢坤宜
武帅
刘涛
王慧
历明
王紫程
谢奥
周雨
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing University of Posts and Telecommunications
Inspur Software Technology Co Ltd
Inspur Tianyuan Communication Information System Co Ltd
Original Assignee
Beijing University of Posts and Telecommunications
Inspur Software Technology Co Ltd
Inspur Tianyuan Communication Information System Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing University of Posts and Telecommunications, Inspur Software Technology Co Ltd, Inspur Tianyuan Communication Information System Co Ltd filed Critical Beijing University of Posts and Telecommunications
Priority to CN202011312378.7A priority Critical patent/CN112416554B/en
Publication of CN112416554A publication Critical patent/CN112416554A/en
Application granted granted Critical
Publication of CN112416554B publication Critical patent/CN112416554B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/48—Program initiating; Program switching, e.g. by interrupt
    • G06F9/4806—Task transfer initiation or dispatching
    • G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
    • G06F9/485—Task life-cycle, e.g. stopping, restarting, resuming execution
    • G06F9/4856—Task life-cycle, e.g. stopping, restarting, resuming execution resumption being on a different machine, e.g. task migration, virtual machine migration
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/48—Program initiating; Program switching, e.g. by interrupt
    • G06F9/4806—Task transfer initiation or dispatching
    • G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
    • G06F9/4881—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The application provides a task migration method, a task migration device, an electronic device and a storage medium, wherein the method comprises the following steps: acquiring task information to be migrated, which is generated by a plurality of user equipment terminals, available bandwidth between each user equipment terminal and a base station and available resources of an edge computing server; the task information to be migrated comprises: computing resource requirements and bandwidth requirements of the tasks to be migrated; generating a plurality of task migration strategies according to the information of the tasks to be migrated, the available bandwidth between each user equipment terminal and the base station and the available resources of the edge computing server within a preset constraint condition; and determining a target task migration strategy according to the total system energy consumption corresponding to each task migration strategy. The total energy consumption of the system of various realizable task migration strategies is estimated, the target task migration strategy is determined, the energy consumption of the mobile edge computing system is prevented from exceeding the standard, the reliability of the mobile edge computing system is improved, and a foundation is laid for improving the task processing efficiency of the mobile edge computing system.

Description

Task migration method and device, electronic equipment and storage medium
Technical Field
The present application relates to the field of distributed computing technologies, and in particular, to a task migration method and apparatus, an electronic device, and a storage medium.
Background
At present, the 5G communication technology promotes the information society to enter the era of interconnection of everything, and the quantity of terminal equipment and data volume of the Internet of things are exponentially increased. In order to meet the data processing requirements, a mobile edge computing system is usually used to perform task migration and processing on massive data generated by a user equipment terminal.
In the prior art, a mobile edge computing system usually performs task migration according to a fixed task migration rule, for example, a task migration policy is determined according to factors such as migration selection of a user equipment terminal, transmission power, size and deadline of a task, computing resources of the user equipment terminal and computing resources of an edge computing server.
However, since the user equipment terminals are various, if task migration is performed on tasks generated by each user equipment terminal only according to a fixed task migration rule, some tasks with low time delay and high reliability cannot be processed in time, and risks such as time delay and over-standard energy consumption of a mobile edge computing system are caused. Therefore, a task migration method capable of relieving energy consumption pressure of a mobile edge computing system when multiple services are concurrent is urgently needed, and the method has important significance for improving task processing efficiency of the mobile edge computing system.
Disclosure of Invention
The application provides a task migration method, a task migration device, an electronic device and a storage medium, which are used for solving the defects that a task migration strategy determined by the task migration method in the prior art may cause the energy consumption of a mobile edge computing system to exceed the standard and the like.
A first aspect of the present application provides a task migration method, including:
acquiring task information to be migrated of tasks to be migrated, which is generated by a plurality of user equipment terminals, available bandwidth between each user equipment terminal and a base station and available resources of the edge computing server; wherein the task information to be migrated includes: computing resource requirements and bandwidth requirements of the tasks to be migrated;
generating a plurality of task migration strategies according to the information of the tasks to be migrated, the available bandwidth between each user equipment terminal and the base station and the available resources of the edge computing server within a preset constraint condition;
and determining a target task migration strategy according to the total system energy consumption corresponding to each task migration strategy.
Optionally, the generating a plurality of task migration policies according to the information of the task to be migrated, the available bandwidth between each user equipment terminal and the base station, and the available resources of the edge computing server includes:
generating a plurality of task migration decisions corresponding to the tasks to be migrated according to the bandwidth requirements of the tasks to be migrated and the available bandwidth between each user equipment terminal and the base station; the task migration decision comprises an execution destination of each task to be migrated, and the execution destination comprises a user equipment terminal and an edge computing server;
when the execution destination of the task to be migrated is the edge computing server, determining that the task to be migrated is a first task;
acquiring a service quality requirement of the first task, and determining the priority of the first task according to the service quality requirement; determining the allocation amount of the available resources of the edge computing server according to the priority of each first task;
and generating a plurality of task migration strategies according to a plurality of task migration decisions corresponding to the tasks to be migrated and the allocation condition of the available resources of the edge computing server.
Optionally, the determining a target task migration policy according to the total system energy consumption corresponding to each task migration policy includes:
determining the energy consumption of the edge computing server according to the transmission power and the transmission delay of the user equipment terminal corresponding to each first task and the preset attribute information of the edge computing server;
determining local energy consumption of each user equipment terminal corresponding to a second task according to the computing resource demand of the second task except the first task and the execution cycle frequency of the user equipment terminal corresponding to the second task;
determining total system energy consumption corresponding to each task migration strategy according to the edge computing server energy consumption and the local energy consumption;
and determining the task migration strategy with the minimum system total energy consumption as the target task migration strategy.
Optionally, the determining, according to the energy consumption of the edge computing server and the local energy consumption, total system energy consumption corresponding to each task migration policy includes:
determining the total energy consumption of the system corresponding to any task migration strategy according to the following formula:
Figure BDA0002790203880000031
wherein E isallRepresenting the total system energy consumption corresponding to the task migration strategy, T representing the total system consumption time, thetaiIndicating an execution destination of an ith task to be migrated, wherein when the execution destination of the ith task to be migrated is an edge computing server, thetaiWhen the execution destination of the ith task to be migrated is the corresponding user equipment terminal, θi＝0，
Figure BDA0002790203880000032
Representing the energy consumption of the edge computing server corresponding to each first task,
Figure BDA0002790203880000033
and local energy consumption corresponding to each second task is represented.
Optionally, the determining a target task migration policy according to the total system energy consumption corresponding to each task migration policy includes:
calculating the reward value of each task migration strategy according to the total energy consumption of the system based on the following formula:
Figure BDA0002790203880000034
where r represents the prize value telocalIndicating the total energy consumption of the system when the execution destinations of the tasks to be migrated are all corresponding user equipment terminals, EallRepresenting the total energy consumption of the system corresponding to the task migration strategy;
and determining the task migration strategy with the highest reward value as the target task migration strategy.
Optionally, the method further includes:
constructing a waiting queue according to a preset caching requirement;
and storing each first task to the waiting queue according to the generation sequence of each first task.
Optionally, the preset constraint condition includes:
the total transmission delay of each task migration strategy is smaller than a preset total transmission delay threshold, the execution cycle frequency of each user equipment terminal is smaller than a preset execution cycle frequency threshold, and/or the execution cycle number of each user equipment terminal is not larger than the execution cycle number of the edge computing server.
A second aspect of the present application provides a task migration apparatus, including:
the system comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring task information to be migrated of tasks generated by a plurality of user equipment terminals, available bandwidth between each user equipment terminal and a base station and available resources of an edge computing server; wherein the task information to be migrated includes: computing resource requirements and bandwidth requirements of the tasks to be migrated;
a generating module, configured to generate a plurality of task migration strategies according to the information of the tasks to be migrated, the available bandwidth between each user equipment terminal and the base station, and the available resources of the edge computing server within a preset constraint condition;
and the determining module is used for determining the target task migration strategy according to the total system energy consumption corresponding to each task migration strategy.
Optionally, the generating module is specifically configured to:
generating a plurality of task migration decisions corresponding to the tasks to be migrated according to the bandwidth requirements of the tasks to be migrated and the available bandwidth between each user equipment terminal and the base station; the task migration decision comprises an execution destination of each task to be migrated, and the execution destination comprises a user equipment terminal and an edge computing server;
when the execution destination of the task to be migrated is the edge computing server, determining that the task to be migrated is a first task;
acquiring a service quality requirement of the first task, and determining the priority of the first task according to the service quality requirement;
determining the allocation amount of the available resources of the edge computing server according to the priority of each first task;
and generating a plurality of task migration strategies according to a plurality of task migration decisions corresponding to the tasks to be migrated and the allocation condition of the available resources of the edge computing server.
Optionally, the determining module is specifically configured to:
determining the energy consumption of the edge computing server according to the transmission power and the transmission delay of the user equipment terminal corresponding to each first task and the preset attribute information of the edge computing server;
determining local energy consumption of each user equipment terminal corresponding to a second task according to the computing resource demand of the second task except the first task and the execution cycle frequency of the user equipment terminal corresponding to the second task;
determining total system energy consumption corresponding to each task migration strategy according to the edge computing server energy consumption and the local energy consumption;
and determining the task migration strategy with the minimum system total energy consumption as the target task migration strategy.
Optionally, the determining module is specifically configured to:
determining the total energy consumption of the system corresponding to any task migration strategy according to the following formula:
Figure BDA0002790203880000041
wherein E isallRepresenting the total system energy consumption corresponding to the task migration strategy, T representing the total system consumption time, thetaiIndicating an execution destination of an ith task to be migrated, wherein when the execution destination of the ith task to be migrated is an edge computing server, thetaiWhen the execution destination of the ith task to be migrated is the corresponding user equipment terminal, θi＝0，
Figure BDA0002790203880000042
Representing the energy consumption of the edge computing server corresponding to each first task,
Figure BDA0002790203880000051
and local energy consumption corresponding to each second task is represented.
Optionally, the determining module is further specifically configured to:
calculating the reward value of each task migration strategy according to the total energy consumption of the system based on the following formula:
Figure BDA0002790203880000052
where r represents the prize value telocalIndicating the total energy consumption of the system when the execution destinations of the tasks to be migrated are all corresponding user equipment terminals, EallRepresenting the total energy consumption of the system corresponding to the task migration strategy;
and determining the task migration strategy with the highest reward value as the target task migration strategy.
Optionally, the generating module is further configured to:
constructing a waiting queue according to a preset caching requirement;
and storing each first task to the waiting queue according to the generation sequence of each first task.
Optionally, the preset constraint condition includes:
the total transmission delay of each task migration strategy is smaller than a preset total transmission delay threshold, the execution cycle frequency of each user equipment terminal is smaller than a preset execution cycle frequency threshold, and/or the execution cycle number of each user equipment terminal is not larger than the execution cycle number of the edge computing server.
A third aspect of the present application provides an electronic device, comprising: at least one processor and memory;
the memory stores computer-executable instructions;
the at least one processor executes computer-executable instructions stored by the memory to cause the at least one processor to perform the method as set forth in the first aspect above and in various possible designs of the first aspect.
A fourth aspect of the present application provides a computer-readable storage medium having stored thereon computer-executable instructions that, when executed by a processor, implement a method as set forth in the first aspect and various possible designs of the first aspect.
This application technical scheme has following advantage:
according to the task migration method, the task migration device, the electronic equipment and the storage medium, task information to be migrated, which is generated by a plurality of user equipment terminals, available bandwidth between each user equipment terminal and a base station and available resources of an edge computing server are obtained; the task information to be migrated comprises: computing resource requirements and bandwidth requirements of the tasks to be migrated; generating a plurality of task migration strategies according to the information of the tasks to be migrated, the available bandwidth between each user equipment terminal and the base station and the available resources of the edge computing server within a preset constraint condition; and determining a target task migration strategy according to the total system energy consumption corresponding to each task migration strategy. According to the task migration method provided by the scheme, the total system energy consumption of various achievable task migration strategies is estimated, so that the task migration strategy with the minimum total system energy consumption is screened out, the risk that the energy consumption of the mobile edge computing system exceeds the standard is avoided, the reliability of the mobile edge computing system is improved, and a foundation is laid for improving the task processing efficiency of the mobile edge computing system.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art according to these drawings.
FIG. 1 is a schematic diagram of a network architecture of a mobile edge computing system according to an embodiment of the present application;
fig. 2 is a schematic flowchart of a task migration method according to an embodiment of the present application;
fig. 3 is a schematic structural diagram of a task migration apparatus according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
With the above figures, there are shown specific embodiments of the present application, which will be described in more detail below. These drawings and written description are not intended to limit the scope of the disclosed concepts in any way, but rather to illustrate the concepts of the disclosure to those skilled in the art by reference to specific embodiments.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The terms referred to in this application are explained first:
and (3) task migration decision: specifically, the task to be migrated is migrated to the edge computing server to be executed, or executed locally (at the user equipment terminal), that is, the execution destination of the task to be migrated.
And (3) task scheduling decision: refers to the allocation of the available resources of the edge compute server for performing the first task destined for the edge compute service.
And (3) task migration strategy: including task migration decisions and task scheduling decisions.
Furthermore, the terms "first", "second", etc. are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. In the description of the following examples, "plurality" means two or more unless specifically limited otherwise.
In the prior art, a mobile edge computing system usually performs task migration according to a fixed task migration rule, for example, a task migration policy is determined according to factors such as migration selection of a user equipment terminal, transmission power, size and deadline of a task, computing resources of the user equipment terminal and computing resources of an edge computing server. However, since the user equipment terminals are various, if task migration is performed on tasks generated by each user equipment terminal only according to a fixed task migration rule, some tasks with low time delay and high reliability cannot be processed in time, and risks such as time delay and over-standard energy consumption of a mobile edge computing system are caused. In order to solve the above problems, a user equipment terminal, a task migration method, a task migration apparatus, an electronic device, and a storage medium according to embodiments of the present application calculate available resources of a server by obtaining task information to be migrated, which is generated by a plurality of user equipment terminals, available bandwidths between each user equipment terminal and a base station, and an edge; the task information to be migrated comprises: computing resource requirements and bandwidth requirements of the tasks to be migrated; generating a plurality of task migration strategies according to the information of the tasks to be migrated, the available bandwidth between each user equipment terminal and the base station and the available resources of the edge computing server within a preset constraint condition; and determining a target task migration strategy according to the total system energy consumption corresponding to each task migration strategy. According to the task migration method provided by the scheme, the total system energy consumption of various achievable task migration strategies is estimated, so that the task migration strategy with the minimum total system energy consumption is screened out, the risk that the energy consumption of the mobile edge computing system exceeds the standard is avoided, the reliability of the mobile edge computing system is improved, and a foundation is laid for improving the task processing efficiency of the mobile edge computing system.
The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present invention will be described below with reference to the accompanying drawings.
First, a network structure of a mobile edge computing system on which the present application is based will be explained:
the task migration method, the task migration device, the electronic equipment and the storage medium are suitable for determining the task migration strategy with the minimum total energy consumption of the system, and the defects that the task migration strategy determined by the task migration method in the prior art may cause the energy consumption of a mobile edge computing system to exceed the standard and the like are overcome. Fig. 1 is a schematic diagram of a network structure of a mobile edge computing system based on the embodiment of the present application, which mainly includes a plurality of user equipment terminals, an edge computing server, a base station, and an electronic device for performing task migration. Specifically, after the user equipment terminals generate the tasks to be migrated, before the generated tasks to be migrated are sent to the edge computing server through the base station, the user equipment terminals send task information to be migrated corresponding to the tasks to be migrated to the electronic equipment, and the electronic equipment determines a target task migration strategy according to the task information to be migrated, available bandwidth between each user equipment terminal and the base station, and available resources of the edge computing server.
The embodiment of the application provides a task migration method, which is used for determining a task migration strategy with the minimum system total energy consumption. The execution subject of the embodiment of the present application is an electronic device, such as a server, a desktop computer, a notebook computer, a tablet computer, and other electronic devices that can be used for task migration.
As shown in fig. 2, a schematic flowchart of a task migration method provided in an embodiment of the present application is shown, where the method includes:
step 201, acquiring information of tasks to be migrated generated by a plurality of user equipment terminals, available bandwidth between each user equipment terminal and a base station, and available resources of an edge computing server.
The task information to be migrated comprises: the computational resource requirements and bandwidth requirements of the task to be migrated.
It should be noted that the computing resource requirement refers to the computing resource consumed when the task is executed, and the bandwidth requirement refers to the bandwidth required when the task is transmitted to the edge computing server through the base station.
Step 202, in a preset constraint condition, generating a plurality of task migration strategies according to the information of the tasks to be migrated, the available bandwidth between each user equipment terminal and the base station and the available resources of the edge computing server.
The preset constraint condition may be that the total transmission delay of each task migration policy is smaller than a preset total transmission delay threshold, the execution cycle frequency of each user equipment terminal is smaller than a preset execution cycle frequency threshold, and/or the execution cycle number of each user equipment terminal is not greater than the execution cycle number of the edge computing server.
Similarly, the preset constraint conditions also include that the allocation proportion of the bandwidth of the base station is within the preset allocation proportion range, the waiting queue is stable and reliable, and the like, and the constraint conditions may be specifically set and combined according to the actual situation, which is not limited in the embodiment of the present application.
Specifically, the task migration policy includes an execution destination of each task to be migrated, and an allocation situation of available resources of the edge computing server when the execution destination is the edge computing server.
And 203, determining a target task migration strategy according to the total system energy consumption corresponding to each task migration strategy.
Specifically, the total system energy consumption is the sum of the edge computing server energy consumption and the local energy consumption, and specifically, the task migration policy with the minimum total system energy consumption is determined as the target task migration policy.
On the basis of the foregoing embodiments, in order to improve the reliability of each generated task migration policy, as an implementable manner, on the basis of the foregoing embodiments, in an embodiment, a plurality of task migration policies are generated according to task information to be migrated, an available bandwidth between each user equipment terminal and a base station, and an available resource of an edge computing server (step 202), including:
step 2021, generating a plurality of task migration decisions corresponding to the tasks to be migrated according to the bandwidth requirements of the tasks to be migrated and the available bandwidth between each user equipment terminal and the base station; the task migration decision comprises an execution destination of each task to be migrated, and the execution destination comprises a user equipment terminal and an edge computing server;
step 2022, when the execution destination of the task to be migrated is the edge computing server, determining that the task to be migrated is the first task;
step 2023, obtaining the qos requirement of the first task, and determining the priority of the first task according to the qos requirement;
step 2024, determining the allocation amount of the available resources of the edge computing server according to the priority of each first task;
step 2025, generating a plurality of task migration policies according to the plurality of task migration decisions corresponding to the tasks to be migrated and the allocation conditions of the available resources of the edge computing server.
Specifically, whether the execution destination of each task to be migrated can be an edge calculation server or not is important to consider whether the bandwidth between each user equipment terminal and the base station can meet the bandwidth requirement or not, and only under the condition that the bandwidth requirement is met, the task to be migrated generated by the user equipment terminal can be transmitted to the edge calculation server and further executed based on the edge calculation server.
It should be explained that the quality of service requirement of the first task may include a task processing time limit and a reliability requirement corresponding to the first task.
Specifically, the priority of each first task may be calculated based on the following formula:
Priority＝α×Time+β×Reliability
wherein Priority represents a Priority corresponding to the first task; the Time represents the task processing Time limit, also called the maximum tolerance value of the task processing event; reliability represents Reliability requirements, specifically the Reliability requirements for an edge computing server; α and β respectively represent coefficient values of the two, where α + β is 1, specifically representing the proportion of the emphasis considered by the moving edge computing system for the two attributes. The task priority refers to a priority value of a task in a mobile edge computing system, and the task with the high priority can be migrated preferentially and obtain more wireless resource bandwidth.
Correspondingly, in an embodiment, the energy consumption of the edge computing server may be determined according to the transmission power and the transmission delay of the user equipment terminal corresponding to each first task and preset attribute information of the edge computing server; determining the local energy consumption of each user equipment terminal corresponding to a second task according to the computing resource demand of the second task except for the first task and the execution cycle frequency of the user equipment terminal corresponding to the second task; determining the total system energy consumption corresponding to each task migration strategy according to the energy consumption of the edge computing server and the local energy consumption; and determining the task migration strategy with the minimum system total energy consumption as a target task migration strategy.
It should be explained that, for the first task, when the ith user equipment migrates the first task ai(t) upon reaching the edge compute server, the entire migration process is typically divided into three steps. Step 1: the ith user equipment terminal uploads a first task AiAnd (t) forwarding the input data to the base station through the wireless private network, and forwarding the data to the edge computing server by the base station. Step 2: task Ai(t) entering a wait queue, waiting for the edge compute server to allocate computing resources to perform the compute task. And step 3: and the edge computing server returns the processed task to the ith user equipment terminal. For the last step of the migration computation, the time required is the return delay of the processing result, denoted tdown. However, in the prior art, the download data rate is generally high, and the data size of the task processing result is much smaller than that of the input data, so that the time delay of step 3 can be ignored in the embodiment of the present application.
Specifically, if a plurality of user equipment terminals select to migrate a first task to the edge computing server within the time slot τ, the bandwidth is proportionally allocated to the edge aggregation segment device for uploading data according to the priority ranking result of each first task. The expression of transmission rates must also add extreme delays and transmission reliability requirements. The embodiment of the application considers the influence of the transmission error rate, and is different from the classical shannon formula. The first task generated by the ith user equipment terminal in the time slot tau is Ai(t) the transfer rate of migration to the edge server can be expressed as:
vi(t)＝γi(t)·BW·(log2(1+δi(t))-ψ)
wherein v isi(t) represents the transmission rate, BW represents the bandwidth allocated to the first task, δi(t) represents a channel parameter variable, γ, between the user equipment terminal and the edge computing serveri(t) A is the task generated by the ith user equipment terminal in the time slot taui(t) the proportion of bandwidth allocated in the migration to the edge calculation server, ψ represents the reliability of the channel.
To be explained, δi(t) may be specifically expressed as:
Figure BDA0002790203880000101
wherein p istxFor migrating the first task A for the ith user equipment terminal in time slot taui(t) transmission power to edge server, g0For migrating the first task A for the ith user equipment terminal in time slot taui(t) channel gain to edge server Process, N0For migrating the first task A for the ith user equipment terminal in time slot taui(t) additive gaussian channel noise power to edge calculation server process.
To be further explained, γiThe value of (t) is affected by the priority and can be expressed as:
Figure BDA0002790203880000111
wherein, gamma is more than or equal to 0i(t) is less than or equal to 1, and ∑i∈Nγi(t)＝1，Pr，i(t) represents the first task Ai(t) corresponding priority. Thetai(t) 0 or 1, for the first task a whose execution destination is the edge compute serveri(t)，θi(t) is 1, otherwise it is 0.
To be further explained, ψ may be specifically expressed as:
Figure BDA0002790203880000112
wherein, VτWhich is the channel dispersion, can be approximated as 1, which indicates the random variation of the channel with respect to a deterministic channel of the same capacity. Q-1(. represents the inverse of the Gaussian Q function, εi(t) denotes a first task A generated by the ith user equipment terminal within the slot τi(t) transmission error rate when migrating to the edge computing server.
Specifically, in step 1, a first task Ai(t) migration from the edge sink terminal equipment to the edge computing server results in a transmission delay of tupWhen the edge compute server performs a task, assuming all available resources are occupied, the expression follows:
Figure BDA0002790203880000113
wherein, bi(t) denotes a first task A that needs to be performedi(t) size of data volume, vi(t) denotes that the first task generated by the ith user equipment terminal is Ai(t) transfer rate of migration to edge server.
For step 2, since the available resources of the edge computing server are limited, all the first tasks cannot be processed at the same time, and the first tasks that arrive and are not processed enter the waiting queue. Suppose there is time required in the edge compute server that is the compute delay t of the edge compute serverexeAnd average latency of the first task in the wait queue
Figure BDA0002790203880000114
Computation delay t of edge computation serverexeCan be expressed as:
Figure BDA0002790203880000115
wherein d isi(t) represents the computing resources required for the task to be migrated (first task) to be executed locally (corresponding user equipment terminal), i.e. the computing resource requirements of the task to be migrated, fs，i(t) denotes the first task Ai(t) calculating an allocation amount of the available resource allocated in the edge calculation server.
It should be explained that, according to the Little law in the prior art, task a of each user equipment terminaliThe average latency experienced by (t) is proportional to its average queue length of the task buffers (waiting queues) in the edge computing servers. Therefore, the average queue length of the task buffer of each user equipment terminal is used as one of the measures of the execution delay, which can be specifically expressed as:
Figure BDA0002790203880000121
wherein the content of the first and second substances,
Figure BDA0002790203880000122
representing a first task A generated by an i-th user equipment terminali(t) average latency, EQi(t)]Indicating the expected value corresponding to the queue length of the waiting queue.
In particular, in the embodiments of the present application
Figure BDA0002790203880000123
A first task A representing the ith user equipment terminali(t) the time delay of the whole process of migrating to the edge computing server, i.e. the transmission time delay, can be specifically expressed as:
Figure BDA0002790203880000124
in particular, in the examples of the present application
Figure BDA0002790203880000125
A first task A representing the ith user equipment terminal in time slot taui(t) the resulting energy consumption in the entire computation process of migration to the edge compute server, which occurs mainly during data transfer and task processing, the expression is as follows:
Figure BDA0002790203880000126
wherein, κserRepresenting the effective switched capacitance of the CPU core in an edge compute server, which is dependent on the chip architecture, i.e., κserAnd calculating the attribute information of the server for the preset edge.
Further, for the second service, the CPU of the ith ue calculates the cycle frequency (execution cycle frequency) as fi lCannot exceed its maximum value
Figure BDA0002790203880000127
By collections
Figure BDA0002790203880000128
The calculation cycle frequency of the local CPU is shown, and the calculation speed of the edge calculation server is assumed to be much faster than the maximum calculation speed of the local CPU, so that the time delay of the second task in the local CPU is the calculation time of the task. In the examples of this application
Figure BDA0002790203880000129
Representing task Ai(t) local latency, expressed as follows:
Figure BDA00027902038800001210
wherein d isi(t) represents the computational resources required for the task to be migrated (second task) to execute locally.
Further, the local energy consumption of the ith ue terminal may be calculated according to the following formula:
Figure BDA00027902038800001211
wherein the content of the first and second substances,
Figure BDA00027902038800001212
means the energy per CPU calculation cycle (execution cycle) to complete the second taskThe source is consumed.
Specifically, in one embodiment, the total system energy consumption corresponding to any task migration policy may be determined according to the following formula:
Figure BDA00027902038800001213
wherein E isallThe total system energy consumption corresponding to the task migration strategy is represented, T represents the total system consumption time, and thetaiIndicating an execution destination of the ith task to be migrated, wherein when the execution destination of the ith task to be migrated is the edge calculation server, thetaiWhen the execution destination of the ith task to be migrated is the corresponding user equipment terminal, θ 1i＝0，
Figure BDA0002790203880000131
Representing the energy consumption of the edge computing server corresponding to each first task,
Figure BDA0002790203880000132
and local energy consumption corresponding to each second task is represented.
Specifically, in an embodiment, in order to improve the management efficiency of the task to be migrated, a waiting queue may be constructed according to a preset cache requirement; and storing each first task to a waiting queue according to the generation sequence of each first task.
Specifically, the waiting queue may be constructed according to a preset caching requirement of the edge computing server.
Illustratively, for each time slot τ, the task generated by the ith ue terminal moves to the task of the edge computing server. Suppose Qi(t) is the queue length of the ith user equipment terminal in the edge computing server within slot τ, which is assumed to have a sufficiently large capacity. Ci(t) represents the task scheduling decision of the edge computing server in the time slot τ, i.e. the number of tasks being processed in the edge computing node by the ith user equipment terminal in the time slot τ. By queue length Q within slot τi(t), unloadingArrived task Bi(t) and task scheduling decisions Ci(t) the queue length Q of the next slot τ can be derivedi(t +1), the expression of which is:
Qi(t+1)＝max{Qi(t)-Ci(t)，0}+Bi(t) t∈{0，1，2，...}
wherein it is assumed that the task buffer is initially empty, i.e. Qi(0) 0, i ∈ N. The computing resources of the computing nodes of the edge computing server may be allocated to the first task migrated from the different user equipment terminal, and therefore the edge computing server scheduling decision should satisfy the following condition:
∑i∈NCi(t)·bi(t)·≤Fser·τ (4)
this means that the number of execution cycles of the user equipment terminal required to complete the target task migration policy should not be greater than the execution cycle of the edge computing server.
The assigned available resources of the edge computing server are also distributed in proportion according to the priority value sequence of each first task in the task set of the ith user equipment in the edge computing server in the time slot tau, and the computing resources distributed to the first task of the ith user equipment terminal in the edge computing server in the time slot tau in which the ith user equipment terminal is in processing are fs，i(t), the expression is as follows:
Figure BDA0002790203880000133
wherein, FserRepresenting the edge compute server available resources.
In particular, in one embodiment, to further improve task migration efficiency and task processing efficiency of the mobile edge computing system. The task migration method provided by the above embodiment may be implemented by using machine learning algorithms such as a DQN algorithm, a WoLF-PHC algorithm, an A3C algorithm, and the like, and the embodiment of the present application is not limited specifically.
Specifically, in an embodiment, in order to adapt to various machine learning algorithms and further improve task migration efficiency, the reward value of each task migration strategy can be calculated according to the total energy consumption of the system based on the following formula:
Figure BDA0002790203880000141
where r represents the prize value telocalIndicating the total energy consumption of the system when the execution destinations of the tasks to be migrated are all corresponding user equipment terminals, EallRepresenting the total energy consumption of the system corresponding to the task migration strategy; and determining the task migration strategy with the highest reward value as a target task migration strategy.
Specifically, a task migration model with deep learning capability can be constructed based on each machine learning algorithm, wherein three elements of the model of task migration are state, action and reward respectively. Wherein, the status should include the available bandwidth between each ue and the base station and the available resources of the edge computing server. The action may specifically include a task migration decision Λ (t) and a decision vector corresponding to the task migration policy c (t). Task migration decision vectors with, for example, N user equipment terminals each
Figure BDA0002790203880000142
And task scheduling decision vector
Figure BDA0002790203880000143
Thus, an action vector (task migration policy) may use a set of task migration decision vectors and task scheduling decision vectors [ θ [ ]1，θ2…，θN，c1，c2…，cN]To indicate. Wherein, the reward means that for each iteration, the agent environment will change the state after executing each possible action, and then obtain a reward value. Generally, the reward function should be related to the desired function. The task migration method optimization problem provided by the embodiment of the application aims to obtain minimum total system energy consumption, and the machine learning aims to obtain maximum long-term returnThe prize value should therefore be inversely related to the amount of total energy consumed by the system.
In practical applications, if there are more and more user equipment terminals, the motion space will increase rapidly, and in order to limit the size of the motion space, we propose a preprocessing step before the machine learning process. For the user equipment terminal, if
Figure BDA0002790203880000144
That is, the local execution cannot meet the requirement of the delay threshold, and only the task to be migrated can be uploaded to the edge computing server for execution, so that the theta corresponding to the task to be migrated can be directly obtainediThe value is fixed to 1, and the distributed edge computing resources meet the preset constraint, so that the possible value of the action can be reduced, the decision space of the reinforcement learning agent is limited, and the task migration efficiency is improved.
Illustratively, when the DQN algorithm is used to implement the task migration method provided in the above embodiment, all Q values are stored in a Q table, and the matrix Q (s, a) will be very large, especially in this environment, the state space is three-dimensional, the number of user equipment terminals also needs one dimension, and three kinds of decisions are included, and the dimension of the Q table has reached a certain height. It is therefore difficult to obtain enough samples to traverse each state, which results in an inefficient algorithm, i.e., a dimensional disaster of Q-learning, and thus, compared to Q-learning algorithm, DQN is no longer decided by storing a complete Q table, but rather estimates Q (s, a) using a neural network.
The specific DQN algorithm flow is as follows: store certain(s)t，at，rt+1，st+1) A pool of experiences wherein stIs the current state, atIs at stAction performed in the state, st+1Is at atNext state reached under action, rt+1Is stState to st+1In the state, a small batch of batch, s is sampled from the reward experience pool given by the environmentt、st+1Put into the network, the function of the state-action value in a and st+1，st+1maxQ(s) oft+1) next _ qvalue, the action value function of the desired decision is calculated as shown in the following equation:
expect _ qvalue ═ forward + γ × next _ qvalue, Loss ═ Loss (qvalue, expected _ qvalue), where the Loss selects the norm Loss, i.e., the squared Loss. The model provided by the embodiment of the application considers one multi-user and two main types of decisions: task migration decision vector
Figure BDA0002790203880000151
And task scheduling decision vector
Figure BDA0002790203880000152
There are two decisions for each decision theta, since thetaiE {0, 1 }. Similarly, for each task scheduling decision, it can be discretized, i.e., ciE.g. {0, 1, 2, 3 }. The number of nodes at the output level of the NN should then be equal to the number of nodes 2+ the number of nodes 4, each node representing the maximum long-term reward for task migration decisions of a certain user equipment terminal. The network can be constructed by adopting a full connection layer, the output part of the network comprises all decision parts, and the embodiment of the application uses a small batch random gradient (MBGD) descent method when updating NN according to the setting of the neural network. It is a neutralization method for BGD and SGD. The idea is as follows: each iteration updates the parameters using a small batch of samples. Suppose hθ(x) Is the function to be fitted, and theta in the J (theta) loss function is a parameter, the value to be solved iteratively. Solving for theta results in a complete function to be fitted. Where m is the number of training sets and j is the number of parameters.
Figure BDA0002790203880000153
Figure BDA0002790203880000154
Firstly, the partial derivative of the objective function is calculated, then the parameter is updated, if the number of training samples used for each parameter updating, namely batch _ size, is specified, then each parameter updating is carried out, and s samples are randomly selected for calculation. The updating method can reduce the number of iterations and practice, and can specifically calculate based on the following formula:
Figure BDA0002790203880000161
Figure BDA0002790203880000162
in the deep reinforcement learning, the learning of DQN weight is mainly concerned, the loss is calculated and the weight is updated through back-prop, which is similar to the learning of a deep neural network. The loss of the DQN algorithm can be calculated according to the following formula:
TDloss＝Rt+γQ(st+1，at+1)-Q(st，at)
where γ represents the discount rate.
Because the deep neural network is slow in convergence, a great number of samples are needed, and if the network is trained only according to environment interaction, the efficiency is very low. Therefore, DQN introduces a pool of experience to perform the playback experience, i.e. the previous experience of interacting with the environment is stored and reused in training. The ReplayBuffer mainly implements two functions: the push module stores the experience, and the sample takes the experience out for training. In the environment configuration, the DQN algorithm can be applied to the mobile edge computing system, the process of task migration and resource allocation is reproduced, and the next state and reward are returned from the environment module.
For example, when the WoLF-PHC algorithm is adopted to implement the task migration method provided in the above embodiment, for a single agent, the Q values of all state-action pairs may be set to 0; initializing a random strategy piiAnd other control conditions, wherein n is 0.At decision time TnAccording to a hybrid strategy piiSelecting state si(Tn) Action a ofi(si(Tn)). Performing action ai(si(Tn) And recording the sample data, calculating the cumulative reward
Figure BDA0002790203880000163
Computation updates
Figure BDA0002790203880000164
Figure BDA0002790203880000165
Updating an average policy
Figure BDA0002790203880000166
In state si (T)n) A temporal action selection probability; updating the average strategy piiIn a state si(Tn) The action selection probability of time. If the algorithm termination condition is met, the learning is finished; otherwise, n is n +1, and redetermining at decision time TnAccording to a hybrid strategy piiSelecting state si(Tn) Action a ofi(si(Tn) Until a termination condition is met, the iteration ends.
Illustratively, when the task migration method provided by the above embodiment is implemented by using the A3C algorithm, since the A3C algorithm is a product of a combination strategy and a cost function, the action strategy pi(s) in the A3C algorithm determines the action of the agent, which means that the output is not a simple action but a probability distribution of the action, and pi (a | s) is a probability of selecting the action a, and the sum is 1. There are two functions for the state reporting of each action policy:
cost function V(s): the reward available for the current state s is the sum of the return available for the next s' and the reward r available during the state transition.
V(s)＝Eπ(s)[r+γ·V(s′)]
Action cost function Q (s, a) State value Return:
Q(s，a)＝r+γ·V(s′)
a new function A (s, a), the merit function, is defined, which is used to calculate the value of each strategy π(s), which expresses how well to select action a in state s. If action a is better than the average, then the merit function is dominant, otherwise it is inferior:
A(s，a)＝Q(s，a)-V(s)
in the conventional motion estimation algorithm, one-step sampling approximation estimation is performed, and in the A3C algorithm, sampling is performed further, and N-step sampling is used to accelerate convergence. Thus:
A(S，A，t)＝rt+γ·rt+1+…+γn-1·rt+n-1V(S′)-V(S)
from the above equation, it can be seen that only v(s) is needed to calculate a (s, a), and this v(s) is easily calculated by NN. Thus, the cost function and the action-cost function can be jointly predicted.
To get a better strategy, optimization updates must be made. The measure of the quality of a strategy uses a function J (pi) to represent the reward of a discount obtainable by a strategy, from s0The function is difficult to directly estimate based on the average of all the rewards obtained by departure, so the gradient of the function is introduced for estimation.
Figure BDA0002790203880000171
Wherein the A3C algorithm creates multiple parallel environments on which multiple agents simultaneously update parameters in the host structure. The agents in parallel do not interfere with each other, and the updating of parameters of the central structure is discontinuously interfered, so that the relevance of updating is reduced, and the convergence is improved. The A3C algorithm employs an asynchronous update strategy that can operate over a fixed empirical time step. It will use these segments to calculate estimates of the reward r and the dominance function a (s, a). Each agent follows the following workflow: acquiring global network parameters; interacting with the environment by following a local strategy that minimizes the number of steps (t _ max: step to final state); calculating value loss and strategy loss; obtaining a gradient from the loss; updating the global network with the gradient; and repeating the working process.
According to the task migration method provided by the embodiment of the application, task information to be migrated, which is generated by a plurality of user equipment terminals, available bandwidth between each user equipment terminal and a base station and available resources of an edge computing server are obtained; the task information to be migrated comprises: computing resource requirements and bandwidth requirements of the tasks to be migrated; generating a plurality of task migration strategies according to the information of the tasks to be migrated, the available bandwidth between each user equipment terminal and the base station and the available resources of the edge computing server within a preset constraint condition; and determining a target task migration strategy according to the total system energy consumption corresponding to each task migration strategy. According to the task migration method provided by the scheme, the total system energy consumption of various achievable task migration strategies is estimated, so that the task migration strategy with the minimum total system energy consumption is screened out, the risk that the energy consumption of the mobile edge computing system exceeds the standard is avoided, the reliability of the mobile edge computing system is improved, and a foundation is laid for improving the task processing efficiency of the mobile edge computing system.
The embodiment of the application provides a task migration device, which is used for executing the task migration method provided by the embodiment.
Fig. 3 is a schematic structural diagram of a task migration apparatus according to an embodiment of the present application. The task migration apparatus 30 includes an acquisition module 301, a generation module 302, and a determination module 303.
The acquiring module 301 is configured to acquire task information to be migrated of tasks to be migrated, which is generated by multiple user equipment terminals, available bandwidths between each user equipment terminal and a base station, and available resources of an edge computing server; the task information to be migrated comprises: computing resource requirements and bandwidth requirements of the tasks to be migrated; a generating module 302, configured to generate a plurality of task migration policies according to the information of the tasks to be migrated, the available bandwidth between each user equipment terminal and the base station, and the available resources of the edge computing server within preset constraint conditions; the determining module 303 is configured to determine a target task migration policy according to the total system energy consumption corresponding to each task migration policy.
Specifically, in an embodiment, the generating module 302 is specifically configured to:
generating a plurality of task migration decisions corresponding to the tasks to be migrated according to the bandwidth requirements of the tasks to be migrated and the available bandwidth between each user equipment terminal and the base station; the task migration decision comprises an execution destination of each task to be migrated, and the execution destination comprises a user equipment terminal and an edge computing server;
when the execution destination of the task to be migrated is the edge computing server, determining that the task to be migrated is a first task;
acquiring a service quality requirement of a first task, and determining the priority of the first task according to the service quality requirement;
determining the allocation amount of the available resources of the edge computing server according to the priority of each first task;
and generating a plurality of task migration strategies according to a plurality of task migration decisions corresponding to the tasks to be migrated and the allocation condition of the available resources of the edge computing server.
Specifically, in an embodiment, the determining module 303 is specifically configured to:
determining the energy consumption of the edge computing server according to the transmission power and the transmission delay of the user equipment terminal corresponding to each first task and the preset attribute information of the edge computing server;
determining the local energy consumption of each user equipment terminal corresponding to a second task according to the computing resource demand of the second task except for the first task and the execution cycle frequency of the user equipment terminal corresponding to the second task;
determining the total system energy consumption corresponding to each task migration strategy according to the energy consumption of the edge computing server and the local energy consumption;
and determining the task migration strategy with the minimum system total energy consumption as a target task migration strategy.
Specifically, in an embodiment, the determining module 303 is specifically configured to:
determining the total energy consumption of the system corresponding to any task migration strategy according to the following formula:
Figure BDA0002790203880000191
wherein E isallThe total system energy consumption corresponding to the task migration strategy is represented, T represents the total system consumption time, and thetaiIndicating an execution destination of the ith task to be migrated, wherein when the execution destination of the ith task to be migrated is the edge calculation server, thetaiWhen the execution destination of the ith task to be migrated is the corresponding user equipment terminal, θ 1i＝0，
Figure BDA0002790203880000192
Representing the energy consumption of the edge computing server corresponding to each first task,
Figure BDA0002790203880000193
and local energy consumption corresponding to each second task is represented.
Specifically, in an embodiment, the determining module 303 is further specifically configured to:
calculating the reward value of each task migration strategy according to the total energy consumption of the system based on the following formula:
Figure BDA0002790203880000194
where r represents the prize value telocalIndicating the total energy consumption of the system when the execution destinations of the tasks to be migrated are all corresponding user equipment terminals, EallRepresenting the total energy consumption of the system corresponding to the task migration strategy;
and determining the task migration strategy with the highest reward value as a target task migration strategy.
Specifically, in an embodiment, the generating module 302 is further configured to:
constructing a waiting queue according to a preset caching requirement;
and storing each first task to a waiting queue according to the generation sequence of each first task.
Specifically, in one embodiment, the preset constraint includes:
the total transmission delay of each task migration strategy is smaller than a preset total transmission delay threshold, the execution cycle frequency of each user equipment terminal is smaller than a preset execution cycle frequency threshold, and/or the execution cycle number of each user equipment terminal is not larger than the execution cycle number of the edge computing server.
With regard to the task migration apparatus in the present embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment related to the method, and will not be elaborated here.
The task migration apparatus provided in the embodiment of the present application is configured to execute the task migration method provided in the above embodiment, and an implementation manner of the task migration apparatus is the same as a principle, and is not described again.
The embodiment of the application provides electronic equipment, which is used for executing the task migration method provided by the embodiment.
Fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application. The electronic device 40 includes: at least one processor 41 and memory 42;
the memory stores computer-executable instructions; the at least one processor executes the memory-stored computer-executable instructions to cause the at least one processor to perform the task migration method provided by the above embodiments.
The electronic device provided in the embodiment of the present application is configured to execute the task migration method provided in the above embodiment, and an implementation manner and a principle of the method are the same, and are not described again.
The embodiment of the present application provides a computer-readable storage medium, in which computer execution instructions are stored, and when a processor executes the computer execution instructions, the task migration method provided in any one of the above embodiments is implemented.
The storage medium including the computer-executable instructions of the embodiments of the present application may be used to store the computer-executable instructions of the task migration method provided in the foregoing embodiments, and an implementation manner of the storage medium is the same as a principle, and is not described again.
It is obvious to those skilled in the art that, for convenience and simplicity of description, the foregoing division of the functional modules is merely used as an example, and in practical applications, the above function distribution may be performed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to perform all or part of the above described functions. For the specific working process of the device described above, reference may be made to the corresponding process in the foregoing method embodiment, which is not described herein again.
Finally, it should be noted that: the above embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present application.

Claims (10)

1. A task migration method is applied to a mobile edge computing system, the mobile edge computing system comprises a plurality of user equipment terminals, an edge computing server and a base station, and the task migration method is characterized by comprising the following steps:
acquiring task information to be migrated of tasks to be migrated, which is generated by a plurality of user equipment terminals, available bandwidth between each user equipment terminal and a base station and available resources of the edge computing server; wherein the task information to be migrated includes: computing resource requirements and bandwidth requirements of the tasks to be migrated;
generating a plurality of task migration strategies according to the information of the tasks to be migrated, the available bandwidth between each user equipment terminal and the base station and the available resources of the edge computing server within a preset constraint condition;
and determining a target task migration strategy according to the total system energy consumption corresponding to each task migration strategy.
2. The task migration method according to claim 1, wherein the generating a plurality of task migration policies according to the to-be-migrated task information, the available bandwidth between each user equipment terminal and the base station, and the available resources of the edge computing server comprises:
generating a plurality of task migration decisions corresponding to the tasks to be migrated according to the bandwidth requirements of the tasks to be migrated and the available bandwidth between each user equipment terminal and the base station; the task migration decision comprises an execution destination of each task to be migrated, and the execution destination comprises a user equipment terminal and an edge computing server;
when the execution destination of the task to be migrated is the edge computing server, determining that the task to be migrated is a first task;
acquiring a service quality requirement of the first task, and determining the priority of the first task according to the service quality requirement;
determining the allocation amount of the available resources of the edge computing server according to the priority of each first task;
and generating a plurality of task migration strategies according to a plurality of task migration decisions corresponding to the tasks to be migrated and the allocation condition of the available resources of the edge computing server.
3. The task migration method according to claim 2, wherein the determining a target task migration policy according to the total system energy consumption corresponding to each task migration policy comprises:
determining the energy consumption of the edge computing server according to the transmission power and the transmission delay of the user equipment terminal corresponding to each first task and the preset attribute information of the edge computing server;
determining local energy consumption of each user equipment terminal corresponding to a second task according to the computing resource demand of the second task except the first task and the execution cycle frequency of the user equipment terminal corresponding to the second task;
determining total system energy consumption corresponding to each task migration strategy according to the edge computing server energy consumption and the local energy consumption;
and determining the task migration strategy with the minimum system total energy consumption as the target task migration strategy.
4. The task migration method according to claim 3, wherein the determining the total system energy consumption corresponding to each task migration policy according to the edge computing server energy consumption and the local energy consumption comprises:
determining the total energy consumption of the system corresponding to any task migration strategy according to the following formula:
Figure FDA0002790203870000021
wherein E isallRepresenting the total system energy consumption corresponding to the task migration strategy, T representing the total system consumption time, thetaiIndicating an execution destination of an ith task to be migrated, wherein when the execution destination of the ith task to be migrated is an edge computing server, thetaiWhen the execution destination of the ith task to be migrated is the corresponding user equipment terminal, θi＝0，
Figure FDA0002790203870000022
Representing the energy consumption of the edge computing server corresponding to each first task,
Figure FDA0002790203870000023
and local energy consumption corresponding to each second task is represented.
5. The task migration method according to claim 1, wherein the determining a target task migration policy according to the total system energy consumption corresponding to each task migration policy comprises:
calculating the reward value of each task migration strategy according to the total energy consumption of the system based on the following formula:
Figure FDA0002790203870000024
where r represents the prize value telocalIndicating the total energy consumption of the system when the execution destinations of the tasks to be migrated are all corresponding user equipment terminals, EallRepresenting the total energy consumption of the system corresponding to the task migration strategy;
and determining the task migration strategy with the highest reward value as the target task migration strategy.
6. The task migration method according to claim 2, further comprising:
constructing a waiting queue according to a preset caching requirement;
and storing each first task to the waiting queue according to the generation sequence of each first task.
7. The task migration method according to claim 1, wherein the preset constraint condition comprises:
the total transmission delay of each task migration strategy is smaller than a preset total transmission delay threshold, the execution cycle frequency of each user equipment terminal is smaller than a preset execution cycle frequency threshold, and/or the execution cycle number of each user equipment terminal is not larger than the execution cycle number of the edge computing server.
8. A task migration apparatus applied to a mobile edge computing system, the mobile edge computing system comprising a plurality of user equipment terminals, an edge computing server and a base station, the task migration apparatus comprising:
the system comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring task information to be migrated of tasks generated by a plurality of user equipment terminals, available bandwidth between each user equipment terminal and a base station and available resources of an edge computing server; wherein the task information to be migrated includes: computing resource requirements and bandwidth requirements of the tasks to be migrated;
a generating module, configured to generate a plurality of task migration strategies according to the information of the tasks to be migrated, the available bandwidth between each user equipment terminal and the base station, and the available resources of the edge computing server within a preset constraint condition;
and the determining module is used for determining the target task migration strategy according to the total system energy consumption corresponding to each task migration strategy.
9. An electronic device, comprising: at least one processor and memory;
the memory stores computer-executable instructions;
the at least one processor executing the computer-executable instructions stored by the memory causes the at least one processor to perform the method of any of claims 1-7.
10. A computer-readable storage medium having computer-executable instructions stored thereon which, when executed by a processor, implement the method of any one of claims 1 to 7.
CN202011312378.7A 2020-11-20 2020-11-20 Task migration method and device, electronic equipment and storage medium Active CN112416554B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011312378.7A CN112416554B (en) 2020-11-20 2020-11-20 Task migration method and device, electronic equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011312378.7A CN112416554B (en) 2020-11-20 2020-11-20 Task migration method and device, electronic equipment and storage medium

Publications (2)

Publication Number Publication Date
CN112416554A true CN112416554A (en) 2021-02-26
CN112416554B CN112416554B (en) 2022-12-02

Family

ID=74778386

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011312378.7A Active CN112416554B (en) 2020-11-20 2020-11-20 Task migration method and device, electronic equipment and storage medium

Country Status (1)

Country Link
CN (1) CN112416554B (en)

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112887435A (en) * 2021-04-13 2021-06-01 中南大学 Method for improving task unloading cooperation rate in edge calculation
CN113014659A (en) * 2021-03-11 2021-06-22 北京邮电大学 Microservice migration method and device, storage medium and electronic equipment
CN113141394A (en) * 2021-03-25 2021-07-20 北京邮电大学 Resource allocation method and device, electronic equipment and storage medium
CN113452751A (en) * 2021-05-20 2021-09-28 国网江苏省电力有限公司信息通信分公司 Cloud edge cooperation-based power internet of things task secure migration system and method
CN113515378A (en) * 2021-06-28 2021-10-19 国网河北省电力有限公司雄安新区供电公司 Method and device for migration and calculation resource allocation of 5G edge calculation task
CN113597013A (en) * 2021-08-05 2021-11-02 哈尔滨工业大学 Cooperative task scheduling method in mobile edge computing under user mobile scene
CN113645637A (en) * 2021-07-12 2021-11-12 中山大学 Method and device for unloading tasks of ultra-dense network, computer equipment and storage medium
CN113689001A (en) * 2021-08-30 2021-11-23 浙江大学 Virtual self-playing method and device based on anti-factual regret minimization
CN114003121A (en) * 2021-09-30 2022-02-01 中国科学院计算技术研究所 Method and device for optimizing energy efficiency of data center server, electronic equipment and storage medium
CN114281426A (en) * 2021-12-21 2022-04-05 中国联合网络通信集团有限公司 Task unloading method and device, electronic equipment and readable storage medium
CN116112976A (en) * 2022-12-20 2023-05-12 暨南大学 Equipment calculation migration method, device, equipment and storage medium
WO2024012295A1 (en) * 2022-07-14 2024-01-18 抖音视界有限公司 Video transmission method, apparatus, and system, device, and medium
CN117873689A (en) * 2024-03-11 2024-04-12 浪潮计算机科技有限公司 Task allocation method, device, equipment and computer readable storage medium
CN117873689B (en) * 2024-03-11 2024-05-31 浪潮计算机科技有限公司 Task allocation method, device, equipment and computer readable storage medium

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019200716A1 (en) * 2018-04-20 2019-10-24 上海无线通信研究中心 Fog computing-oriented node computing task scheduling method and device thereof
CN111031102A (en) * 2019-11-25 2020-04-17 哈尔滨工业大学 Multi-user, multi-task mobile edge computing system cacheable task migration method

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019200716A1 (en) * 2018-04-20 2019-10-24 上海无线通信研究中心 Fog computing-oriented node computing task scheduling method and device thereof
CN111031102A (en) * 2019-11-25 2020-04-17 哈尔滨工业大学 Multi-user, multi-task mobile edge computing system cacheable task migration method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
朱霞等: "基于边缘计算的单任务迁移策略研究", 《金陵科技学院学报》 *

Cited By (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113014659A (en) * 2021-03-11 2021-06-22 北京邮电大学 Microservice migration method and device, storage medium and electronic equipment
CN113141394A (en) * 2021-03-25 2021-07-20 北京邮电大学 Resource allocation method and device, electronic equipment and storage medium
CN112887435A (en) * 2021-04-13 2021-06-01 中南大学 Method for improving task unloading cooperation rate in edge calculation
CN113452751A (en) * 2021-05-20 2021-09-28 国网江苏省电力有限公司信息通信分公司 Cloud edge cooperation-based power internet of things task secure migration system and method
CN113515378A (en) * 2021-06-28 2021-10-19 国网河北省电力有限公司雄安新区供电公司 Method and device for migration and calculation resource allocation of 5G edge calculation task
CN113645637A (en) * 2021-07-12 2021-11-12 中山大学 Method and device for unloading tasks of ultra-dense network, computer equipment and storage medium
CN113597013A (en) * 2021-08-05 2021-11-02 哈尔滨工业大学 Cooperative task scheduling method in mobile edge computing under user mobile scene
CN113597013B (en) * 2021-08-05 2024-03-22 哈尔滨工业大学 Collaborative task scheduling method under user mobile scene in mobile edge calculation
CN113689001B (en) * 2021-08-30 2023-12-05 浙江大学 Virtual self-playing method and device based on counter-facts regretation minimization
CN113689001A (en) * 2021-08-30 2021-11-23 浙江大学 Virtual self-playing method and device based on anti-factual regret minimization
CN114003121A (en) * 2021-09-30 2022-02-01 中国科学院计算技术研究所 Method and device for optimizing energy efficiency of data center server, electronic equipment and storage medium
CN114003121B (en) * 2021-09-30 2023-10-31 中国科学院计算技术研究所 Data center server energy efficiency optimization method and device, electronic equipment and storage medium
CN114281426B (en) * 2021-12-21 2023-05-16 中国联合网络通信集团有限公司 Task unloading method and device, electronic equipment and readable storage medium
CN114281426A (en) * 2021-12-21 2022-04-05 中国联合网络通信集团有限公司 Task unloading method and device, electronic equipment and readable storage medium
WO2024012295A1 (en) * 2022-07-14 2024-01-18 抖音视界有限公司 Video transmission method, apparatus, and system, device, and medium
CN116112976A (en) * 2022-12-20 2023-05-12 暨南大学 Equipment calculation migration method, device, equipment and storage medium
CN116112976B (en) * 2022-12-20 2024-05-03 暨南大学 Equipment calculation migration method, device, equipment and storage medium
CN117873689A (en) * 2024-03-11 2024-04-12 浪潮计算机科技有限公司 Task allocation method, device, equipment and computer readable storage medium
CN117873689B (en) * 2024-03-11 2024-05-31 浪潮计算机科技有限公司 Task allocation method, device, equipment and computer readable storage medium

Also Published As

Publication number Publication date
CN112416554B (en) 2022-12-02

Similar Documents

Publication Publication Date Title
CN112416554B (en) Task migration method and device, electronic equipment and storage medium
CN113242568B (en) Task unloading and resource allocation method in uncertain network environment
CN111031102B (en) Multi-user, multi-task mobile edge computing system cacheable task migration method
CN107911478B (en) Multi-user calculation unloading method and device based on chemical reaction optimization algorithm
CN113612843A (en) MEC task unloading and resource allocation method based on deep reinforcement learning
CN113950066A (en) Single server part calculation unloading method, system and equipment under mobile edge environment
CN112422644B (en) Method and system for unloading computing tasks, electronic device and storage medium
CN111835827A (en) Internet of things edge computing task unloading method and system
CN112291335B (en) Optimized task scheduling method in mobile edge calculation
CN110069341B (en) Method for scheduling tasks with dependency relationship configured according to needs by combining functions in edge computing
Mostafavi et al. A stochastic approximation approach for foresighted task scheduling in cloud computing
CN113645637B (en) Method and device for unloading tasks of ultra-dense network, computer equipment and storage medium
CN113778691B (en) Task migration decision method, device and system
CN114205353B (en) Calculation unloading method based on hybrid action space reinforcement learning algorithm
CN116541106A (en) Computing task unloading method, computing device and storage medium
Hu et al. Dynamic task offloading in MEC-enabled IoT networks: A hybrid DDPG-D3QN approach
CN113489787B (en) Method and device for collaborative migration of mobile edge computing service and data
CN115408072A (en) Rapid adaptation model construction method based on deep reinforcement learning and related device
CN112905315A (en) Task processing method, device and equipment in Mobile Edge Computing (MEC) network
CN116954866A (en) Edge cloud task scheduling method and system based on deep reinforcement learning
CN116915869A (en) Cloud edge cooperation-based time delay sensitive intelligent service quick response method
CN116471303A (en) Industrial Internet of things edge computing unloading and task migration method and industrial Internet of things system
CN116996941A (en) Calculation force unloading method, device and system based on cooperation of cloud edge ends of distribution network
CN114785692B (en) Communication network flow balancing method and device for aggregation regulation of virtual power plants
CN114968402A (en) Edge calculation task processing method and device and electronic equipment

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant