CN112860429A - Cost-efficiency optimization system and method for task unloading in mobile edge computing system - Google Patents

Cost-efficiency optimization system and method for task unloading in mobile edge computing system Download PDF

Info

Publication number
CN112860429A
CN112860429A CN202011595152.2A CN202011595152A CN112860429A CN 112860429 A CN112860429 A CN 112860429A CN 202011595152 A CN202011595152 A CN 202011595152A CN 112860429 A CN112860429 A CN 112860429A
Authority
CN
China
Prior art keywords
task
edge server
cost
unloading
base station
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
CN202011595152.2A
Other languages
Chinese (zh)
Other versions
CN112860429B (en
Inventor
李秀华
吴凡凡
李辉
徐峥辉
明钊
范琪琳
文俊浩
熊庆宇
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.)
Chongqing University
Original Assignee
Chongqing University
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 Chongqing University filed Critical Chongqing University
Priority to CN202011595152.2A priority Critical patent/CN112860429B/en
Publication of CN112860429A publication Critical patent/CN112860429A/en
Application granted granted Critical
Publication of CN112860429B publication Critical patent/CN112860429B/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
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • 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
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5061—Partitioning or combining of resources
    • G06F9/5072—Grid computing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/04—Constraint-based CAD
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2209/00—Indexing scheme relating to G06F9/00
    • G06F2209/50—Indexing scheme relating to G06F9/50
    • G06F2209/502—Proximity
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2209/00—Indexing scheme relating to G06F9/00
    • G06F2209/50—Indexing scheme relating to G06F9/50
    • G06F2209/509—Offload
    • 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
    • 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
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Abstract

The invention discloses a cost-effective optimization system and method for task offloading in a mobile edge computing system. The method comprises the following steps: 1) building a mobile edge computing system; 2) calculating energy costs and time costs; 3) setting unloading decision variables of tasks and establishing a user service experience gain model; 4) determining task R using a user service experience gain modeliAn offload decision; 5) calculating cost effectiveness; 6) a task offload profile based on task offload priority is determined. The system comprises a mobile edge computing system, an energy cost and time cost computing module, a single task unloading decision generating module, a task unloading cost efficiency computing module, a task unloading scheme generating module anda database. The invention can realize the maximization of energy-time cost efficiency of the mobile edge computing system.

Description

Cost-efficiency optimization system and method for task unloading in mobile edge computing system
Technical Field
The invention relates to the technical field of edge computing, in particular to a cost-efficiency optimization system and method for task unloading in a mobile edge computing system.
Background
With the wave of digital transformation and the arrival of the world of everything interconnection, information technologies such as 5G, big data, artificial intelligence and the like are rapidly developed. In the face of explosive growth of network data, a communication network bears huge pressure; meanwhile, new application service scenes such as unmanned driving, virtual reality/augmented reality, remote medical treatment and the like have higher requirements on service delay and power consumption. Faced with these challenges, conventional centralized network architectures (e.g., cloud computing) have been unable to meet the needs of users due to the heavy backhaul link load and long latency. Mobile edge computing, as a new network architecture, can open the capabilities of the core network into the edge network. Compared with cloud computing, the mobile edge computing realizes sinking of resources and services to edge positions, so that interaction time delay is reduced, network burden is reduced, service types are enriched, service processing is optimized, and service quality and user experience are improved. In an MEC scene, the mobile device unloads a computing task with time delay sensitivity and high computing resource demand to an edge server for execution, so as to achieve the purposes of shortening service time delay and improving service experience.
In a mobile edge computing system, the system incurs energy and time costs while the user gains in the service experience by offloading tasks into the edge network. Currently, related research in task offloading only considers the service experience gain of the user, such as reduction of energy consumption of the user equipment or reduction of service delay, and ignores the system cost, including energy cost and time cost, generated by task offloading work. In particular, energy costs arise from data transfer of the communication link and task execution by the edge server/user device, and time costs arise from the occupation of the communication device as well as the computing device by task offload work. In order to solve the limitation of related research, the comprehensive optimization between the user service experience gain and the system cost has considerable innovation and important working significance.
In addition, the task offloading problem in hot spot area scenarios is poorly studied. In a hot spot area scene, the following three challenges exist to be solved: 1) a multi-user multi-base station scenario. In a hot spot area scenario, some areas are in signal coverage areas of multiple base stations, and a user task offloading request in the area has multiple offloading strategies available for consideration, which greatly increases the complexity of the task offloading problem. Therefore, it is significant to make a reasonable task unloading scheme from the overall system perspective. 2) High concurrency task requests. In a hot spot area scene, task unloading requests in a mobile edge computing system have high density and high concurrency, so the system needs an efficient task unloading decision scheme. 3) Task heterogeneity. The task heterogeneity is that under different service scenes, users have different sensitivities to energy consumption and time delay, for example, under service scenes such as cloud games and virtual reality, compared with energy consumption of user equipment, users are more sensitive to service time delay, and the service time delay has a greater influence on user service experience; for services such as model training and big data analysis, the key factors influencing user service experience are insufficient computing capacity of user equipment and excessive energy consumption for task execution.
Disclosure of Invention
The invention aims to provide a cost-effective optimization method for task unloading in a mobile edge computing system, which comprises the following steps:
1) and building a mobile edge computing system.
The mobile edge computing system includes a number of users, a number of base stations, and a number of edge servers.
The edge server is deployed at a location that is distant from base station i. Base stations and edge servers with a linear distance equal to l are denoted as a base station-edge server group.
All base station-edge server groups as a set
Figure BDA0002870131080000021
Wherein, the operation capability of the jth BS-EDM group is recorded as
Figure BDA0002870131080000022
Figure BDA0002870131080000023
And
Figure BDA0002870131080000024
respectively representing the communication capability of the base station in the jth base station-edge server group and the computing capability of the edge server.
Clustering users in hot spot regions in a mobile edge computing system
Figure BDA0002870131080000025
The hot spot area is a cross coverage area of the communication signals of the base station. Wherein the unloading task of the ith user is recorded as
Figure BDA0002870131080000026
fiAnd ciRespectively representing the amount of data that needs to be transferred in the communication link for task offloading and the number of CPU run cycles needed for task execution. Alpha is alphaiAnd betaiRespectively represents the sensitivity of the service to the service energy consumption and the service delay of the task, and alphai+βi＝1。
2) Computation task RiEnergy costs required for local execution on a user equipment
Figure BDA0002870131080000027
And time cost Ti L。
Task RiEnergy costs required for local execution on a user equipment
Figure BDA0002870131080000028
And time cost Ti LRespectively as follows:
Figure BDA0002870131080000029
Figure BDA00028701310800000210
in the formula (I), the compound is shown in the specification,
Figure BDA00028701310800000211
and
Figure BDA00028701310800000212
respectively representing the energy consumed and the computing power consumed by the ith user equipment in one CPU operation cycle. c. CiIndicating the number of CPU run cycles required for task execution.
3) Computation task RiEnergy cost required to offload to jth BS-edge server group for execution
Figure BDA00028701310800000213
And time cost
Figure BDA00028701310800000214
Task RiEnergy cost required to offload to jth BS-edge server group for execution
Figure BDA00028701310800000215
And time cost
Figure BDA00028701310800000216
Respectively as follows:
Figure BDA0002870131080000031
Figure BDA0002870131080000032
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000033
respectively representing the energy cost of data transmission and the energy cost of task execution during the task unloading process.
Figure BDA0002870131080000034
Running the energy consumption of one CPU cycle for the jth edge server. f. ofiAnd ciRespectively representing the amount of data that needs to be transferred in the communication link for task offloading and the number of CPU run cycles needed for task execution.
Wherein the communication capability
Figure BDA0002870131080000035
Computing power
Figure BDA0002870131080000036
Task RiData transmission speed v in a communication linki,jRespectively as follows:
Figure BDA0002870131080000037
Figure BDA0002870131080000038
Figure BDA0002870131080000039
in the formula, piAnd N0Signal transmission power and noise power, g, of the ith user equipment, respectivelyi,jThe channel gain from the ith user equipment to the jth base station.
4) Setting task RiUnloading blockPolicy variable oiAnd establishing a user service experience gain model.
Figure BDA00028701310800000310
When task RiIs unloaded to the decision variable oiWhen j, task RiAnd unloading to the jth base station-edge server group. When task RiIs unloaded to the decision variable oiWhen 0, task RiUnload failure, task RiLocally on the ith user equipment.
The user service experience gain model is as follows:
Figure BDA00028701310800000311
in the formula (I), the compound is shown in the specification,
Figure BDA00028701310800000312
representing a task RiThe power consumption generated by the ith user equipment when the local execution is carried out.
Figure BDA00028701310800000313
Representing a task RiService delay generated by the ith user equipment when the local execution is carried out.
Figure BDA00028701310800000314
Representing a task RiIs unloaded to oiThe power consumption generated by the ith user equipment when the base station-edge server group is executed.
Figure BDA00028701310800000315
Representing a task RiIs unloaded to oiService delay generated by the ith user equipment when the base station-edge server group is executed.
Figure BDA00028701310800000316
A gain value is experienced for the user service.
5) Computation task RiAt the time of unloadingThe carrier decision is oiTime and cost efficiency
Figure BDA0002870131080000041
Task RiAt an offload decision of oiTime and cost efficiency
Figure BDA0002870131080000042
As follows:
Figure BDA0002870131080000043
where η is the coefficient that balances the energy cost term and the time cost term.
6) And determining a task unloading scheme based on the task unloading priority, establishing communication connection for the task unloading request according to the task unloading scheme, and allocating communication resources and computing resources.
Determining a task offloading scheme based on task offloading priorities, the steps comprising:
6.1) initializing variables, including task set variables
Figure BDA0002870131080000044
Base station-edge server group working capacity variable
Figure BDA0002870131080000045
Unallocated task set variables
Figure BDA0002870131080000046
Offloading decision variables
Figure BDA0002870131080000047
Variables for recording workload conditions for base station-edge server groups
Figure BDA0002870131080000048
Wherein
Figure BDA0002870131080000049
Representing the computational workload of the jth base station-edge server group.
Figure BDA00028701310800000410
Representing the communication workload of the jth base station-edge server group.
6.2) taking the base station-edge server group with the minimum workload as a working group, wherein the sequence number χ of the working group is as follows:
Figure BDA00028701310800000411
in the formula (I), the compound is shown in the specification,
Figure BDA00028701310800000412
the sequence number of the base station-edge server group with the minimum workload in the base station-edge server group is solved.
6.3) set of unallocated tasks
Figure BDA00028701310800000413
Classifying to obtain a locally executed task set thetaLAnd the set of tasks Θ remotely executed off-load to the edge serverR. Wherein, the task set thetaRUser service experience gain k corresponding to remote execution of tasks in (1)i,χGreater than 1.
Task set ΘLAnd task set ΘRAs follows:
Figure BDA00028701310800000414
Figure BDA00028701310800000415
determining a current task R using a user service experience gain modeliIf the current task R is determined to be uninstallediExecuting locally, then the current task RiWrite task set ΘLIn the middle and on the contrary, the current task R is processediWrite task set ΘRIn (1).
Determining task R using a user service experience gain modeliThe method of offloading decision(s) of (1) is: computation task RiOffloading user service experience gain values performed on jth edge server
Figure BDA0002870131080000051
If it is
Figure BDA0002870131080000052
Then task RiExecuting locally, offloading decision variable o i0. Otherwise, task RiOff-load to edge server execution, off-load decision variable oi＝j。
6.4) set of tasks ΘRWherein task R establishes an offloading priorityiOf offload priority PiAs follows:
Figure BDA0002870131080000053
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000054
representing a task RiThe ratio of the power consumption generated by the ith ue in both cases of local execution and execution on the χ -th edge server.
Figure BDA0002870131080000055
Representing a task RiThe ratio of service delay incurred by the ith UE in both cases of executing locally and on the x-th edge server.
Figure BDA0002870131080000056
And
Figure BDA0002870131080000057
respectively represent tasks RiExecute locally and execute on x-th edgeThe difference between the energy cost and the time cost generated by the system in the two cases is executed on the edge server.
Figure BDA0002870131080000058
Respectively represent tasks RiThe energy and time costs incurred by executing the lower system on the x-th edge server.
6.5) determining the task with the maximum unloading priority at the current moment
Figure BDA0002870131080000059
And will unload the task with the largest priority
Figure BDA00028701310800000510
Off-loading to workgroup χ for execution, i.e. task
Figure BDA00028701310800000511
Offload decision
Figure BDA00028701310800000512
Offloading the task with the greatest priority
Figure BDA00028701310800000513
As follows:
Figure BDA00028701310800000514
in the formula (I), the compound is shown in the specification,
Figure BDA00028701310800000515
and the task sequence number with the largest unloading priority in the solving task is shown.
6.6) the first
Figure BDA00028701310800000516
After each task is unloaded to the x base station-edge server group, the system state variable is updated, namely:
Figure BDA00028701310800000517
Figure BDA00028701310800000518
Figure BDA00028701310800000519
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000061
respectively representing the updated unallocated task set, the computation workload and the communication workload of the chi-th base station-edge server group;
Figure BDA0002870131080000062
representing the computation workload and the communication workload of the chi-th base station-edge server group before updating;
Figure BDA0002870131080000063
indicating the unloading of
Figure BDA0002870131080000064
The amount of data that an individual task needs to transmit in a communication link;
6.7) return to step 6.2) until
Figure BDA0002870131080000065
The idle system resources for the empty set or all base station-edge server groups are not sufficient to handle the task offload request.
6.8) repeating the step 6.2) to the step 6.7) to obtain a task unloading decision set
Figure BDA0002870131080000066
The task unloading decision set O is the task unloading scheme.
The system based on the cost-efficiency optimization method for task unloading in the mobile edge computing system comprises the mobile edge computing system, an energy cost and time cost computing module, a single-task unloading decision generating module, a task unloading cost-efficiency computing module, a task unloading scheme generating module and a database.
The energy cost and time cost calculation module calculates a task RiEnergy costs required for local execution on a user equipment
Figure BDA0002870131080000067
Task RiTime cost T required for local execution on user equipmenti LTask RiEnergy cost required to offload to jth BS-edge server group for execution
Figure BDA0002870131080000068
And task RiTime cost required for offloading to jth BS-edge server group for execution
Figure BDA0002870131080000069
And sending the data to a single task unloading decision generation module and a task unloading cost efficiency calculation module.
The single task offload decision generation module stores a user service experience gain model.
The single task offload decision generation module determines task R using a user service experience gain modeliAnd sending the unloading decision to a task unloading cost efficiency calculation module.
The task offloading cost-efficiency calculation module calculates a task RiAt an offload decision of oiTime and cost efficiency
Figure BDA00028701310800000610
And the task unloading scheme generating module determines a task unloading scheme based on the task unloading priority, establishes communication connection for the task unloading request according to the task unloading scheme, and allocates communication resources and computing resources.
The database stores data of an energy cost and time cost calculation module, a single task unloading decision generation module, a task unloading cost efficiency calculation module and a task unloading scheme generation module.
The technical effect of the invention is undoubted, and the invention comprehensively considers the energy cost, the time cost and the user service experience gain of task unloading in the mobile edge computing system, the communication resource constraint of a base station in the system and the computing resource constraint of an edge server, optimizes the decision scheme of task unloading and realizes the maximization of the energy-time cost efficiency of the mobile edge computing system.
Drawings
FIG. 1 is a model diagram of a moving edge computing system;
FIG. 2 is a flow chart of an algorithm for a task offload scheme based on task offload priority in accordance with the present invention;
FIG. 3 is a graph comparing cost effectiveness results for different methods with different numbers of user devices according to the present invention;
FIG. 4 is a graph comparing cost effectiveness results of different methods of the present invention at different cost factors η.
Detailed Description
The present invention is further illustrated by the following examples, but it should not be construed that the scope of the above-described subject matter is limited to the following examples. Various substitutions and alterations can be made without departing from the technical idea of the invention and the scope of the invention is covered by the present invention according to the common technical knowledge and the conventional means in the field.
Example 1:
a method for cost-effective optimization of task offloading in a mobile edge computing system, comprising the steps of:
1) and building a mobile edge computing system.
The mobile edge computing system includes a number of users, a number of base stations, and a number of edge servers.
The edge server is deployed at a location that is distant from base station i. Base stations and edge servers with a linear distance equal to l are denoted as a base station-edge server group.
All base station-edge server groups as a set
Figure BDA0002870131080000071
Wherein, the operation capability of the jth BS-EDM group is recorded as
Figure BDA0002870131080000072
Figure BDA0002870131080000073
And
Figure BDA0002870131080000074
respectively representing the communication capability of the base station in the jth base station-edge server group and the computing capability of the edge server.
Clustering users in hot spot regions in a mobile edge computing system
Figure BDA0002870131080000075
The hot spot area is a cross coverage area of the communication signals of the base station. Wherein the unloading task of the ith user is recorded as
Figure BDA0002870131080000076
fiAnd ciRespectively representing the amount of data that needs to be transferred in the communication link for task offloading and the number of CPU run cycles needed for task execution. Alpha is alphaiAnd betaiRespectively represents the sensitivity of the service to the service energy consumption and the service delay of the task, and alphai+βi＝1。
2) Computation task RiEnergy costs required for local execution on a user equipment
Figure BDA0002870131080000077
And time cost Ti L。
Task RiEnergy costs required for local execution on a user equipment
Figure BDA0002870131080000078
And time cost Ti LRespectively as follows:
Figure BDA0002870131080000081
Figure BDA0002870131080000082
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000083
and
Figure BDA0002870131080000084
respectively representing the energy consumed and the computing power consumed by the ith user equipment in one CPU operation cycle. c. CiIndicating the number of CPU run cycles required for task execution.
3) Computation task RiEnergy cost required to offload to jth BS-edge server group for execution
Figure BDA0002870131080000085
And time cost
Figure BDA0002870131080000086
Task RiEnergy cost required to offload to jth BS-edge server group for execution
Figure BDA0002870131080000087
And time cost
Figure BDA0002870131080000088
Respectively as follows:
Figure BDA0002870131080000089
Figure BDA00028701310800000810
in the formula (I), the compound is shown in the specification,
Figure BDA00028701310800000811
respectively representing the energy cost of data transmission and the energy cost of task execution during the task unloading process.
Figure BDA00028701310800000812
Running the energy consumption of one CPU cycle for the jth edge server. f. ofiAnd ciRespectively representing the amount of data that needs to be transferred in the communication link for task offloading and the number of CPU run cycles needed for task execution.
Wherein the communication capability
Figure BDA00028701310800000813
Computing power
Figure BDA00028701310800000814
Task RiData transmission speed v in a communication linki,jRespectively as follows:
Figure BDA00028701310800000815
Figure BDA00028701310800000816
Figure BDA00028701310800000817
in the formula, piAnd N0Signal transmission power and noise power, g, of the ith user equipment, respectivelyi,jThe channel gain from the ith user equipment to the jth base station.
4) Setting task RiIs unloaded to the decision variable oiAnd establishing a user service experience gain model.
Figure BDA00028701310800000818
When task RiIs unloaded to the decision variable oiWhen j, task RiAnd unloading to the jth base station-edge server group. When task RiIs unloaded to the decision variable oiWhen 0, task RiUnload failure, task RiLocally on the ith user equipment.
The user service experience gain model is as follows:
Figure BDA0002870131080000091
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000092
representing a task RiThe power consumption generated by the ith user equipment when the local execution is carried out.
Figure BDA0002870131080000093
Representing a task RiService delay generated by the ith user equipment when the local execution is carried out.
Figure BDA0002870131080000094
Representing a task RiIs unloaded to oiThe power consumption generated by the ith user equipment when the base station-edge server group is executed.
Figure BDA0002870131080000095
Representing a task RiIs unloaded to oiService delay generated by the ith user equipment when the base station-edge server group is executed.
Figure BDA0002870131080000096
A gain value is experienced for the user service.
5) Computation task RiAt an offload decision of oiTime and cost efficiency
Figure BDA0002870131080000097
Task RiAt an offload decision of oiTime and cost efficiency
Figure BDA0002870131080000098
As follows:
Figure BDA0002870131080000099
where η is the coefficient that balances the energy cost term and the time cost term.
6) And determining a task unloading scheme based on the task unloading priority, establishing communication connection for the task unloading request according to the task unloading scheme, and allocating communication resources and computing resources.
Determining a task offloading scheme based on task offloading priorities, the steps comprising:
6.1) initializing variables, including task set variables
Figure BDA00028701310800000910
Base station-edge server group working capacity variable
Figure BDA00028701310800000911
Unallocated task set variables
Figure BDA00028701310800000912
Offloading decision variables
Figure BDA00028701310800000913
Variables for recording workload conditions for base station-edge server groups
Figure BDA00028701310800000914
Wherein
Figure BDA00028701310800000915
Representing the computational workload of the jth base station-edge server group.
Figure BDA00028701310800000916
Representing the communication workload of the jth base station-edge server group.
6.2) taking the base station-edge server group with the minimum workload as a working group, wherein the sequence number χ of the working group is as follows:
Figure BDA00028701310800000917
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000101
the sequence number of the base station-edge server group with the minimum workload in the base station-edge server group is solved.
6.3) set of unallocated tasks
Figure BDA0002870131080000102
Classifying to obtain a locally executed task set thetaLAnd the set of tasks Θ remotely executed off-load to the edge serverR. Wherein, the task set thetaRUser service experience gain k corresponding to remote execution of tasks in (1)i,χGreater than 1.
Task set ΘLAnd task set ΘRAs follows:
Figure BDA0002870131080000103
Figure BDA0002870131080000104
determining a current task R using a user service experience gain modeliIf the current task R is determined to be uninstallediExecuting locally, then the current task RiWrite task set ΘLIn the middle and on the contrary, the current task R is processediWrite task set ΘRIn (1).
Leveraging user service experience augmentationModel-based determination task RiThe method of offloading decision(s) of (1) is: computation task RiOffloading user service experience gain values performed on jth edge server
Figure BDA0002870131080000105
If it is
Figure BDA0002870131080000106
Then task RiExecuting locally, offloading decision variable o i0. Otherwise, task RiOff-load to edge server execution, off-load decision variable oi＝j。
6.4) set of tasks ΘRWherein task R establishes an offloading priorityiOf offload priority PiAs follows:
Figure BDA0002870131080000107
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000108
representing a task RiThe ratio of the power consumption generated by the ith ue in both cases of local execution and execution on the χ -th edge server.
Figure BDA0002870131080000109
Representing a task RiThe ratio of service delay incurred by the ith UE in both cases of executing locally and on the x-th edge server.
Figure BDA00028701310800001010
And
Figure BDA00028701310800001011
respectively represent tasks RiThe difference between the energy cost and the time cost incurred by the system in both cases of executing locally and on the x-th edge server.
Figure BDA00028701310800001012
Respectively represent tasks RiThe energy and time costs incurred by executing the lower system on the x-th edge server.
6.5) determining the task with the maximum unloading priority at the current moment
Figure BDA00028701310800001013
And will unload the task with the largest priority
Figure BDA00028701310800001014
Off-loading to workgroup χ for execution, i.e. task
Figure BDA00028701310800001015
Offload decision
Figure BDA00028701310800001016
Offloading the task with the greatest priority
Figure BDA00028701310800001017
As follows:
Figure BDA0002870131080000111
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000112
and the task sequence number with the largest unloading priority in the solving task is shown.
6.6) the first
Figure BDA0002870131080000113
After each task is unloaded to the x base station-edge server group, the system state variable is updated, namely:
Figure BDA0002870131080000114
Figure BDA0002870131080000115
Figure BDA0002870131080000116
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000117
respectively representing the updated unallocated task set, the computation workload and the communication workload of the chi-th base station-edge server group;
Figure BDA0002870131080000118
representing the computation workload and the communication workload of the chi-th base station-edge server group before updating;
Figure BDA0002870131080000119
indicating the unloading of
Figure BDA00028701310800001110
The amount of data that an individual task needs to transmit in a communication link;
6.7) return to step 6.2) until
Figure BDA00028701310800001111
The idle system resources for the empty set or all base station-edge server groups are not sufficient to handle the task offload request.
6.8) repeating the step 6.2) to the step 6.7) to obtain a task unloading decision set
Figure BDA00028701310800001112
The task unloading decision set O is the task unloading scheme.
Example 2:
the system based on the cost-efficiency optimization method for task unloading in the mobile edge computing system comprises the mobile edge computing system, an energy cost and time cost computing module, a single-task unloading decision generating module, a task unloading cost-efficiency computing module, a task unloading scheme generating module and a database.
The edge server is deployed at a location that is distant from base station i. Base stations and edge servers with a linear distance equal to l are denoted as a base station-edge server group.
All base station-edge server groups as a set
Figure BDA00028701310800001113
Wherein, the operation capability of the jth BS-EDM group is recorded as
Figure BDA00028701310800001114
Figure BDA00028701310800001115
And
Figure BDA00028701310800001116
respectively representing the communication capability of the base station in the jth base station-edge server group and the computing capability of the edge server.
Clustering users in hot spot regions in a mobile edge computing system
Figure BDA00028701310800001117
The hot spot area is a cross coverage area of the communication signals of the base station. Wherein the unloading task of the ith user is recorded as
Figure BDA00028701310800001118
fiAnd ciRespectively representing the amount of data that needs to be transferred in the communication link for task offloading and the number of CPU run cycles needed for task execution. Alpha is alphaiAnd betaiRespectively represents the sensitivity of the service to the service energy consumption and the service delay of the task, and alphai+βi＝1。
The energy cost and time cost calculation module calculates a task RiEnergy costs required for local execution on a user equipment
Figure BDA0002870131080000121
Task RiTime cost T required for local execution on user equipmenti LTask RiEnergy cost required to offload to jth BS-edge server group for execution
Figure BDA0002870131080000122
And task RiTime cost required for offloading to jth BS-edge server group for execution
Figure BDA0002870131080000123
And sending the data to a single task unloading decision generation module and a task unloading cost efficiency calculation module.
Task RiEnergy costs required for local execution on a user equipment
Figure BDA0002870131080000124
And time cost Ti LRespectively as follows:
Figure BDA0002870131080000125
Figure BDA0002870131080000126
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000127
and
Figure BDA0002870131080000128
respectively representing the energy consumed and the computing power consumed by the ith user equipment in one CPU operation cycle. c. CiIndicating the number of CPU run cycles required for task execution.
Task RiEnergy cost required to offload to jth BS-edge server group for execution
Figure BDA0002870131080000129
And time cost
Figure BDA00028701310800001210
Respectively as follows:
Figure BDA00028701310800001211
Figure BDA00028701310800001212
in the formula (I), the compound is shown in the specification,
Figure BDA00028701310800001213
respectively representing the energy cost of data transmission and the energy cost of task execution during the task unloading process.
Figure BDA00028701310800001214
Running the energy consumption of one CPU cycle for the jth edge server. f. ofiAnd ciRespectively representing the amount of data that needs to be transferred in the communication link for task offloading and the number of CPU run cycles needed for task execution.
Wherein the communication capability
Figure BDA00028701310800001215
Computing power
Figure BDA00028701310800001216
Task RiData transmission speed v in a communication linki,jRespectively as follows:
Figure BDA00028701310800001217
Figure BDA00028701310800001218
Figure BDA00028701310800001219
in the formula, piAnd N0Signal transmission power and noise power, g, of the ith user equipment, respectivelyi,jThe channel gain from the ith user equipment to the jth base station.
The single task offload decision generation module stores a user service experience gain model.
The user service experience gain model is as follows:
Figure BDA0002870131080000131
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000132
representing a task RiThe power consumption generated by the ith user equipment when the local execution is carried out.
Figure BDA0002870131080000133
Representing a task RiService delay generated by the ith user equipment when the local execution is carried out.
Figure BDA0002870131080000134
Representing a task RiIs unloaded to oiThe power consumption generated by the ith user equipment when the base station-edge server group is executed.
Figure BDA0002870131080000135
Representing a task RiIs unloaded to oiService delay generated by the ith user equipment when the base station-edge server group is executed.
Figure BDA0002870131080000136
A gain value is experienced for the user service.
The single task offload decision generation module determines task R using a user service experience gain modeliOffload decision-makingAnd sending the data to a task unloading cost efficiency calculation module.
Determining task R using a user service experience gain modeliThe process of offloading decision(s) of (1) is: computation task RiOffloading user service experience gain values performed on jth edge server
Figure BDA0002870131080000137
If it is
Figure BDA0002870131080000138
Then task RiExecuting locally, offloading decision variable o i0. Otherwise, task RiOff-load to edge server execution, off-load decision variable oi＝j。
The task offloading cost-efficiency calculation module calculates a task RiAt an offload decision of oiTime and cost efficiency
Figure BDA0002870131080000139
Task RiAt an offload decision of oiTime and cost efficiency
Figure BDA00028701310800001310
As follows:
Figure BDA00028701310800001311
where η is the coefficient that balances the energy cost term and the time cost term.
And the task unloading scheme generating module determines a task unloading scheme based on the task unloading priority, establishes communication connection for the task unloading request according to the task unloading scheme, and allocates communication resources and computing resources.
Determining a task offloading scheme based on task offloading priorities, the steps comprising:
1) initializing variables, including task set variables
Figure BDA00028701310800001312
Base station-edge server group working capacity variable
Figure BDA00028701310800001313
Unallocated task set variables
Figure BDA00028701310800001314
Offloading decision variables
Figure BDA0002870131080000141
Variables for recording workload conditions for base station-edge server groups
Figure BDA0002870131080000142
Wherein
Figure BDA0002870131080000143
Representing the computational workload of the jth base station-edge server group.
Figure BDA0002870131080000144
Representing the communication workload of the jth base station-edge server group.
2) And taking the base station-edge server group with the minimum workload as a working group, wherein the sequence number χ of the working group is as follows:
Figure BDA0002870131080000145
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000146
the sequence number of the base station-edge server group with the minimum workload in the base station-edge server group is solved.
3) For unallocated task set
Figure BDA0002870131080000147
Classifying to obtain a locally executed task set thetaLAnd off-load to an edge server for remote executionService set thetaR. Wherein, the task set thetaRUser service experience gain k corresponding to remote execution of tasks in (1)i,χGreater than 1.
Task set ΘLAnd task set ΘRAs follows:
Figure BDA0002870131080000148
Figure BDA0002870131080000149
4) set theta for taskRWherein task R establishes an offloading priorityiOf offload priority PiAs follows:
Figure BDA00028701310800001410
in the formula (I), the compound is shown in the specification,
Figure BDA00028701310800001411
representing a task RiThe ratio of the power consumption generated by the ith ue in both cases of local execution and execution on the χ -th edge server.
Figure BDA00028701310800001412
Representing a task RiThe ratio of service delay incurred by the ith UE in both cases of executing locally and on the x-th edge server.
Figure BDA00028701310800001413
And
Figure BDA00028701310800001414
respectively represent tasks RiThe difference between the energy cost and the time cost incurred by the system in both cases of executing locally and on the x-th edge server.
5) Determining task with maximum unloading priority at current moment
Figure BDA00028701310800001415
And will unload the task with the largest priority
Figure BDA00028701310800001416
Off-loading to workgroup χ for execution, i.e. task
Figure BDA00028701310800001417
Offload decision
Figure BDA00028701310800001418
Offloading the task with the greatest priority
Figure BDA0002870131080000151
As follows:
Figure BDA0002870131080000152
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000153
and the task sequence number with the largest unloading priority in the solving task is shown.
6) First, the
Figure BDA0002870131080000154
After each task is unloaded to the x base station-edge server group, the system state variable is updated, namely:
Figure BDA0002870131080000155
Figure BDA0002870131080000156
Figure BDA0002870131080000157
7) returning to the step 2) until
Figure BDA0002870131080000158
The idle system resources for the empty set or all base station-edge server groups are not sufficient to handle the task offload request.
8) Repeating the steps 2) to 7) to obtain a task unloading decision set
Figure BDA0002870131080000159
The task unloading decision set O is the task unloading scheme. The database stores data of an energy cost and time cost calculation module, a single task unloading decision generation module, a task unloading cost efficiency calculation module and a task unloading scheme generation module.
The energy cost and time cost calculation module, the single task unloading decision generation module, the task unloading cost efficiency calculation module and the task unloading scheme generation module can be communicated with each other.
Example 3:
referring to fig. 1 to 2, a cost-effective optimization method for task offloading in a mobile edge computing system mainly includes the following steps:
1) a multi-user multi-base station mobile edge computing system is modeled.
In the mobile edge computing system, communication resources and computing resources are quantified by bandwidth size and CPU operating period, respectively. Edge servers are deployed near each base station to form base station-edge server groups, and all the base station-edge server groups are recorded as a set
Figure BDA00028701310800001510
Wherein, the operation capability of the jth BS-EDM group is recorded as
Figure BDA00028701310800001511
Figure BDA00028701310800001512
And
Figure BDA00028701310800001513
respectively representing the communication capability of the base station in the jth base station-edge server group and the computing capability of the edge server. The users in the hot spot area (shown shaded in FIG. 1) in a mobile edge computing system are grouped together as
Figure BDA00028701310800001514
Wherein the unloading task of the ith user is recorded as
Figure BDA00028701310800001515
Wherein f isiAnd ciRespectively representing the data volume required to be transmitted in a communication link during task unloading and the number of CPU operation cycles required by task execution; alpha is alphaiAnd betaiRespectively represents the sensitivity of the service to which the task belongs to the service energy consumption and the service delay, and satisfies alphai+βi＝1。
2) Modeling the energy and time costs incurred by task offloading.
Task RiIs recorded as an unload decision variable
Figure BDA0002870131080000161
And is
Figure BDA0002870131080000162
When o isiWhen j, the task R is representediUnloading to the jth base station-edge server group for execution; when o isiWhen equal to 0, it means that task R isiUnload failure, task RiCan only be executed locally on the ith user equipment. Aiming at two modes of task remote execution and local execution, the energy and time cost of the invention are respectively modeled, and the model is as follows:
2.1) local execution
Task RiThe energy cost and time cost required for local execution are respectively:
Figure BDA0002870131080000163
Figure BDA0002870131080000164
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000165
and
Figure BDA0002870131080000166
respectively representing the energy consumed by the ith user equipment in one CPU operation cycle and the computing power thereof.
2.2) remote execution
Since one base station-edge server group needs to serve a plurality of unloading tasks at the same time, communication resources and computing resources of the base station-edge server group need to be shared in a plurality of tasks. Specifically, task RiAt offload decision oiIn the case of j, the allocated communication capacity and computing capacity are:
Figure BDA0002870131080000167
Figure BDA0002870131080000168
based on the allocated communication resources, task RiThe data transmission speed in the communication link can be calculated by the following formula:
Figure BDA0002870131080000169
in the formula, piAnd N0Signal transmission power and noise power, g, of the ith user equipment, respectivelyi,jFor channel increase from ith user equipment to jth base stationIt is beneficial to. It should be noted that, in the present invention, it is assumed that the data size of the task execution result is small, and the downlink transmission time of the execution result can be ignored compared with the uplink transmission time of the task data. Thus, the time cost and energy cost of the whole process of task offloading are respectively:
Figure BDA0002870131080000171
Figure BDA0002870131080000172
in the formula (7), the reaction mixture is,
Figure BDA0002870131080000173
respectively representing the energy cost of data transmission and the energy cost of task execution in the task unloading process, wherein
Figure BDA0002870131080000174
Running the energy consumption of one CPU cycle for the jth edge server.
3) A user service experience gain model is modeled.
In order to quantify the gain effect of task unloading on user service experience, the invention provides a quantitative evaluation method for the gain effect of service experience based on a relative evaluation method. Based on two important factors that affect the user service experience: energy consumption and time delay, the invention compares the energy consumption and time delay generated by the local execution and the remote execution of the task to obtain the gain effect of the task unloading on the user service experience. At task RiIs oiThen, the service experience gain value generated by the task offloading behavior is calculated as follows:
Figure BDA0002870131080000175
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000176
respectively represent tasks RiEnergy consumption and service delay generated by the ith user equipment during local execution;
Figure BDA0002870131080000177
respectively represent tasks RiIs unloaded to oiThe energy consumption and service delay generated by the ith user equipment when the base station-edge server group is executed. As can be seen from equation (8), when the task is executed locally, the service experience gain value is 1. Therefore, if a task is offloaded to be executed on an edge server and the expected service experience gain value is less than 1, the task should adopt a local execution policy for efficient utilization of system resources.
4) Modeling task offloading cost efficiency model
Based on the energy and time cost model of task unloading in step 2) and the service experience gain model in step 3), a task R can be obtainediAt an offload decision of oiTime and cost efficiency
Figure BDA0002870131080000178
The calculation is as follows:
Figure BDA0002870131080000179
where η is a coefficient that balances the energy cost term and the time cost term, and
Figure BDA00028701310800001710
larger means that the task offload is more cost effective.
5) A task offload scheme based on task offload priority.
Based on the cost efficiency model in the step 4), with the cost efficiency of executing unloading by maximizing all tasks as an optimization target, the invention provides an unloading strategy making algorithm with time complexity of polynomial time, and the algorithm comprises the following specific steps:
5.1) initializing variables, specifically: task set variables:
Figure BDA0002870131080000181
base station-edge server group working capacity variable
Figure BDA0002870131080000182
Unallocated task set variables
Figure BDA0002870131080000183
Offloading decision variables
Figure BDA0002870131080000184
Variables for recording workload conditions for base station-edge server groups
Figure BDA0002870131080000185
Wherein
Figure BDA0002870131080000186
Respectively representing the computation workload and the communication workload of the jth base station-edge server group.
And 5.2) in order to ensure that the unloading task is executed more efficiently and the load balance between the base station-edge server groups, determining the base station-edge server group with the minimum workload as a working group in the task unloading process. The sequence number of the base station-edge server group is calculated as follows:
Figure BDA0002870131080000187
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000188
the sequence number of the base station-edge server group with the minimum workload in the base station-edge server group is solved.
5.3) set of unallocated tasks
Figure BDA0002870131080000189
Classifying according to the task distanceWhether the service experience gain with the value larger than 1 can be obtained by program execution or not is judged, and the locally executed task set theta is obtained by classificationLAnd the set of tasks Θ remotely executed off-load to the edge serverR，ΘLAnd ΘRIs represented as follows:
Figure BDA00028701310800001810
Figure BDA00028701310800001811
5.4) from the overall view of the moving edge computing system, the set theta isRTask R of (1) to establish an offloading priorityiThe offload priority of (a) is calculated as follows:
Figure BDA00028701310800001812
in the formula (I), the compound is shown in the specification,
Figure BDA00028701310800001813
and
Figure BDA00028701310800001814
respectively represent tasks RiThe power consumption ratio and service delay ratio of the ith user equipment in the two cases of local execution and execution on the chi-edge server;
Figure BDA00028701310800001815
and
Figure BDA00028701310800001816
respectively represent tasks RiThe difference between the energy cost and the time cost incurred by the system in both cases of executing locally and on the x-th edge server.
5.5) based on the unloading priority of the task calculated in the step 5.4), the task with the largest unloading priority in the iteration can be obtained, and the task number is expressed as follows:
Figure BDA0002870131080000191
in the formula (I), the compound is shown in the specification,
Figure BDA0002870131080000192
and the task sequence number with the largest unloading priority in the solving task is shown. Task that gets the maximum offload priority (sequence number is
Figure BDA0002870131080000193
) It is offloaded to the base station-edge server group (with sequence number χ) with the minimum workload determined in step 5.2). The task offloading decision scheme is represented as follows:
Figure BDA0002870131080000194
5.6) the first
Figure BDA0002870131080000195
After each task is unloaded to the x base station-edge server group, the system state variables are updated, and the specific updating contents are as follows:
Figure BDA0002870131080000196
Figure BDA0002870131080000197
Figure BDA0002870131080000198
5.7) returning to step 5.2), and continuing to process the unallocated task set
Figure BDA0002870131080000199
Up to
Figure BDA00028701310800001910
There are insufficient free system resources to handle the task offload request for the empty set or for all base station-edge server groups.
5.8) obtaining a task unloading decision set by iterating the step 5.2) to the step 5.7)
Figure BDA00028701310800001911
And accordingly, communication connection is established for the task unloading request, and communication resources and computing resources are distributed.
Finally, in order to verify the effectiveness and high efficiency of the method, the invention carries out simulation experiments and provides four comparison methods, which specifically comprise the following steps:
comparative method 1: in step 5.4), only the energy cost and the time cost of task unloading are considered, and the task with the minimum cost value is unloaded preferentially in each iteration of the algorithm;
control method 2: in the step 5.4), only the user service experience gain effect of task unloading is considered, and the task with the maximum user service experience gain value is unloaded preferentially in each iteration of the algorithm;
control method 3: randomly generating a task unloading strategy;
control method 4: in step 5.4), a greedy strategy is used for making task unloading priority, and the task with the highest cost efficiency is unloaded in each iteration of the algorithm.
Setting simulation experiment parameters: the number M of the base station-edge server groups is 3, the communication bandwidths of the three base stations are 10MHz, and the deployed edge servers respectively have the computing capacities of 3GHz, 6GHz and 9 GHz. The features of the offloading task are as follows: the amount of communication data is randomly distributed over an interval [300,800 ]]KB; the number of CPU operating cycles required is randomly distributed over intervals [100,1000](ii) a The sensitivity to energy consumption and time delay is randomly distributed in the interval [0,1 ]]. Transmission power p of each user equipmentiThe signal-to-noise ratio of the communication is 8Hz/w and is 0.5 w.
The task unloading process is simulated through MATLAB, and the simulation result is shown in figures 3 and 4. As shown in fig. 3, in the case of η ═ 1, as the number of ues increases, the ratio of the average cost efficiency of the method of the present invention to the 4 comparison methods is: 8.41%, 115.17%, 118.15%, 65.34%; in the case where η is 2, the excess ratio is: 5.85%, 117.57%, 174.89%, 56.13%. As shown in fig. 4, in the case of N300, as η increases, the ratio of the average cost efficiency under the method of the present invention to the 4 comparison methods is: 7.70%, 120.40%, 189.58%, 72.96%; in the case of N being 400, the excess ratios are 9.86%, 104.41%, 188.72%, 58.37%, respectively.

Claims (10)

1. A method for cost-effective optimization of task offloading in a mobile edge computing system, comprising the steps of:
1) and building a mobile edge computing system.
2) Computation task RiSaid energy cost required for local execution on a user equipment
Figure FDA00028701310700000115
And time cost Ti L；
3) Computation task RiEnergy cost required to offload to jth BS-edge server group for execution
Figure FDA0002870131070000011
And time cost
Figure FDA0002870131070000012
4) Setting task RiIs unloaded to the decision variable oiEstablishing a user service experience gain model;
5) computation task RiAt an offload decision of oiTime and cost efficiency
Figure FDA0002870131070000013
6) And determining a task unloading scheme based on the task unloading priority, establishing communication connection for the task unloading request according to the task unloading scheme, and allocating communication resources and computing resources.
2. The method of claim 1, wherein the task offload cost-effective optimization method comprises: the mobile edge computing system comprises a plurality of users, a plurality of base stations and a plurality of edge servers;
the edge server is deployed at a position which is far from a base station l; marking the base station and the edge server with the straight line distance equal to l as a base station-edge server group;
all base station-edge server groups as a set
Figure FDA0002870131070000014
Wherein, the operation capability of the jth BS-EDM group is recorded as
Figure FDA0002870131070000015
Figure FDA0002870131070000016
And
Figure FDA0002870131070000017
respectively representing the communication capacity of the base station in the jth base station-edge server group and the computing capacity of the edge server;
clustering users in hot spot regions in a mobile edge computing system
Figure FDA0002870131070000018
The hot spot area is a cross coverage area of a base station communication signal; wherein the unloading task of the ith user is recorded as
Figure FDA0002870131070000019
fiAnd ciRespectively representing the data volume required to be transmitted in a communication link during task unloading and the number of CPU operation cycles required by task execution;αiand betaiRespectively represents the sensitivity of the service to the service energy consumption and the service delay of the task, and alphai+βi＝1；
Figure FDA00028701310700000110
3. The method of claim 1, wherein task R is a cost-effective optimization of task offloading in a mobile edge computing systemiEnergy costs required for local execution on a user equipment
Figure FDA00028701310700000111
And time cost Ti LRespectively as follows:
Figure FDA00028701310700000112
Figure FDA00028701310700000113
in the formula, deltai LAnd
Figure FDA00028701310700000114
respectively representing the energy and the computing power consumed in one CPU operation cycle of the ith user equipment; c. CiIndicating the number of CPU run cycles required for task execution.
4. The method of claim 1, wherein the task offload cost-effective optimization method comprises: task RiEnergy cost required to offload to jth BS-edge server group for execution
Figure FDA0002870131070000021
And time cost
Figure FDA0002870131070000022
Respectively as follows:
Figure FDA0002870131070000023
Figure FDA0002870131070000024
in the formula (I), the compound is shown in the specification,
Figure FDA0002870131070000025
respectively representing the energy cost of data transmission and the energy cost of task execution in the task unloading process;
Figure FDA0002870131070000026
running the energy consumption of one CPU cycle for the jth edge server; f. ofiAnd ciRespectively representing the data volume required to be transmitted in a communication link during task unloading and the number of CPU operation cycles required by task execution;
wherein the computing power
Figure FDA0002870131070000027
Task RiData transmission speed v in a communication linki,jRespectively as follows:
Figure FDA0002870131070000028
Figure FDA0002870131070000029
in the formula, piAnd N0Signal transmission power and noise power, g, of the ith user equipment, respectivelyi,jChannel gain from the ith user equipment to the jth base station;
communication capability
Figure FDA00028701310700000210
As follows:
Figure FDA00028701310700000211
in the formula (I), the compound is shown in the specification,
Figure FDA00028701310700000212
indicating the communication capability of the base station in the jth base station-edge server group.
5. The method of claim 1, wherein the user service experience gain model is as follows:
Figure FDA00028701310700000213
in the formula (I), the compound is shown in the specification,
Figure FDA00028701310700000214
representing a task RiThe energy consumption generated by the ith user equipment during local execution;
Figure FDA00028701310700000215
representing a task RiService delay generated by the ith user equipment during local execution;
Figure FDA00028701310700000216
representing a task RiIs unloaded to oiEnergy consumption generated by the ith user equipment when the base station-edge server group is executed;
Figure FDA0002870131070000031
representing a task RiIs unloaded to oiService time delay generated by the ith user equipment when the base station-edge server group is executed;
Figure FDA0002870131070000032
a gain value is experienced for the user service.
6. The method of claim 1, wherein the task R is determined using a user service experience gain modeliThe method of offloading decision(s) of (1) is: computation task RiOffloading user service experience gain values performed on jth edge server
Figure FDA0002870131070000033
If it is
Figure FDA0002870131070000034
Then task RiExecuting locally, offloading decision variable oi0; otherwise, task RiOff-load to edge server execution, off-load decision variable oi＝j。
7. The method of claim 6, wherein task R is a cost-effective optimization of task offloading in a mobile edge computing systemiAt an offload decision of oiTime and cost efficiency
Figure FDA0002870131070000035
As follows:
Figure FDA0002870131070000036
where η is the coefficient that balances the energy cost term and the time cost term.
8. The method of claim 1, wherein the task offload cost-effective optimization method comprises: determining a task offloading scheme based on task offloading priorities, the steps comprising:
1) initializing variables, including task set variables
Figure FDA0002870131070000037
Base station-edge server group working capacity variable
Figure FDA0002870131070000038
Unallocated task set variables
Figure FDA0002870131070000039
Offloading decision variables
Figure FDA00028701310700000310
Variables for recording workload conditions for base station-edge server groups
Figure FDA00028701310700000311
Wherein
Figure FDA00028701310700000312
Representing the computational workload of the jth base station-edge server group;
Figure FDA00028701310700000313
Figure FDA00028701310700000314
representing a communication workload of a jth base station-edge server group;
2) and taking the base station-edge server group with the minimum workload as a working group, wherein the sequence number χ of the working group is as follows:
Figure FDA00028701310700000315
in the formula (I), the compound is shown in the specification,
Figure FDA0002870131070000041
the sequence number of the base station-edge server group with the minimum workload in the base station-edge server group is solved;
3) for unallocated task set
Figure FDA0002870131070000042
Classifying to obtain a locally executed task set thetaLAnd the set of tasks Θ remotely executed off-load to the edge serverR(ii) a Wherein, the task set thetaRUser service experience gain corresponding to remote execution of tasks in (1)
Figure FDA0002870131070000043
Greater than 1;
task set ΘLAnd task set ΘRAs follows:
Figure FDA0002870131070000044
Figure FDA0002870131070000045
determining a current task R using a user service experience gain modeliIf the current task R is determined to be uninstallediExecuting locally, then the current task RiWrite task set ΘLIn the middle and on the contrary, the current task R is processediWrite task set ΘRIn (1).
4) Set theta for taskRWherein task R establishes an offloading priorityiOf offload priority PiAs follows:
Figure FDA0002870131070000046
in the formula (I), the compound is shown in the specification,
Figure FDA0002870131070000047
representing a task RiThe ratio of the power consumption generated by the ith UE in both cases of local execution and execution on the χ -th edge server;
Figure FDA0002870131070000048
representing a task RiThe ratio of service delays incurred by the ith UE in both cases of executing locally and on the x-th edge server;
Figure FDA0002870131070000049
and
Figure FDA00028701310700000410
respectively represent tasks RiThe difference between the energy cost and the time cost generated by the system in the two cases of local execution and the execution on the x-th edge server;
Figure FDA00028701310700000411
respectively represent tasks RiEnergy and time costs of executing the down system on the χ -th edge server;
5) determining task with maximum unloading priority at current moment
Figure FDA00028701310700000412
And will unload the task with the largest priority
Figure FDA00028701310700000413
Off-loading to workgroup χ for execution, i.e. task
Figure FDA00028701310700000414
Offload decision
Figure FDA00028701310700000415
Offloading the task with the greatest priority
Figure FDA00028701310700000416
As follows:
Figure FDA00028701310700000417
in the formula (I), the compound is shown in the specification,
Figure FDA00028701310700000418
the task sequence number which represents the maximum unloading priority in the task solution is shown;
6) first, the
Figure FDA00028701310700000419
After each task is unloaded to the x base station-edge server group, the system state variable is updated, namely:
Figure FDA0002870131070000051
Figure FDA0002870131070000052
Figure FDA0002870131070000053
in the formula (I), the compound is shown in the specification,
Figure FDA0002870131070000054
respectively representing the updated unallocated task set, the computation workload and the communication workload of the chi-th base station-edge server group;
Figure FDA0002870131070000055
representing the computation workload and the communication workload of the chi-th base station-edge server group before updating;
Figure FDA0002870131070000056
indicating the unloading of
Figure FDA0002870131070000057
The amount of data that an individual task needs to transmit in a communication link;
7) returning to the step 2) until
Figure FDA0002870131070000058
The idle system resources for the empty set or all base station-edge server groups are not sufficient to process the task offload request;
8) repeating the steps 2) to 7) to obtain a task unloading decision set
Figure FDA0002870131070000059
The task unloading decision set O is the task unloading scheme.
9. The method of claim 8, wherein the task offload cost-effective optimization method comprises: when task RiIs unloaded to the decision variable oiWhen j, task RiUnloading to the jth base station-edge server group for execution; when task RiIs unloaded to the decision variable oiWhen 0, task RiUnload failure, task RiExecuting locally on the ith user equipment;
Figure FDA00028701310700000510
10. a system for a cost-effective optimization method of task offloading in a mobile edge computing system according to any of claims 1 to 9, characterized by: the system comprises a mobile edge computing system, an energy cost and time cost computing module, a single task unloading decision generating module, a task unloading cost efficiency computing module, a task unloading scheme generating module and a database;
the energy cost and time cost calculation module calculates a task RiEnergy costs required for local execution on a user equipment
Figure FDA00028701310700000511
Task RiTime cost T required for local execution on user equipmenti LTask RiEnergy cost required to offload to jth BS-edge server group for execution
Figure FDA00028701310700000512
And task RiTime cost required for offloading to jth BS-edge server group for execution
Figure FDA00028701310700000513
And sending the data to a single task unloading decision generation module and a task unloading cost efficiency calculation module;
the single task unloading decision generation module stores a user service experience gain model;
the single task offload decision generation module determines task R using a user service experience gain modeliAnd sending the unloading decision to a task unloading cost efficiency calculation module;
the task offloading cost-efficiency calculation module calculates a task RiAt an offload decision of oiTime and cost efficiency
Figure FDA00028701310700000514
The task unloading scheme generation module determines a task unloading scheme based on task unloading priority, establishes communication connection for the task unloading request according to the task unloading scheme, and allocates communication resources and computing resources;
the database stores data of an energy cost and time cost calculation module, a single task unloading decision generation module, a task unloading cost efficiency calculation module and a task unloading scheme generation module.
CN202011595152.2A 2020-12-29 2020-12-29 Cost-effective optimization system and method for task offloading in mobile edge computing system Active CN112860429B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011595152.2A CN112860429B (en) 2020-12-29 2020-12-29 Cost-effective optimization system and method for task offloading in mobile edge computing system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011595152.2A CN112860429B (en) 2020-12-29 2020-12-29 Cost-effective optimization system and method for task offloading in mobile edge computing system

Publications (2)

Publication Number Publication Date
CN112860429A true CN112860429A (en) 2021-05-28
CN112860429B CN112860429B (en) 2023-06-16

Family

ID=75998098

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011595152.2A Active CN112860429B (en) 2020-12-29 2020-12-29 Cost-effective optimization system and method for task offloading in mobile edge computing system

Country Status (1)

Country Link
CN (1) CN112860429B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113534829A (en) * 2021-06-11 2021-10-22 南京邮电大学 Daily detecting system that patrols of unmanned aerial vehicle based on edge calculation
CN113852692A (en) * 2021-09-24 2021-12-28 中国移动通信集团陕西有限公司 Service determination method, device, equipment and computer storage medium
CN113891359A (en) * 2021-10-11 2022-01-04 中国联合网络通信集团有限公司 Data processing method, device and storage medium
KR20230032754A (en) * 2021-08-31 2023-03-07 연세대학교 산학협력단 Apparatus and Method for Task Offloading of MEC-Based Wireless Network
CN117032832A (en) * 2023-08-25 2023-11-10 重庆邮电大学 Minimum task calling cost unloading method based on mobile edge calculation

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190141593A1 (en) * 2017-11-03 2019-05-09 Hon Hai Precision Industry Co., Ltd. Methods and related devices for optimizing a mobile edge computing (mec) system
CN109814951A (en) * 2019-01-22 2019-05-28 南京邮电大学 The combined optimization method of task unloading and resource allocation in mobile edge calculations network
CN110377353A (en) * 2019-05-21 2019-10-25 湖南大学 Calculating task uninstalling system and method
WO2020023115A1 (en) * 2018-07-27 2020-01-30 Futurewei Technologies, Inc. Task offloading and routing in mobile edge cloud networks
CN111199740A (en) * 2019-12-31 2020-05-26 重庆大学 Unloading method for accelerating automatic voice recognition task based on edge calculation
CN111240701A (en) * 2019-12-31 2020-06-05 重庆大学 Task unloading optimization method for end-edge-cloud collaborative computing
CN111372314A (en) * 2020-03-12 2020-07-03 湖南大学 Task unloading method and task unloading device based on mobile edge computing scene

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190141593A1 (en) * 2017-11-03 2019-05-09 Hon Hai Precision Industry Co., Ltd. Methods and related devices for optimizing a mobile edge computing (mec) system
WO2020023115A1 (en) * 2018-07-27 2020-01-30 Futurewei Technologies, Inc. Task offloading and routing in mobile edge cloud networks
CN109814951A (en) * 2019-01-22 2019-05-28 南京邮电大学 The combined optimization method of task unloading and resource allocation in mobile edge calculations network
CN110377353A (en) * 2019-05-21 2019-10-25 湖南大学 Calculating task uninstalling system and method
CN111199740A (en) * 2019-12-31 2020-05-26 重庆大学 Unloading method for accelerating automatic voice recognition task based on edge calculation
CN111240701A (en) * 2019-12-31 2020-06-05 重庆大学 Task unloading optimization method for end-edge-cloud collaborative computing
CN111372314A (en) * 2020-03-12 2020-07-03 湖南大学 Task unloading method and task unloading device based on mobile edge computing scene

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
K. ZHANG ET AL.: "Energy-efficient offloading for mobile edge computing in 5g heterogeneous networks", 《IEEE ACCESS》 *
PAN Y ET AL.: "Energy-Efficient NOMA-Based Mobile Edge Computing Offloading", 《IEEE COMMUNICATIONS LETTERS》 *
李邱苹 等: "移动边缘计算中的计算卸载和资源管理方案", 《电信科学》 *
薛建彬 等: "移动边缘计算中基于能量收集的能效优化方案", 《北京邮电大学学报》, vol. 43, no. 5 *
薛建彬;安亚宁;: "基于边缘计算的新型任务卸载与资源分配策略", 计算机工程与科学, no. 06 *

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113534829A (en) * 2021-06-11 2021-10-22 南京邮电大学 Daily detecting system that patrols of unmanned aerial vehicle based on edge calculation
CN113534829B (en) * 2021-06-11 2024-04-05 南京邮电大学 Unmanned aerial vehicle daily patrol detecting system based on edge calculation
KR20230032754A (en) * 2021-08-31 2023-03-07 연세대학교 산학협력단 Apparatus and Method for Task Offloading of MEC-Based Wireless Network
KR102562732B1 (en) 2021-08-31 2023-08-01 연세대학교 산학협력단 Apparatus and Method for Task Offloading of MEC-Based Wireless Network
CN113852692A (en) * 2021-09-24 2021-12-28 中国移动通信集团陕西有限公司 Service determination method, device, equipment and computer storage medium
CN113852692B (en) * 2021-09-24 2024-01-30 中国移动通信集团陕西有限公司 Service determination method, device, equipment and computer storage medium
CN113891359A (en) * 2021-10-11 2022-01-04 中国联合网络通信集团有限公司 Data processing method, device and storage medium
CN113891359B (en) * 2021-10-11 2023-06-20 中国联合网络通信集团有限公司 Data processing method, device and storage medium
CN117032832A (en) * 2023-08-25 2023-11-10 重庆邮电大学 Minimum task calling cost unloading method based on mobile edge calculation
CN117032832B (en) * 2023-08-25 2024-03-08 重庆邮电大学 Minimum task calling cost unloading method based on mobile edge calculation

Also Published As

Publication number Publication date
CN112860429B (en) 2023-06-16

Similar Documents

Publication Publication Date Title
CN112860429A (en) Cost-efficiency optimization system and method for task unloading in mobile edge computing system
CN110493360B (en) Mobile edge computing unloading method for reducing system energy consumption under multiple servers
CN113950066B (en) Single server part calculation unloading method, system and equipment under mobile edge environment
CN109413724B (en) MEC-based task unloading and resource allocation scheme
CN111586720B (en) Task unloading and resource allocation combined optimization method in multi-cell scene
CN107766135B (en) Task allocation method based on particle swarm optimization and simulated annealing optimization in moving cloud
CN110198278B (en) Lyapunov optimization method for vehicle networking cloud and edge joint task scheduling
CN111800828B (en) Mobile edge computing resource allocation method for ultra-dense network
CN113242568A (en) Task unloading and resource allocation method in uncertain network environment
CN111372314A (en) Task unloading method and task unloading device based on mobile edge computing scene
CN110096362B (en) Multitask unloading method based on edge server cooperation
CN110928654A (en) Distributed online task unloading scheduling method in edge computing system
CN109756912B (en) Multi-user multi-base station joint task unloading and resource allocation method
CN114189892A (en) Cloud-edge collaborative Internet of things system resource allocation method based on block chain and collective reinforcement learning
CN113286329B (en) Communication and computing resource joint optimization method based on mobile edge computing
CN110233755B (en) Computing resource and frequency spectrum resource allocation method for fog computing in Internet of things
CN111711962B (en) Cooperative scheduling method for subtasks of mobile edge computing system
CN113286317A (en) Task scheduling method based on wireless energy supply edge network
WO2022242468A1 (en) Task offloading method and apparatus, scheduling optimization method and apparatus, electronic device, and storage medium
CN113573363A (en) MEC calculation unloading and resource allocation method based on deep reinforcement learning
CN114153515B (en) Highway internet of vehicles task unloading algorithm based on 5G millimeter wave communication
CN114885422A (en) Dynamic edge computing unloading method based on hybrid access mode in ultra-dense network
He et al. An offloading scheduling strategy with minimized power overhead for internet of vehicles based on mobile edge computing
CN111930435A (en) Task unloading decision method based on PD-BPSO technology
CN115955479A (en) Task rapid scheduling and resource management method in cloud edge cooperation system

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