CN112105062A - Mobile edge computing network energy consumption minimization strategy method under time-sensitive condition - Google Patents

Mobile edge computing network energy consumption minimization strategy method under time-sensitive condition Download PDF

Info

Publication number
CN112105062A
CN112105062A CN202010975206.1A CN202010975206A CN112105062A CN 112105062 A CN112105062 A CN 112105062A CN 202010975206 A CN202010975206 A CN 202010975206A CN 112105062 A CN112105062 A CN 112105062A
Authority
CN
China
Prior art keywords
user
optimal
calculation
energy consumption
computing
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
CN202010975206.1A
Other languages
Chinese (zh)
Other versions
CN112105062B (en
Inventor
赵明雄
罗佳
余俊杰
李文涛
包聆言
邓彪
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.)
Yunnan University YNU
Original Assignee
Yunnan University YNU
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 Yunnan University YNU filed Critical Yunnan University YNU
Priority to CN202010975206.1A priority Critical patent/CN112105062B/en
Publication of CN112105062A publication Critical patent/CN112105062A/en
Application granted granted Critical
Publication of CN112105062B publication Critical patent/CN112105062B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/16—Central resource management; Negotiation of resources or communication parameters, e.g. negotiating bandwidth or QoS [Quality of Service]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/02—Traffic management, e.g. flow control or congestion control
    • H04W28/08—Load balancing or load distribution
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W72/00—Local resource management
    • H04W72/50—Allocation or scheduling criteria for wireless resources
    • H04W72/52—Allocation or scheduling criteria for wireless resources based on load
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Computational Mathematics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Algebra (AREA)
  • Quality & Reliability (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention relates to a mobile edge computing network energy consumption minimization strategy method under a time-sensitive condition, which is characterized by comprising the following steps of: the method comprises the following steps: an initialization stage: the nodes in the network obtain the basic configuration information of the network at the present stage and carry out the initialization of the relevant parameters; step two: establishing an energy consumption minimum system optimization model; step three: according to the constraint conditions, obtaining an initial optimal unloading ratio; step four: returning to the step two under the condition of meeting the constraint condition, and entering the next round of iterative loop calculation according to the block coordinate descent algorithm; step five: and after the circulation reaches the maximum limit times, obtaining the optimal unloading ratio, the computing resources and the subcarrier allocation expression of different computing tasks of each user node. Under the condition of meeting the QoE requirement of a user on a time-efficient computing task, the invention optimizes three strong coupling optimization variables of unloading ratio, computing resource allocation and subcarrier allocation by decoupling optimization problems, thereby realizing the minimization of the total energy consumption of the whole mobile edge computing network.

Description

Mobile edge computing network energy consumption minimization strategy method under time-sensitive condition
Technical Field
The invention relates to the technical field of information and communication engineering, in particular to an orthogonal frequency division multiple access technology, a partial unloading technology of mobile edge calculation and a resource allocation technology in a wireless communication system.
Background
Mobile Edge Computing (MEC) provides a user with a high quality of experience (QoE) by placing a server close to the end user. MEC helps to save energy compared to local computation, but also causes communication delays. There are several studies in sequence aiming at the problem of meeting the user quality experience requirements in a multi-user environment and reducing the energy consumption of the mobile device as much as possible. Dai Yueyue studied the minimum of the total energy consumption of the MEC system from the perspective of user association optimization in Joint computing and user association in Multi task Mobile edge. Ren Jinke realizes the energy consumption minimization of MEC network based on time division multiple access in the text of space Optimization for Resource Allocation in Mobile-Edge Computation Offloading. Laizhong Cui in the "Joint Optimization of Energy Consumption and Latency in Mobile Edge Computing for Internet of Things" a balance between Energy Consumption minimization and time delay is considered, and the scenario is modeled as a constrained multi-objective Optimization problem. Point Paymard in Joint Optimization of Energy Consumption and Latency in Mobile Edge Computing for Internet of Things researches the problems of profit maximization, delay minimization and the like of operators and Energy Consumption minimization with or without calculation delay requirements based on orthogonal frequency division multiple access technology. Khalii, Ata in the section of Joint resource allocation and of streaming decision in mobile edge computing considers the QoE requirement of time-sensitive computing task, but only considers the complete offloading technology.
Disclosure of Invention
The invention aims to solve the defects of the problems and provides an invention for minimizing the energy consumption of a mobile edge network system, which is improved by comprehensively considering a partial unloading technology and a resource allocation technology under the condition of meeting time sensitivity in the mobile edge network based on orthogonal frequency division multiple access, and finally obtaining the optimal resource allocation and unloading ratio to achieve the energy consumption minimization of the system by decoupling a plurality of strong coupling optimization variables and optimization problems related to the energy consumption minimization problem.
The invention is realized by adopting the following technical scheme.
A mobile edge computing network energy consumption minimization strategy method based on orthogonal frequency division multiple access and under the time-sensitive condition comprises the following steps:
the method comprises the following steps: an initialization stage: the nodes in the network at this stage obtain the basic configuration information of the network and initialize the related parameters including local computing resources and computing power, the subcarrier set in the network, the current cycle, the maximum cycle,
Figure RE-GDA0002767953690000021
Initial values of auxiliary variables such as α, β, γ, etc.;
step two: establishing an energy consumption minimum system optimization model according to an optimization total target of the minimum system energy consumption and the constrained conditions such as time delay, maximum computing resources and processing speed, and substituting related parameters into the optimization model;
step three: according to the constraint condition, obtaining an initial optimal unloading ratio, and further calculating to obtain the calculation resource allocated to the user K, the subcarrier set allocated to the user K, the transmission energy of the subcarriers in the set and other calculation related parameters according to the optimal unloading ratio;
step four: returning to the step two according to the obtained optimal load ratio, calculation resources and subcarrier allocation strategy of the current round under the condition of meeting the constraint condition, and entering the next round of iterative loop calculation according to a block coordinate descent algorithm;
step five: and after the circulation reaches the maximum limit times, obtaining the optimal unloading ratio, the computing resources and the subcarrier allocation expression of different computing tasks of each user node.
The system optimization model is as follows:
Figure RE-GDA0002767953690000031
Figure RE-GDA0002767953690000032
Figure RE-GDA0002767953690000033
Figure RE-GDA0002767953690000034
Figure RE-GDA0002767953690000035
Figure RE-GDA0002767953690000036
Figure RE-GDA0002767953690000037
Ekrepresenting the total energy consumption of user k, λkData occupation total data R processed by MEC server on behalf of user task loadkProportion of (a) tk，lTime of local calculation on behalf of the user, tk，offTime of use, f, on behalf of MEC server load computationk，mRepresenting the calculation frequency assigned by the MEC server to user k, F representing the available total CPU calculation frequency of the MEC server, T representing a time slot length, xk，nRepresenting whether subcarrier n is allocated to user k, is a binary variable; wherein
Figure RE-GDA0002767953690000038
The load calculation time tk，offThe method mainly comprises two parts, namely uplink transmission time of data from a user k to an MEC server and data processing time of the MEC server; wherein R iskRepresents the total data to be processed; r iskRepresenting the transmission rate of the MEC server to obtain the user load data:
Figure RE-GDA0002767953690000039
the simplified system model is: decoupling the problem P into sub-problems Po and Ps, and further performing decoupling calculation on k users respectively to obtain an optimal unloading ratio and calculation resources of the corresponding user k and a subcarrier allocation strategy;
solving for the optimal unload ratio λ for user kk ★The transformation is to minimize the problem as follows:
Figure RE-GDA00027679536900000310
s.t.0≤λk≤1，
Figure RE-GDA00027679536900000311
and is provided with
Figure RE-GDA00027679536900000312
The Ps problem can be modeled as follows, which will be further decoupled by the following steps:
Figure RE-GDA0002767953690000041
s.t.(7d)-(7g)
Figure RE-GDA0002767953690000042
introducing a non-negative auxiliary variable
Figure RE-GDA00027679536900000413
The Ps problem was converted to Ps 1:
Figure RE-GDA0002767953690000043
s.t.(7b)-(7g)，(11b)
Figure RE-GDA0002767953690000044
further, the problem Ps1 is converted into an unconstrained lagrange multiplier:
Figure RE-GDA0002767953690000045
wherein
Figure RE-GDA0002767953690000046
Three of gamma are non-negative Langerhans and are recorded
Figure RE-GDA0002767953690000047
To satisfy the constraint of 0 ≦
Figure RE-GDA0002767953690000048
The set of f (a) to (b),
Figure RE-GDA0002767953690000049
to satisfy the constraint
Figure RE-GDA00027679536900000410
Eyes of a user
Figure RE-GDA00027679536900000411
The set of x of (a); then the lagrangian dual function can be defined as follows:
Figure RE-GDA00027679536900000412
namely the Lagrangian dual problem to be solved is
max g(α，β，γ)
s.t.α≥0，β≥0，γ≥0.
The Ps problem decoupling is completed.
Solving the optimal allocation strategy of the computing resources:
given that the BCD algorithm KKT condition is satisfied, the optimal solution for computing resource allocation that is easily obtained satisfies the following sufficient requirements:
Figure RE-GDA0002767953690000051
wherein
Figure RE-GDA00027679536900000516
Is a convex function, and
Figure RE-GDA0002767953690000052
with fk，mMonotonically increasing and having f greater than or equal to 0k，m≤F。
The following dichotomy thought is adopted to solve the optimal computing resource allocation strategy
Figure RE-GDA0002767953690000053
The upper limit value of the calculation frequency for initializing the user k to be distributed is
Figure RE-GDA0002767953690000054
Lower limit value of
Figure RE-GDA0002767953690000055
And constants for controlling accuracy∈3；
The following procedure was repeated:
order to
Figure RE-GDA0002767953690000056
Based on the above sufficient requirements, calculating
Figure RE-GDA0002767953690000057
If it is not
Figure RE-GDA0002767953690000058
Then order
Figure RE-GDA0002767953690000059
Otherwise make
Figure RE-GDA00027679536900000510
Repeating the above process until
Figure RE-GDA00027679536900000511
Obtaining the optimal computing resource allocation strategy after the circulation is finished
Figure RE-GDA00027679536900000512
Solving the subcarrier allocation strategy can be converted into solving by the following Lagrangian function:
Figure RE-GDA00027679536900000513
wherein,
Figure RE-GDA00027679536900000514
assuming that subcarrier n is allocated to user k, we can get:
Figure RE-GDA00027679536900000515
wherein,
Figure RE-GDA0002767953690000061
Figure RE-GDA0002767953690000062
wherein
Figure RE-GDA0002767953690000063
To minimize each
Figure RE-GDA0002767953690000064
Nth subcarrier xnThe optimal solution of whether to assign to user k may be expressed as follows:
Figure RE-GDA0002767953690000065
thus, the optimal subcarrier set currently allocated to the user k is obtained.
Updating a related calculation parameter phi:
first from the { f obtained★，X★H, update pk，nLet us order
Figure RE-GDA0002767953690000066
The φ optimization problem can be modeled as Ps 2:
Figure RE-GDA0002767953690000067
Figure RE-GDA00027679536900000614
Figure RE-GDA0002767953690000068
wherein
Figure RE-GDA0002767953690000069
NkIndicating the selection of the set of subcarriers allocated to user k.
Figure RE-GDA00027679536900000610
For k users, Ps2 is decoupled into Ps 3:
Figure RE-GDA00027679536900000611
Figure RE-GDA00027679536900000612
Figure RE-GDA00027679536900000613
the following can be obtained:
Figure RE-GDA0002767953690000071
wherein,
Figure RE-GDA0002767953690000072
Figure RE-GDA0002767953690000073
wherein
Figure RE-GDA0002767953690000074
Up to lagrange function
Figure RE-GDA0002767953690000078
Converge to obtain phik ★。
Updating the relevant auxiliary parameters α, β, γ:
has obtained f according to the above steps★，X★，φk ★Therefore, it is possible to start solving the lagrange dual function, which is a convex function, so that the dual variables (α, β, β, γ) are updated according to the gradient descent method:
Figure RE-GDA0002767953690000075
Figure RE-GDA0002767953690000076
wherein
Figure RE-GDA0002767953690000077
Is the corresponding step size of the dual variable in the loop;
the loop ends with the update (α, β, γ) until the three parameters converge.
The optimal duty ratio, the optimal computing resources, the optimal subcarrier set and the values of the related auxiliary parameters of the current round of circulation are obtained.
The method has the advantages that the method for minimizing the energy consumption of the mobile edge computing network based on the orthogonal frequency division multiple access under the time-sensitive condition comprehensively considers the user QoE requirement, the partial unloading ratio, the computing resource, the subcarrier allocation and other strong coupling variables of the time-sensitive application. Under the condition of meeting the QoE requirement of a user on a time-efficient computing task, three strong coupling optimization variables of the unloading ratio, the computing resource allocation and the subcarrier allocation are optimized through a decoupling optimization problem, an optimal partial unloading strategy and a resource allocation strategy are obtained, and the total energy consumption of the whole mobile edge computing network is minimized.
The invention is further explained below with reference to the drawings and the detailed description.
Drawings
FIG. 1 is an exemplary diagram of a network topology of the present invention;
FIG. 2 is a simulation diagram of the total energy consumption of the system varying with the number of users;
FIG. 3 is a simulation of the total energy consumption of a system as a function of the severity of user quality experience (QoE) requirements;
FIG. 4 is a simulation diagram of the variation of total energy consumption of the system with the number of subcarriers;
FIG. 5 is a block diagram of the logical thinking of the method of the present invention.
Fig. 2-5 illustrate the average unload rate of the proposed algorithm PA and fixed unload ratio method FR, local calculation method LC, and PA for the present invention. The horizontal axis represents the number of users, the left vertical axis represents energy consumption (unit: joules), and the right vertical axis represents unloading rate.
Detailed Description
The present invention assumes that the topology of the entire network is as shown in fig. 1. Consider a wireless network with a base station integrated with an MEC server that calculates tasks for user load data. The network is provided with K user terminals, each user terminal has the same maximum transmission capacity and is independently distributed in a circular area with the MEC server as the center of a circle and the radius of 30 meters. Each user node in the network is equipped with a single antenna. Note the book
Figure RE-GDA0002767953690000081
For user set, note
Figure RE-GDA0002767953690000082
Each subcarrier has a bandwidth of B for the set of orthogonal subcarriers, and 1 subcarrier can only be allocated to 1 user. The path loss function obeys a variance of | β | d for network transmissions within the simulation environment-2Where β represents a short-term channel fading coefficient and d represents a distance between two nodes.
Using four parameters { Rk，ck，λk，tkTo users
Figure RE-GDA0002767953690000091
Is described, RkRepresenting the total data to be processed. c. CkRepresenting the number of revolutions at which the CPU processes 1 bit of data. f. ofkRepresenting a user
Figure RE-GDA0002767953690000092
CPU frequency of fk，mRepresenting the calculation frequency of the MEC server to the user k, and satisfying that the total calculation frequency allocated to all the users is less than the CPU calculation frequency F of the MEC server, i.e. sigmak∈Kfk，m≤F。λk∈[0，1]Data occupation total data R processed by MEC server on behalf of user task loadkThe ratio of (a) to (b). t is tkRepresenting the maximum tolerable delay (for user k, t)kMust not be greater than the channel coherence time to ensure that the radio channel can remain exclusive in a time slot of length T, and each TkNot necessarily the same).
Time delay t of user kkTime-of-use t dividable into user local calculationsk，lAnd elapsed time t of MEC server load calculationk，off. Since the user local computation and the computation of the MEC server are performed synchronously, tk＝ max{tk，l，tk，off}。
The user local computation time may be expressed as:
Figure RE-GDA0002767953690000093
the load calculation may be expressed in terms of:
Figure RE-GDA0002767953690000094
the load calculation time mainly comprises two parts, namely the uplink transmission time of data from the user k to the MEC server, and the data processing time of the MEC server. Wherein r iskIndicating MEC server to obtain user load dataTransmission rate of (2):
Figure RE-GDA0002767953690000095
wherein g isK，nRepresenting the signal gain between user k and the base station,
Figure RE-GDA0002767953690000096
representing the variance of gaussian white noise for the base station subcarrier n. p is a radical ofk，nRepresenting the transmission energy of user k on subcarrier n, and for simplifying the calculation
Figure RE-GDA0002767953690000101
Wherein
Figure RE-GDA0002767953690000102
Representing the maximum transmission energy, NkIndicating the number of sub-carriers allocated to user k, p if user k does not load data to MEC server k，n0. Binary variable xk，nIndicating whether subcarrier n is allocated to user k.
Calculating the total energy consumption E for processing a taskkCan be divided into local calculation of energy consumption Ek，lAnd MEC server load computing energy consumption Ek，offTwo parts are as follows:
Ek＝Ek，l+Ek，off＝Ek，l+Ek，u+Ek，m (4)
local computing energy consumption Ek，lCan be expressed as after substituting into (1)
Figure RE-GDA0002767953690000103
Given processor computation rate (i.e., CPU frequency) fkThe total energy consumption of the processor can be expressed as
Figure RE-GDA0002767953690000104
(Joule per second), KkRepresenting calculated energy efficiency coefficients of a user processor(associated with the processor chip of user k).
Energy consumption E due to MEC load calculationk，offIncluding the energy consumption of upstream transmission and the energy consumption of remote computing load data.
Figure RE-GDA0002767953690000105
Given fk，mComputing frequency assigned to user k on behalf of MEC server, where kmRepresenting the computational energy efficiency coefficient of an MEC Server (associated with the MEC Server processor chip)
According to the above conditions, the optimization model of the system minimum energy consumption problem solved by the invention is P.
Figure RE-GDA0002767953690000106
Figure RE-GDA0002767953690000107
Figure RE-GDA0002767953690000108
Figure RE-GDA0002767953690000109
Figure RE-GDA0002767953690000111
Figure RE-GDA0002767953690000112
Figure RE-GDA0002767953690000113
The strategy of the invention decomposes the P problem into a Po sub-problem solving the unloading strategy and a Ps problem solving the sub-carrier and computing resource allocation. The following steps are solved according to the BCD method.
First step to solve the optimal load ratio λ★. Given an initial computing power distribution strategy f and a subcarrier distribution strategy X, in order to obtain an optimal load ratio lambda★The Po problem needs to be solved. The Po optimization problem is modeled as:
Figure RE-GDA0002767953690000114
s.t.(7b)(7c)
to obtain the minimum total energy consumption in the network, only the minimum energy consumption of the kth user needs to be solved, which is equivalent to decoupling the Po problem into k subproblems and recording the k subproblems as Po 1. The Po1 problem can be described as:
Figure RE-GDA0002767953690000115
s.t.0≤λk≤1， (9b)
Figure RE-GDA0002767953690000116
wherein given (f, X), the constraint according to (9b) (9c), and
Figure RE-GDA0002767953690000117
the derivation result can obtain the optimal load ratio lambda of the user kk ★The value of (A) is as follows:
Figure RE-GDA0002767953690000118
the optimum load ratio λ obtained according to (10)★. Optimal allocation strategy to obtain allocation of computing resources of MEC server and allocation of subcarriers to each userSolving the problem of Ps, i.e. solving the objective function sigmak∈KEk，offThus, Ps the system optimization model is:
Figure RE-GDA0002767953690000121
Figure RE-GDA0002767953690000122
due to rkIs a denominator, can not directly convert the Ps problem into the dual domain problem, thereby introducing a non-negative auxiliary variable
Figure RE-GDA0002767953690000123
The Ps problem was converted to Ps 1:
Figure RE-GDA0002767953690000124
s.t.(7b)-(7g)，(11b)
Figure RE-GDA0002767953690000125
further, the problem Ps1 is converted into an unconstrained lagrange multiplier:
Figure RE-GDA0002767953690000126
wherein
Figure RE-GDA0002767953690000127
Three of gamma are non-negative Langerhans and are recorded
Figure RE-GDA0002767953690000128
To satisfy the set of f of the constraint (7d),
Figure RE-GDA0002767953690000129
set of x to satisfy constraints (7f) and (7 g). Then the lagrangian dual function can be defined as follows:
Figure RE-GDA00027679536900001210
namely the Lagrangian dual problem to be solved is
max g(α，β，γ) (15)
s.t.α≥0，β≥0，γ≥0.
And secondly, solving a calculation resource optimal allocation strategy according to the load ratio and the transmission energy consumption of the current cycle.
Under the condition of meeting the BCD algorithm KKT condition, the optimal solution of computing resource allocation is easy to obtain and meets the following sufficient necessary conditions:
Figure RE-GDA0002767953690000131
wherein
Figure RE-GDA0002767953690000132
Is a convex function, and
Figure RE-GDA0002767953690000133
with fk，mMonotonically increasing and having f greater than or equal to 0k，mF is less than or equal to F. Therefore, the following dichotomy thought can be adopted to solve the optimal computing resource allocation strategy
Figure RE-GDA0002767953690000134
The upper limit value of the calculation frequency for initializing the user k to be distributed is
Figure RE-GDA0002767953690000135
Lower limit value of
Figure RE-GDA0002767953690000136
And a constant e for control accuracy3。
The following procedure was repeated:
order to
Figure RE-GDA0002767953690000137
According to (16) calculating
Figure RE-GDA0002767953690000138
If it is not
Figure RE-GDA0002767953690000139
Then order
Figure RE-GDA00027679536900001310
Otherwise make
Figure RE-GDA00027679536900001311
Repeating the above process until
Figure RE-GDA00027679536900001312
Obtaining the optimal computing resource allocation strategy after the circulation is finished
Figure RE-GDA00027679536900001313
Thirdly, solving an optimal subcarrier allocation strategy according to the load ratio and the transmission energy consumption of the current cycle and the calculation resource allocation strategy:
rewriting the optimal calculation power distribution strategy obtained in the second step
Figure RE-GDA00027679536900001314
Figure RE-GDA00027679536900001315
Wherein,
Figure RE-GDA00027679536900001316
based on (17), assuming that subcarrier n is allocated to user k, we can get:
Figure RE-GDA00027679536900001317
wherein,
Figure RE-GDA0002767953690000141
the problem is further decomposed into an independently solvable sub-problem:
Figure RE-GDA0002767953690000142
wherein
Figure RE-GDA0002767953690000143
To minimize each
Figure RE-GDA0002767953690000144
xnThe optimal solution of (a) can be expressed as follows:
Figure RE-GDA0002767953690000145
thus, the optimal subcarrier set currently allocated to the user k is obtained.
And fourthly, updating the related calculation parameter phi.
First obtained from the above step { f★，X★H, update pk，nLet us order
Figure RE-GDA0002767953690000146
Further solving an optimal auxiliary variable phi by an optimization problem as follows★. The φ optimization problem can be modeled as Ps 2:
Figure RE-GDA0002767953690000147
s.t.(11b)
Figure RE-GDA0002767953690000148
wherein
Figure RE-GDA0002767953690000149
NkRepresents the set of subcarriers selected (22) for allocation to user k. Further, for k users, Ps2 is decoupled into Ps 3:
Figure RE-GDA00027679536900001410
s.t. (23b)
Figure RE-GDA00027679536900001411
the following can be obtained:
Figure RE-GDA00027679536900001412
wherein,
Figure RE-GDA0002767953690000151
Figure RE-GDA0002767953690000152
wherein
Figure RE-GDA0002767953690000153
Up to (13) interior Lagrangian function
Figure RE-GDA0002767953690000154
Converging, otherwise continuously executing the fourth step and the fifth step, and updating the related auxiliary parameters alpha, beta, gamma
Has obtained f according to the above steps★，X★，φk ★It is therefore possible to start solving (15) the proposed dual function, which is a convex function, so that the dual variables (α, β, β, γ) are updated according to the gradient descent method:
Figure RE-GDA0002767953690000155
Figure RE-GDA0002767953690000156
Figure RE-GDA0002767953690000157
wherein
Figure RE-GDA0002767953690000158
Are the corresponding step sizes of the dual variables in the loop.
According to the fifth step of continuously cycling (28), updating (alpha, beta, gamma) until the three parameters are converged, and then cycling is finished.
So far, the optimal load ratio, the optimal computing resource, the optimal subcarrier set and the values of the related auxiliary parameters of the current round of circulation are obtained, the first step is returned, and the obtained parameter values are used as the initial values of the next round of circulation to be computed until the maximum circulation time limit is reached.
Figure RE-GDA0002767953690000159
Figure RE-GDA0002767953690000161
The above description is only a part of specific embodiments of the present invention (since the formula of the present invention belongs to the numerical range, the embodiments are not exhaustive, and the protection scope of the present invention is subject to the numerical range and other technical point ranges), and the detailed contents or common knowledge known in the schemes are not described too much. It should be noted that the above-mentioned embodiments do not limit the present invention in any way, and all technical solutions obtained by means of equivalent substitution or equivalent transformation for those skilled in the art are within the protection scope of the present invention. The scope of the claims of the present application shall be determined by the contents of the claims, and the description of the embodiments and the like in the specification shall be used to explain the contents of the claims.

Claims (10)

1. A method for minimizing energy consumption of a mobile edge computing network under a time-sensitive condition is characterized by comprising the following steps:
the method comprises the following steps: an initialization stage: in the present phase, the nodes in the network obtain the basic configuration information of the network and initialize the relevant parameters, including the local computing resources and computing power, the subcarrier set in the network, the current cycle, the maximum cycle and the initial value of the auxiliary variable;
step two: establishing an energy consumption minimum system optimization model according to an optimization total target of the minimum system energy consumption and the constrained conditions such as time delay, maximum computing resources and processing speed, and substituting related parameters into the optimization model;
step three: according to the constraint condition, obtaining an initial optimal unloading ratio, and further calculating to obtain the calculation resource allocated to the user K, the subcarrier set allocated to the user K, the transmission energy of the subcarriers in the set and other calculation related parameters according to the optimal unloading ratio;
step four: returning to the step two according to the obtained optimal load ratio, calculation resources and subcarrier allocation strategy of the current round under the condition of meeting the constraint condition, and entering the next round of iterative loop calculation according to a block coordinate descent algorithm;
step five: and after the circulation reaches the maximum limit times, obtaining the optimal unloading ratio, the computing resources and the subcarrier allocation expression of different computing tasks of each user node.
2. The method according to claim 1, wherein the second step specifically comprises: the system optimization model is as follows:
Figure RE-FDA0002767953680000021
Figure RE-FDA0002767953680000022
Figure RE-FDA0002767953680000023
Figure RE-FDA0002767953680000024
Figure RE-FDA0002767953680000025
Figure RE-FDA0002767953680000026
Figure RE-FDA0002767953680000027
Ekrepresenting the total energy consumption of user k, λkData occupation total data R processed by MEC server on behalf of user task loadkProportion of (a) tk，lTime of local calculation on behalf of the user, tk，offTime of use, f, on behalf of MEC server load computationk，mRepresenting the calculation frequency assigned by the MEC server to user k, F representing the available total CPU calculation frequency of the MEC server, T representing a time slot length, xk，nRepresenting whether subcarrier n is allocated to user k, is a binary variable;
wherein
Figure RE-FDA0002767953680000028
Time for load calculation tk，offThe method comprises two parts, namely the uplink transmission time of data from a user k to an MEC server, and the time for the MEC server to process the data; wherein R iskRepresents the total data to be processed; r iskRepresenting the transmission rate of the MEC server to obtain the user load data:
Figure RE-FDA0002767953680000029
3. the method according to claim 1, wherein the third step is specifically: decoupling the problem P into sub-problems Po and Ps, and further performing decoupling calculation on k users respectively to obtain an optimal unloading ratio and calculation resources of the corresponding user k and a subcarrier allocation strategy;
solving for the optimal unload ratio for user k
Figure RE-FDA00027679536800000210
The transformation is to minimize the problem as follows:
Figure RE-FDA00027679536800000211
s.t.0≤λk≤1,
Figure RE-FDA00027679536800000212
and is provided with
Figure RE-FDA0002767953680000031
4. The method according to claim 3, wherein the third step is specifically:
the Ps problem is modeled as follows, which will be further decoupled by the following steps:
Figure RE-FDA0002767953680000032
s.t.(7d)-(7g)
Figure RE-FDA0002767953680000033
introducing a non-negative auxiliary variable
Figure RE-FDA0002767953680000034
The Ps problem was converted to Ps 1:
Figure RE-FDA0002767953680000035
s.t.(7b)-(7g),(11b)
Figure RE-FDA0002767953680000036
further, the problem Ps1 is converted into an unconstrained lagrange multiplier:
Figure RE-FDA0002767953680000037
Figure RE-FDA0002767953680000038
wherein
Figure RE-FDA0002767953680000039
Three of gamma are non-negative Langerhans and are recorded
Figure RE-FDA00027679536800000310
To satisfy the constraint
Figure RE-FDA00027679536800000311
Figure RE-FDA00027679536800000312
F, x is a set satisfying the constraint
Figure RE-FDA00027679536800000313
And is
Figure RE-FDA00027679536800000314
The set of x of (a); then the lagrange dual function is defined as follows:
Figure RE-FDA00027679536800000315
namely the Lagrangian dual problem to be solved is
maxg(α，β，γ)
s.t.α≥0，β≥0，γ≥0.
The Ps problem decoupling is completed.
5. The method according to claim 4, wherein the third step is specifically: solving the optimal allocation strategy of the computing resources:
given that the BCD algorithm KKT condition is satisfied, the optimal solution for computing resource allocation that is easily obtained satisfies the following sufficient requirements:
Figure RE-FDA0002767953680000041
wherein
Figure RE-FDA0002767953680000042
Is a convex function, and
Figure RE-FDA0002767953680000043
with fk，mMonotonically increasing and having f greater than or equal to 0k，m≤F。
6. The method according to claim 5, wherein the third step is specifically:
the following dichotomy thought is adopted to solve the optimal computing resource allocation strategy
Figure RE-FDA0002767953680000044
The upper limit value of the calculation frequency for initializing the user k to be distributed is
Figure RE-FDA0002767953680000045
Lower limit value of
Figure RE-FDA0002767953680000046
And a constant e for control accuracy3；
The following procedure was repeated:
order to
Figure RE-FDA0002767953680000047
Based on the above sufficient requirements, calculating
Figure RE-FDA0002767953680000048
If it is not
Figure RE-FDA0002767953680000049
Then order
Figure RE-FDA00027679536800000410
Otherwise make
Figure RE-FDA00027679536800000411
Repeating the above process until
Figure RE-FDA00027679536800000412
Obtaining optimal computing resource allocation strategy after circulation is finished
Figure RE-FDA00027679536800000413
7. The method according to claim 6, wherein the third step is specifically:
solving the subcarrier allocation strategy is converted into solving by the following Lagrange function:
Figure RE-FDA0002767953680000051
wherein,
Figure RE-FDA0002767953680000052
assuming that subcarrier n is allocated to user k, we get:
Figure RE-FDA0002767953680000053
wherein,
Figure RE-FDA0002767953680000054
Figure RE-FDA0002767953680000055
wherein
Figure RE-FDA0002767953680000056
To minimize each
Figure RE-FDA0002767953680000057
Nth subcarrier xnThe optimal solution of whether to assign to user k is expressed as follows:
Figure RE-FDA0002767953680000058
thus, the optimal subcarrier set currently allocated to the user k is obtained.
8. The method according to claim 7, wherein the third step is specifically:
updating a related calculation parameter phi:
first from the { f obtained★，X★H, update pk，nLet us order
Figure RE-FDA0002767953680000059
The φ optimization problem is modeled as Ps 2:
Figure RE-FDA00027679536800000510
Figure RE-FDA0002767953680000061
Figure RE-FDA0002767953680000062
wherein,
Figure RE-FDA0002767953680000063
indicating the selection of the set of subcarriers allocated to user k.
Figure RE-FDA0002767953680000064
9. The method according to claim 8, wherein the third step is specifically: for k users, Ps2 is decoupled into Ps 3:
Figure RE-FDA0002767953680000065
Figure RE-FDA0002767953680000066
Figure RE-FDA0002767953680000067
obtaining:
Figure RE-FDA0002767953680000068
wherein,
Figure RE-FDA0002767953680000069
Figure RE-FDA00027679536800000610
wherein
Figure RE-FDA00027679536800000611
Up to lagrange function
Figure RE-FDA00027679536800000612
Converge to obtain
Figure RE-FDA00027679536800000613
10. The method according to claim 9, wherein the third step is specifically:
updating the relevant auxiliary parameters α, β, γ:
has obtained f according to the above steps★，X★，
Figure RE-FDA0002767953680000071
Solving a lagrange dual function, the dual variables (α, β, β, γ) being updated according to a gradient descent method, since the function is a convex function:
Figure RE-FDA0002767953680000072
Figure RE-FDA0002767953680000073
Figure RE-FDA0002767953680000074
wherein
Figure RE-FDA0002767953680000075
ξkθ is the corresponding step size of the dual variable in the loop;
the loop ends with the update (α, β, γ) until the three parameters converge.
CN202010975206.1A 2020-09-16 2020-09-16 Mobile edge computing network energy consumption minimization strategy method under time-sensitive condition Active CN112105062B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010975206.1A CN112105062B (en) 2020-09-16 2020-09-16 Mobile edge computing network energy consumption minimization strategy method under time-sensitive condition

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010975206.1A CN112105062B (en) 2020-09-16 2020-09-16 Mobile edge computing network energy consumption minimization strategy method under time-sensitive condition

Publications (2)

Publication Number Publication Date
CN112105062A true CN112105062A (en) 2020-12-18
CN112105062B CN112105062B (en) 2022-05-31

Family

ID=73760252

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010975206.1A Active CN112105062B (en) 2020-09-16 2020-09-16 Mobile edge computing network energy consumption minimization strategy method under time-sensitive condition

Country Status (1)

Country Link
CN (1) CN112105062B (en)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112887128A (en) * 2021-01-12 2021-06-01 西安电子科技大学 Resource allocation joint optimization method, system, storage medium and application
CN113556764A (en) * 2021-07-30 2021-10-26 云南大学 Method and system for determining calculation rate based on mobile edge calculation network
CN113727371A (en) * 2021-08-06 2021-11-30 北京科技大学 IRS (inter-Range instrumentation) assisted MEC (Multi-media communication) network wireless and computing resource allocation method and device
CN113923223A (en) * 2021-11-15 2022-01-11 安徽大学 User allocation method with low time cost in edge environment
CN114745384A (en) * 2022-04-08 2022-07-12 云南大学 Resource allocation method based on temperature sensing in mobile edge computing network
CN114845308A (en) * 2022-03-25 2022-08-02 国网安徽省电力有限公司信息通信分公司 Cross-MEC resource management method considering power multi-service dynamic requirements
CN114866557A (en) * 2022-05-06 2022-08-05 东北林业大学 Computing resource sharing platform and method under edge computing and V2V converged network
CN114885345A (en) * 2021-02-05 2022-08-09 中国科学院沈阳自动化研究所 System computing energy efficiency maximization resource allocation method based on differential evolution algorithm
CN116506877A (en) * 2023-06-26 2023-07-28 北京航空航天大学 Distributed collaborative computing method for mobile crowd sensing

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180183855A1 (en) * 2016-12-28 2018-06-28 Intel Corporation Application computation offloading for mobile edge computing
CN109413615A (en) * 2018-09-14 2019-03-01 重庆邮电大学 The energy delay compromise proposal of Energy-aware unloading under car networking based on MEC
US20190138934A1 (en) * 2018-09-07 2019-05-09 Saurav Prakash Technologies for distributing gradient descent computation in a heterogeneous multi-access edge computing (mec) networks
CN109756378A (en) * 2019-01-12 2019-05-14 大连理工大学 A kind of intelligence computation discharging method under In-vehicle networking
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
CN110418356A (en) * 2019-06-18 2019-11-05 深圳大学 A kind of calculating task discharging method, device and computer readable storage medium
CN111124666A (en) * 2019-11-25 2020-05-08 哈尔滨工业大学 Efficient and safe multi-user multi-task unloading method in mobile Internet of things
CN111262944A (en) * 2020-01-20 2020-06-09 北京大学 Method and system for hierarchical task offloading in heterogeneous mobile edge computing network

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180183855A1 (en) * 2016-12-28 2018-06-28 Intel Corporation Application computation offloading for mobile edge computing
US20190138934A1 (en) * 2018-09-07 2019-05-09 Saurav Prakash Technologies for distributing gradient descent computation in a heterogeneous multi-access edge computing (mec) networks
CN109413615A (en) * 2018-09-14 2019-03-01 重庆邮电大学 The energy delay compromise proposal of Energy-aware unloading under car networking based on MEC
CN109756378A (en) * 2019-01-12 2019-05-14 大连理工大学 A kind of intelligence computation discharging method under In-vehicle networking
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
CN110418356A (en) * 2019-06-18 2019-11-05 深圳大学 A kind of calculating task discharging method, device and computer readable storage medium
CN111124666A (en) * 2019-11-25 2020-05-08 哈尔滨工业大学 Efficient and safe multi-user multi-task unloading method in mobile Internet of things
CN111262944A (en) * 2020-01-20 2020-06-09 北京大学 Method and system for hierarchical task offloading in heterogeneous mobile edge computing network

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
SONALI CHOUHAN: "Energy Optimal Partial Computation Offloading Framework for Mobile Devices in Multi-access Edge Computing", 《2019 INTERNATIONAL CONFERENCE ON SOFTWARE, TELECOMMUNICATIONS AND COMPUTER NETWORKS》 *
吴柳青等: "基于边-端协同的任务卸载资源分配联合优化算法", 《电信科学》 *
周文晨等: "移动边缘计算中分布式的设备发射功率优化算法", 《西安交通大学学报》 *

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112887128A (en) * 2021-01-12 2021-06-01 西安电子科技大学 Resource allocation joint optimization method, system, storage medium and application
CN114885345B (en) * 2021-02-05 2023-08-18 中国科学院沈阳自动化研究所 System computing energy efficiency maximization resource allocation method based on differential evolution algorithm
CN114885345A (en) * 2021-02-05 2022-08-09 中国科学院沈阳自动化研究所 System computing energy efficiency maximization resource allocation method based on differential evolution algorithm
CN113556764A (en) * 2021-07-30 2021-10-26 云南大学 Method and system for determining calculation rate based on mobile edge calculation network
CN113556764B (en) * 2021-07-30 2022-05-31 云南大学 Method and system for determining calculation rate based on mobile edge calculation network
CN113727371A (en) * 2021-08-06 2021-11-30 北京科技大学 IRS (inter-Range instrumentation) assisted MEC (Multi-media communication) network wireless and computing resource allocation method and device
CN113923223A (en) * 2021-11-15 2022-01-11 安徽大学 User allocation method with low time cost in edge environment
CN113923223B (en) * 2021-11-15 2024-02-06 安徽大学 User allocation method with low time cost in edge environment
CN114845308A (en) * 2022-03-25 2022-08-02 国网安徽省电力有限公司信息通信分公司 Cross-MEC resource management method considering power multi-service dynamic requirements
CN114845308B (en) * 2022-03-25 2023-02-21 国网安徽省电力有限公司信息通信分公司 Cross-MEC resource management method considering power multi-service dynamic requirements
CN114745384A (en) * 2022-04-08 2022-07-12 云南大学 Resource allocation method based on temperature sensing in mobile edge computing network
CN114745384B (en) * 2022-04-08 2023-06-20 云南大学 Resource allocation method based on temperature sensing in mobile edge computing network
CN114866557B (en) * 2022-05-06 2024-01-23 东北林业大学 Computing resource sharing platform and method under edge computing and V2V fusion network
CN114866557A (en) * 2022-05-06 2022-08-05 东北林业大学 Computing resource sharing platform and method under edge computing and V2V converged network
CN116506877A (en) * 2023-06-26 2023-07-28 北京航空航天大学 Distributed collaborative computing method for mobile crowd sensing
CN116506877B (en) * 2023-06-26 2023-09-26 北京航空航天大学 Distributed collaborative computing method for mobile crowd sensing

Also Published As

Publication number Publication date
CN112105062B (en) 2022-05-31

Similar Documents

Publication Publication Date Title
CN112105062B (en) Mobile edge computing network energy consumption minimization strategy method under time-sensitive condition
WO2022121097A1 (en) Method for offloading computing task of mobile user
CN111245651B (en) Task unloading method based on power control and resource allocation
CN109413724B (en) MEC-based task unloading and resource allocation scheme
CN111132191B (en) Method for unloading, caching and resource allocation of joint tasks of mobile edge computing server
CN109151864B (en) Migration decision and resource optimal allocation method for mobile edge computing ultra-dense network
CN108924936B (en) Resource allocation method of unmanned aerial vehicle-assisted wireless charging edge computing network
CN109951869B (en) Internet of vehicles resource allocation method based on cloud and mist mixed calculation
CN111182570B (en) User association and edge computing unloading method for improving utility of operator
CN111586696A (en) Resource allocation and unloading decision method based on multi-agent architecture reinforcement learning
Shang et al. Deep learning-assisted energy-efficient task offloading in vehicular edge computing systems
CN111130911B (en) Calculation unloading method based on mobile edge calculation
CN109756912B (en) Multi-user multi-base station joint task unloading and resource allocation method
CN114189936A (en) Cooperative edge computing task unloading method based on deep reinforcement learning
CN110233755B (en) Computing resource and frequency spectrum resource allocation method for fog computing in Internet of things
Shang et al. Computation offloading and resource allocation in NOMA–MEC: A deep reinforcement learning approach
CN111565380B (en) NOMA-MEC-based hybrid unloading method in Internet of vehicles
CN113377533A (en) Dynamic computation unloading and server deployment method in unmanned aerial vehicle assisted mobile edge computation
CN115103326B (en) Internet of vehicles task unloading and resource management method and device based on alliance game
CN114554495B (en) Federal learning-oriented user scheduling and resource allocation method
CN110177383B (en) Efficiency optimization method based on task scheduling and power allocation in mobile edge calculation
CN116405979A (en) Millimeter wave mobile edge computing networking resource allocation method
CN114615705B (en) Single-user resource allocation strategy method based on 5G network
CN113556760B (en) Mobile edge computing cost benefit optimization method, system and application
Kim et al. Minimum overhead beamforming and resource allocation in d2d edge networks

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