CN109905888B - Joint optimization migration decision and resource allocation method in mobile edge calculation - Google Patents
Joint optimization migration decision and resource allocation method in mobile edge calculation Download PDFInfo
- Publication number
- CN109905888B CN109905888B CN201910215945.8A CN201910215945A CN109905888B CN 109905888 B CN109905888 B CN 109905888B CN 201910215945 A CN201910215945 A CN 201910215945A CN 109905888 B CN109905888 B CN 109905888B
- Authority
- CN
- China
- Prior art keywords
- migration
- user
- computing node
- node
- 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.)
- Active
Links
Images
Abstract
The invention discloses a method for joint optimization migration decision and resource allocation in mobile edge calculation. The optimization problem is based on power constraint, total migration constraint and transmission delay constraint, the optimization design of migration calculated amount is achieved, and migration calculation expenditure is the sum of migration transmission energy consumption and time consumption expenditure from a user to each calculation node and the energy consumption and time consumption expenditure of calculation migration data of each calculation node. And then solving the optimal transmission power on each subcarrier by Lagrange dual theorem, then solving the optimal migration data distribution to each edge server by adopting a multivariate linear programming solving method such as a penalty function method and the like, and finally performing iterative calculation until convergence to obtain the corresponding migration data amount to each calculation node and the transmission power on each subcarrier. The iterative optimization algorithm can effectively converge and improve the system performance.
Description
Technical Field
The invention belongs to the technical field of wireless communication, and relates to a joint optimization migration decision and resource allocation method in mobile edge computing.
Background
With the popularity of intelligent mobile terminals, people are increasingly expecting more compute-intensive applications to run on mobile terminals. The powerful computing power required by these computationally intensive applications, as well as the stringent requirements for latency and limited resources of the device create a huge conflict, which also becomes a bottleneck to improve the satisfaction of the user experience.
To enhance the computing power of the wireless terminal device, a mobile edge calculation is proposed. Compared to cloud computing, which migrates computing tasks to remote clouds for computing, moving edge computing can be viewed as a "closer-to-the-ground cloud". The mobile edge server is positioned at the edge of the wireless network and is closer to the user, so that the surrounding user can be effectively served. Mobile edge computing allows a mobile terminal to migrate computing tasks to nearby mobile edge servers, such as cell sites and WiFi access points. Compared to cloud computing, mobile edge computing can handle and reduce the amount of data at the edge of the network. Meanwhile, the mobile edge computing has the characteristics of low delay and position perception, and can improve the service quality for streaming media and real-time application programs. Mobile edge computing can accomplish real-time big data analysis, support dense distributed data collection points, and hold advantages in entertainment, advertising, personal computing, and other applications.
In the existing migration strategy of the mobile edge, most of the consideration is the strategy of migrating the whole task to a single edge computing node or the cooperative migration between the user and the migration node, so that the energy consumption or the time consumption is minimum. And the cooperation of a plurality of edge nodes is less considered, the service is provided for the user together, and the time consumption and the energy consumption expenditure of the user are most economical.
Disclosure of Invention
The purpose of the invention is as follows: the invention provides a low-complexity and easy-to-implement method for jointly optimizing migration decision and resource allocation in mobile edge calculation, which is used for reducing the expenditure of a user in migration calculation.
The technical scheme is as follows: in order to achieve the purpose, the invention adopts the following technical scheme:
a joint optimization migration decision and resource allocation method in mobile edge calculation comprises the following steps:
(1) performing mathematical modeling on the design of migration decision and resource allocation to obtain a corresponding optimization problem, wherein the optimization problem comprises three groups of constraint conditions of power, migration data volume and transmission delay, the power constraint is that the transmission power and the total transmission power distributed by a user on each subchannel meet a set value range, the migration data volume constraint is that the migration data volume distributed to each computing node by the user meets the set value range and the sum of the migration data volumes is equal to the total migration computing file data volume, and the transmission delay constraint limits that the transmission time of the user migration data to each computing node does not exceed the maximum transmission delay; the optimization target of the optimization problem is to minimize migration calculation expenditure and obtain corresponding migration data volume to each calculation node and the transmitting power on each subcarrier; the migration calculation expenditure is migration transmission energy consumption and time consumption expenditure from a user to each calculation node, and the sum or weighted sum of the energy consumption and the time consumption expenditure of the migration data calculated by each calculation node;
(2) fixing a group of initial values of migration data quantity which accords with the constraint condition and reaches each computing node, and then solving a new problem that the migration computing expense is minimized only by taking the transmitting power on each subcarrier as a variable; solving the distribution condition of the transmitting power on each subcarrier by the new problem, fixing the transmitting power value on each subcarrier, and forming a new problem and solving the new problem by taking the migration data volume of each computing node as a variable in the original problem.
(3) And (3) taking the optimal solution of the next optimization problem obtained in the step (2) as an iteration point, and performing iterative solution to obtain a stable solution of the original optimization problem.
As a preferred embodiment, the step (2) comprises:
(2.1) setting a group of initial values of migration data quantity of each computing node which accords with constraint conditions, converting an original joint non-convex problem into a fractional optimization problem of each subcarrier transmitting power, and converting the original joint non-convex problem into a convex problem through Lagrange dual conditions, thereby obtaining an optimal solution of each subcarrier transmitting power;
and (2.2) fixing the transmitting power value on each subcarrier, solving a multivariate linear programming problem by taking the migration data volume of each computing node as an optimization variable, and obtaining the optimal solution of the migration data volume of each computing node.
As a preferred embodiment, the values of the transmission power allocated by the user on each subchannel satisfy:
wherein p isnThe transmission power allocated to the sub-carrier n, P represents the maximum transmission power of the user, and the set of sub-carriers is recorded as
The migration data volume value allocated to each computing node by the user meets the following requirements:
wherein beta iskThe bit number of the migration file from the user to a computing node k is represented, D represents the size of the task, and the set of the computing nodes is recorded as
The transmission time value of the user migration data to each computing node meets the following conditions:
wherein R isk(pn) Meter for indicating user to moving edgeCalculating the Transmission Rate, T, of node kmaxIndicating the maximum transmission delay.
As a preferred embodiment, the energy consumption expenditure for migration transmission from the user to the computing node k is the product of the energy consumption per joule for migration from the user side and the energy consumption for migration transmission, where the energy consumption for migration transmission to the computing node k is expressed as:
where δ is the local operating power constant of the user.
As a preferred embodiment, the expense of migration transmission time consumption from the user to the computing node k is the product of the price consumed by the user for migration per second and the migration transmission time consumption, where the migration transmission time consumption to the computing node k is expressed as:
as a preferred embodiment, the expenditure of the computing node k is the product of the price per joule of energy consumption and the energy consumption for computing the migration data, wherein the energy consumption for computing the migration data by the computing node k is expressed as:
ek＝κfk 2βkω
wherein κ represents the effective capacitance coefficient, fkThe CPU calculation rotating speed of the calculation node k is shown, and omega represents the number of turns of the operation required by the CPU to calculate 1 bit.
As a preferred embodiment, the expenditure of the computing node k is the product of the price of the time consumed per second and the time consumed for computing the migration data, wherein the time consumed for computing the migration data by the computing node k is expressed as:
as a preferred embodiment, the optimization goal of the optimization problem is to minimize migration computation expenditure, and obtain corresponding migration data amount to each computing node and transmission power size on each subcarrier, which are expressed as:
wherein upsilon isE，υTRepresenting the price of energy consumption per joule and time consumption per second of user terminal migration;
representing the calculation energy consumption per joule and the time consumption price per second of the calculation node k; p is the set of subcarrier transmit powers, denoted as p ═ p1,p2,...,pN]，p*Distributing values for the optimal subcarrier transmitting power; β is a set of file sizes migrated to each node, and is expressed as β ═ β1,β2,...,βK]，β*And (4) optimally allocating the size of the file migrated to each node.
As a preferred embodiment, in step (2), the migration data amount from the user to each computing node is fixed, and the optimization problem of the transmission power on each subcarrier is solved as a fractional optimization problem, where the optimization target is expressed as:
wherein h isk(p) removing constant denominator terms from the optimization problem constructed in step (1), gk(p) is the denominator term, rho, excluding the constant value in the optimization problem constructed in step (1)kAnd (3) the multiplier constant which cannot be removed due to the determination of the initial value of the migration volume in the optimization problem constructed in the step (1).
As a preferred embodiment, in step (2), the transmission power on each subcarrier is fixed, and the optimization problem of solving the migration data amount of each computing node is a linear programming problem, where the optimization target is expressed as:
wherein the content of the first and second substances,
a multiplier constant removed for determining an initial value of the non-factorial carrier transmit power in the optimization problem constructed in step (1).
Has the advantages that: the joint optimization migration decision and resource allocation method applicable to the mobile edge computing can give full play to the cooperation of multiple computing nodes, effectively reduce the migration computation expenditure of users, and obviously reduce the migration computation expenditure of users compared with the migration data amount uniformly allocated to each computing node and the transmitting power on each subcarrier.
Drawings
Fig. 1 is a flowchart of a joint optimization migration decision and resource allocation method in mobile edge computing according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of an optimal allocation algorithm for migration data and subcarrier transmission power in the embodiment of the present invention.
FIG. 3 shows a comparison of the optimization costs before and after the embodiment of the present invention.
FIG. 4 shows the convergence rate of the algorithm of the present invention.
Detailed Description
The present invention is further illustrated by the following description in conjunction with the accompanying drawings and the specific embodiments, it is to be understood that the embodiments are illustrative only and not limiting of the scope of the invention, and that various equivalent modifications of the invention will occur to those skilled in the art upon reading the present invention and fall within the scope of the appended claims.
For an OFDM uplink mobile edge computing system with a single user and multiple computing nodes, K computing nodes with communication functions exist, and the user has an intensive computing task gammauAnd the task can be divided by migrating the task to a computing node to complete the computation. Where D represents the size of the task and ω represents the number of turns of the operation required for the CPU to calculate 1 bit. As shown in FIG. 1, the method of the present invention mainly comprises a mathematical modeling stage and optimizationAnd a problem solving stage. The method comprises the steps of firstly carrying out mathematical modeling on a migration decision and resource allocation to obtain a corresponding optimization problem, wherein the optimization problem comprises a power constraint condition, a migration data volume constraint condition and a transmission delay constraint condition. Setting an initial value of one variable according with the condition, converting the original optimization problem into a new optimization problem and solving the new optimization problem, setting the solved variable value as the initial value to obtain a new optimization problem, and solving the new optimization problem to obtain an optimal solution. And taking the optimal solution of the last optimization problem as an iteration point for solving the stable solution, and carrying out iterative solution to obtain the stable solution of the original optimization problem.
In the embodiment, 4 computing nodes around a user have 64 subcarriers in wireless communication, the number of subcarriers from the user to each base station is uniformly and randomly distributed, the channel bandwidth is 40MHz, the effective capacitance coefficient is 1e-27, 10 circles of 1-bit operation are calculated by a CPU, the task size is 500+ z 100 bits, wherein z belongs to {0, 1.., 10}, and the rotation speed of the CPU of the computing node per second is [10000,12000,12000,11000 } respectively]The noise density of the channel is-102 dB/Hz, and the energy consumption and time consumption of the user terminal are respectively 105Yuan/Joule, 0.1 Yuan/second, and the energy consumption and time consumption charge of the computing node end are respectively 1060.1 yuan/second, the maximum transmitting power of the user is 1W, and the maximum transmission delay is the minimum delay from the full transmitting power of the full subcarriers of the user to a certain computing node and is expressed as
I.e., D500 + z 100, z ∈ {0, 1.., 10}, N ═ 64, K ═ 4, K ═ 1e-27, and f ═ 10000,12000,12000,11000 [ D10000,12000,12000,11000 ]]，ω＝10，B＝40MHz，P＝1W，σ2＝-102dB/Hz，υE＝106，υT＝0.1，
The following describes in detail a joint optimization migration decision and resource allocation method in mobile edge computing according to the present invention with reference to a specific system model, which mainly includes the following steps:
step (1): and carrying out mathematical modeling on the migration decision and the resource allocation to obtain a corresponding optimization problem.
Record the subcarrier set as
Let the transmission power of the nth subcarrier be pn. The set of compute nodes is written as
Let the k-th compute node be assigned a computation amount of βk. In general, it is desirable to design the subcarrier transmit power and the amount of migration to the respective computing node such that the computational effort is as small as possible, with a design transmit power p ═ p1,p2,...,pN]And the calculated distribution amount beta ═ beta1,β2,...,βK]The basic idea is as follows:
the range of the transmission power, the migration amount and the transmission delay is restricted, and the expenditure of the calculation task is minimized. The total calculation expenditure of the system migration is the sum or weighted sum of the energy consumption and the time consumption expenditure of the migration transmission from the user to each calculation node and the calculation of the migration data by each calculation node. In this example, the user's expenditure is expressed as
Wherein Ek，TkTransferring energy consumption and transferring time for transferring the user to the computing node k; upsilon isE，υTRepresenting the price of energy consumed per joule and time consumed per second for the client migration.
The expenditure of a compute node is expressed as
Wherein ek，tkTo calculateCalculating the energy consumption and the time consumption of migration data of the node k;
representing the computational energy consumption per joule and the time consumption price per second of the computing node k.
The expense of the migration calculation is determined by the total expense of the system, the total expense is the sum of the expense of the user and the expense of each calculation node and is expressed as
Mathematically, designing the transmit power p and the allocation computation β can be modeled as an optimization problem as follows:
the optimization problem modeled in this step includes three sets of constraint conditions, the first set of corresponding power constraint conditions, that is, the value range of the transmission power distributed by the user on each sub-channel, which can be expressed mathematically as
The second group of corresponding migration data volume allocation constraint conditions, that is, the migration data volume value range allocated to each computing node by the user, can be expressed as mathematically
The third group of corresponding transmission delay constraints, i.e. the value ranges of the transmission delays from the users to the respective computing nodes, can be expressed mathematically as
Wherein R isk(pn) Representing the transmission rate of the user to the mobile edge computing node k,
wherein vk,nAn indication of whether a subcarrier n is allocated to a compute node k, and vk,n∈{0,1}。vk,n1, denotes that subcarrier n is allocated to compute node k; if v isk,n0, meaning that subcarrier n is not allocated to compute node k; r isk,n(pn) Represents the maximum transmission rate per second (bps) that the computing node k can achieve on subcarrier n, and is denoted as
Wherein
hk,nRadio access channel factor, p, on subcarrier n for user and computing node knThe allocated transmission power on subcarrier n.
The optimization goal of the optimization problem is to minimize the migration calculation expenditure and obtain the corresponding migration data amount p to each calculation node and the transmission power β on each subcarrier.
Step (2): randomly setting a group of initial values of migration quantities to each computing node which meet constraint conditions, and then solving a new problem of minimized migration computing expenditure which only takes the transmitting power on each subcarrier as a variable; solving the distribution condition of the transmitting power by the new problem, fixing the transmitting power value on each subcarrier, and forming a new problem and solving the new problem by taking the migration quantity of each computing node as a variable in the original problem.
As shown in fig. 2, step (2) in this embodiment is mainly divided into 2 sub-steps.
Step (2.1): setting initial values of a group of optimization variables (migration data quantity of each computing node), converting the original joint non-convex problem (1) into a fractional optimization problem of another group of optimization variables (emission power of each subcarrier), and converting the fractional optimization problem into a convex problem through Lagrange dual conditions, thereby obtaining an optimal solution of the emission power of each subcarrier.
Step (2.2): and (3) setting the transmitting power on each subcarrier obtained by calculation in the step (2.1) as a known quantity to be fixed, changing the problem (1) into a new problem containing the migration quantity to each calculation node as an optimization variable, and solving the multivariate linear programming problem by adopting a penalty function method to obtain the optimal solution of the migration data quantity of each calculation node.
The solution idea in step (2) is explained in detail based on the mathematical representation.
The first step is as follows: after a group of initial values of migration quantities to each computing node which meet constraint conditions are randomly set and are brought into a problem (1), an optimization problem is constructed as follows
Wherein the content of the first and second substances,
the denominator term of the fixed value is removed from the original optimization problem (1),
the denominator term of the fixed value is removed from the original optimization problem (1),
the multiplier constant which cannot be removed due to the determination of the initial value of the migration volume in the original optimization problem (1) is adopted. The problem (2) is introduced with an auxiliary variable χ ═ χ1,...,χKConverts the problem to problem (3).
Depending on the KKT condition, an auxiliary variable λ ═ λ is introduced1,...,λKGet problems to
To solve the problem (4), the solving steps are as follows:
step (2.1.1): let i equal 1, C(i)When p is 0, p is taken as the condition(i)Is an initial value;
step (2.1.2): calculating lambda(i)Hexix-(i)，
Step (2.1.3): at a known lambda(i)Hexix-(i)Then, calculate the problem (4), and then the objective function and the constraint of the problem (4) are both convex functions, and can use Lagrange multiplier method to obtain p(i+1)；
Step (2.1.4): if it is
Wherein, epsilon is 1e-4, p*＝p(i+1)The algorithm terminates; otherwise, i is i +1, and the procedure returns to step (1.2).
The second step is that: setting the transmitting power on each subcarrier obtained by the first step of the step as a known quantity to be brought into the problem (1), and constructing an optimization problem as follows
Wherein the content of the first and second substances,
the problem (5) is a multivariate linear programming problem, feasible solving methods are solving by lagrange multiplier method, penalty function method and other methods, the example is introduced by taking the penalty function method as an example, and the solving steps are as follows:
step (2.2.1): the constraint is changed to a penalty term for the objective function. The penalty term given by the equality constraint of problem (5) is ej +(β)＝(H1(β)-N)2J is 1; for the inequality constraint of problem (5), the step function is first defined as:
then the inequality constraint of problem (5) yields a penalty term of:
in conjunction with the penalty term, the problem (5) becomes:
wherein M ism＞0，M0＜M1＜...＜MmAnd is provided with
According to the calculation experience, M can be takenm+1＝cMm,c∈[2,50]。
Step (2.2.2): let q be 1 and M be 0, and select M0Greater than 0, c is greater than or equal to 2, allowable error epsilon and initial point beta[0]；
Step (2.2.3): by beta[q-1]As a starting point, solving the unconstrained optimization problem (6) by adopting a Newton method and obtaining the optimal solution of the unconstrained optimization problem as beta[q]；
Step (2.2.4): computing
Step (2.2.5): if mu < epsilon, output beta(l)＝β[q]The algorithm terminates; otherwise Mm+1＝cMmM +1, q +1, skip step (2.2.3).
Wherein, the Newton method of the step (2.2.3) comprises the following steps:
step (2.2.3.1): selecting an initial point estimate beta{0}Determining an allowable error epsilon, and enabling f to be 0;
step (2.2.3.2): calculating the objective function at beta{f}Gradient of lambda (. beta.) of{f})；
Step (2.2.3.3): if | | | Λ (β){f}) If | is less than or equal to epsilon, then beta*＝β{f}The algorithm terminates; otherwise, continuing to execute the step (2.2.3.4);
step (2.2.3.4): constructing the Newton Direction D(f)＝-H-1(β{f})Λ(β{f}) In which H is-1(β{f}) Expressing the objective function at beta{f}The inverse of the Hessian matrix of (a); and updates the dot column beta{f+1}＝β{f}+D(f)Let f be f +1, go to step 2.2.3.2.
And (3): and taking the optimal solution of the next optimization problem obtained in the second step as an iteration point, and performing iteration solution to obtain a stable solution of the original optimization problem.
In this embodiment, the step (3) mainly includes 2 sub-steps as follows:
step (3.1): setting the allowable error e' to 1e-6, l to 0, from p obtained in step (2)*And beta*Computing
Step (3.2): if l is 0, l is l +1, skipping to step (3.1); whether or notThen, if | C(l)-C(l-1)If | < epsilon', step (3) ends, otherwise, if l ═ l +1, step (3.1) is skipped.
As shown in fig. 3, in order to verify the effect of the method of the present invention, the present invention uses the above method to perform the comparison between the migration amount on each computing node and the transmission power distribution on each subcarrier and the uniform distribution of the migration amount on each computing node and the transmission power on each subcarrier. Taking the user calculation task size as 1500 bits as an example, the optimized time consumption and energy consumption expenditure are respectively 98% and 1.9% of the average resource allocation.
The convergence rate of the algorithm is shown in fig. 4, and taking the calculation task size of 500 bits as an example, the method can converge faster.
The above description is only of the preferred embodiments of the present invention, and it should be noted that: it will be apparent to those skilled in the art that several contemplated modifications and adaptations can be made without departing from the principles of the invention and these are intended to be included within the scope of the invention.
Claims (7)
1. The joint optimization migration decision and resource allocation method in the mobile edge calculation is characterized by comprising the following steps:
(1) performing mathematical modeling on the design of migration decision and resource allocation to obtain a corresponding optimization problem, wherein the optimization problem comprises three groups of constraint conditions of power, migration data volume and transmission delay, the power constraint is that the transmission power and the total transmission power distributed by a user on each subchannel meet a set value range, the migration data volume constraint is that the migration data volume distributed to each computing node by the user meets the set value range and the sum of the migration data volumes is equal to the total migration computing file data volume, and the transmission delay constraint limits that the transmission time of the user migration data to each computing node does not exceed the maximum transmission delay; the optimization target of the optimization problem is to minimize migration calculation expenditure and obtain corresponding migration data volume to each calculation node and the transmitting power on each subcarrier; the migration calculation expenditure is migration transmission energy consumption and time consumption expenditure from a user to each calculation node, and the sum or weighted sum of the energy consumption and the time consumption expenditure of the migration data calculated by each calculation node;
(2) fixing a group of initial values of migration data quantity which accords with the constraint condition and reaches each computing node, and then solving a new problem that the migration computing expense is minimized only by taking the transmitting power on each subcarrier as a variable; solving the distribution condition of the transmitting power on each subcarrier by the new problem, fixing the transmitting power value on each subcarrier, and forming a new problem and solving the new problem by taking the migration data volume of each computing node as a variable in the original problem; the method comprises the following steps: (2.1) setting a group of initial values of migration data quantity of each computing node which accords with constraint conditions, converting an original joint non-convex problem into a fractional optimization problem of each subcarrier transmitting power, and converting the original joint non-convex problem into a convex problem through Lagrange dual conditions, thereby obtaining an optimal solution of each subcarrier transmitting power;
(2.2) fixing the transmitting power value on each subcarrier, taking the migration data volume of each computing node as an optimization variable, and solving a multivariate linear programming problem to obtain an optimal solution of the migration data volume of each computing node;
the method comprises the following steps of fixing migration data volume of users to each computing node, solving a fractional optimization problem of transmitting power on each subcarrier, and expressing an optimization target as follows:
wherein p ═ p1,p2,...,pN]For transmitting sets of power, p, for subcarriersnFor the transmission power allocated on subcarrier N, N is in the range of {1k(p) removing constant denominator terms from the optimization problem constructed in step (1), gk(p) is the denominator term, rho, excluding the constant value in the optimization problem constructed in step (1)kA multiplier constant which cannot be removed due to the determination of the initial value of the migration quantity in the optimization problem constructed in the step (1) is set, N is the number of subcarriers, and K is the number of calculation nodes;
wherein the transmission power on each subcarrier is fixed, the linear programming problem of the migration data volume of each computing node is solved, and the optimization target is expressed as:
wherein β ═ β1,β2,...,βK]Set of file sizes, β, for migration to respective nodesiFor the number of bits of the migration file from the user to the computing node i, K belongs to { 1., K },
determining and removing a multiplier constant for an initial value of the transmission power of the carrier which cannot be factored in the optimization problem constructed in the step (1), wherein K is the number of the calculation nodes;
(3) and (3) taking the optimal solution of the next optimization problem obtained in the step (2) as an iteration point, and performing iterative solution to obtain a stable solution of the original optimization problem.
2. The method for joint optimization migration decision and resource allocation in mobile edge computing according to claim 1, wherein the values of the transmission power allocated by the user on each subchannel satisfy:
wherein p isnThe transmission power allocated to the sub-carrier n, P represents the maximum transmission power of the user, and the set of sub-carriers is recorded as
The migration data volume value allocated to each computing node by the user meets the following requirements:
wherein beta iskThe bit number of the migration file from the user to a computing node k is represented, D represents the size of the task, and the set of the computing nodes is recorded as
The transmission time value of the user migration data to each computing node meets the following conditions:
wherein R isk(pn) Representing the transmission rate, T, of a user to a mobile edge computing node kmaxIndicating the maximum transmission delay.
3. The method according to claim 1, wherein the expenditure of energy consumption for migration transmission from the user to the computing node k is a product of a price per joule of energy consumption for migration from the user end and energy consumption for migration transmission, wherein the energy consumption for migration transmission to the computing node k is expressed as:
where δ is the local operating power constant of the user, βkNumber of bits of migration file, p, for user to compute node knFor transmission power allocated on subcarrier n, Rk(pn) And the transmission rate from the user to a mobile edge computing node K is shown, wherein N is the number of subcarriers, and K is the number of computing nodes.
4. The method for joint optimization of migration decision and resource allocation in mobile edge computing according to claim 1, wherein the expenditure of time for migration transmission from the user to the computing node k is a product of a price consumed by the user for migration per second and the time consumed for migration transmission, wherein the time consumed for migration transmission to the computing node k is expressed as:
wherein, betakFor the number of file bits, R, migrated from user to compute node kk(pn) Representing the transmission rate of the user to the mobile edge computing node k.
5. The method for joint optimization of migration decision and resource allocation in mobile edge computing according to claim 1, wherein the expenditure of the computing node k is the product of the price per joule of energy consumption and the energy consumption of the computing migration data, wherein the energy consumption of computing node k for computing the migration data is expressed as:
ek＝κfk 2βkω
wherein κ represents the effective capacitance coefficient, fkCPU calculated speed, beta, representing calculation node kkAnd the bit number of the migration file from the user to the computing node k is shown as omega, and the number of turns of the operation required by the CPU to compute 1 bit is shown.
6. The method for joint optimization of migration decision and resource allocation in mobile edge computing according to claim 1, wherein the expenditure of the computing node k is the product of the price of time consumption per second and the time consumption for computing the migration data, wherein the time consumption for computing the migration data by the computing node k is expressed as:
wherein f iskCPU calculated speed, beta, representing calculation node kkAnd the bit number of the migration file from the user to the computing node k is shown as omega, and the number of turns of the operation required by the CPU to compute 1 bit is shown.
7. The method for joint optimization migration decision and resource allocation in mobile edge computing according to claim 1, wherein the optimization objective of the optimization problem is to minimize migration computation expenditure and obtain corresponding migration data amount to each computing node and transmission power size on each subcarrier, which are expressed as:
wherein p isnFor the transmission power allocated to the subcarrier n, k denotes the effective capacitance coefficient, fkCPU calculated speed, beta, representing calculation node kkThe bit number of the migration file from the user to the computing node k is shown, and omega represents the number of turns of 1 bit required to be operated by the CPU; ek，TkTransferring energy consumption and transfer time upsilon for transferring from user to computing node kE，υTRepresenting the price of energy consumption per joule and time consumption per second of user terminal migration; e.g. of the typek，tkTo compute the power and time consumption of migrating data for node k,
representing the calculation energy consumption per joule and the time consumption price per second of the calculation node k; p is the set of subcarrier transmit powers, denoted as p ═ p1,p2,...,pN]，p*Distributing values for the optimal subcarrier transmitting power; β is a set of file sizes migrated to each node, and is expressed as β ═ β1,β2,...,βK]，β*And N is the number of subcarriers, and K is the number of calculation nodes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910215945.8A CN109905888B (en) | 2019-03-21 | 2019-03-21 | Joint optimization migration decision and resource allocation method in mobile edge calculation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910215945.8A CN109905888B (en) | 2019-03-21 | 2019-03-21 | Joint optimization migration decision and resource allocation method in mobile edge calculation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109905888A CN109905888A (en) | 2019-06-18 |
| CN109905888B true CN109905888B (en) | 2021-09-07 |
Family
ID=66952829
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910215945.8A Active CN109905888B (en) | 2019-03-21 | 2019-03-21 | Joint optimization migration decision and resource allocation method in mobile edge calculation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109905888B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110535936B (en) * | 2019-08-27 | 2022-04-26 | 南京邮电大学 | Energy efficient fog computing migration method based on deep learning |
| CN111338760B (en) * | 2020-02-27 | 2023-04-25 | 长沙市源本信息科技有限公司 | Service instance cross-node telescoping method and device for edge computing |
| CN111491332B (en) * | 2020-04-20 | 2021-08-27 | 中国石油大学(北京) | Dynamic service migration method and device |
| CN111796880B (en) * | 2020-07-01 | 2021-06-04 | 电子科技大学 | Unloading scheduling method for edge cloud computing task |
| CN112187872B (en) * | 2020-09-08 | 2021-07-30 | 重庆大学 | Content caching and user association optimization method under mobile edge computing network |
| CN112148380B (en) * | 2020-09-16 | 2022-04-12 | 鹏城实验室 | Resource optimization method in mobile edge computing task unloading and electronic equipment |
| CN112118312B (en) * | 2020-09-17 | 2021-08-17 | 浙江大学 | Network burst load evacuation method facing edge server |
| CN112291335B (en) * | 2020-10-27 | 2021-11-02 | 上海交通大学 | Optimized task scheduling method in mobile edge calculation |
| CN114581160B (en) * | 2022-05-05 | 2022-09-02 | 支付宝(杭州)信息技术有限公司 | Resource allocation method, distributed computing system and equipment |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017176329A1 (en) * | 2016-04-05 | 2017-10-12 | Intel IP Corporation | Devices and methods for mec v2x |
| CN107249218A (en) * | 2017-06-05 | 2017-10-13 | 东南大学 | Radio Resource and the combined distributing method of cloud resource in a kind of MEC |
| CN107333267A (en) * | 2017-06-23 | 2017-11-07 | 电子科技大学 | A kind of edge calculations method for 5G super-intensive networking scenes |
| CN107819840A (en) * | 2017-10-31 | 2018-03-20 | 北京邮电大学 | Distributed mobile edge calculations discharging method in the super-intensive network architecture |
| CN107995660A (en) * | 2017-12-18 | 2018-05-04 | 重庆邮电大学 | Support Joint Task scheduling and the resource allocation methods of D2D- Edge Servers unloading |
| CN108540406A (en) * | 2018-07-13 | 2018-09-14 | 大连理工大学 | A kind of network discharging method based on mixing cloud computing |
| WO2018165934A1 (en) * | 2017-03-16 | 2018-09-20 | Intel Corporation | Systems, methods and devices for user plane traffic forwarding |
| CN109151864A (en) * | 2018-09-18 | 2019-01-04 | 贵州电网有限责任公司 | A kind of migration decision and resource optimal distribution method towards mobile edge calculations super-intensive network |
| CN109413724A (en) * | 2018-10-11 | 2019-03-01 | 重庆邮电大学 | A kind of task unloading and Resource Allocation Formula based on MEC |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170272365A1 (en) * | 2016-03-15 | 2017-09-21 | Hon Hai Precision Industry Co., Ltd | Method and appratus for controlling network traffic |
| US10440096B2 (en) * | 2016-12-28 | 2019-10-08 | Intel IP Corporation | Application computation offloading for mobile edge computing |
| CN108574728B (en) * | 2017-03-08 | 2021-05-04 | 中兴通讯股份有限公司 | Method and apparatus for traffic path change detection for moving edge computation |
-
2019
- 2019-03-21 CN CN201910215945.8A patent/CN109905888B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017176329A1 (en) * | 2016-04-05 | 2017-10-12 | Intel IP Corporation | Devices and methods for mec v2x |
| WO2018165934A1 (en) * | 2017-03-16 | 2018-09-20 | Intel Corporation | Systems, methods and devices for user plane traffic forwarding |
| CN107249218A (en) * | 2017-06-05 | 2017-10-13 | 东南大学 | Radio Resource and the combined distributing method of cloud resource in a kind of MEC |
| CN107333267A (en) * | 2017-06-23 | 2017-11-07 | 电子科技大学 | A kind of edge calculations method for 5G super-intensive networking scenes |
| CN107819840A (en) * | 2017-10-31 | 2018-03-20 | 北京邮电大学 | Distributed mobile edge calculations discharging method in the super-intensive network architecture |
| CN107995660A (en) * | 2017-12-18 | 2018-05-04 | 重庆邮电大学 | Support Joint Task scheduling and the resource allocation methods of D2D- Edge Servers unloading |
| CN108540406A (en) * | 2018-07-13 | 2018-09-14 | 大连理工大学 | A kind of network discharging method based on mixing cloud computing |
| CN109151864A (en) * | 2018-09-18 | 2019-01-04 | 贵州电网有限责任公司 | A kind of migration decision and resource optimal distribution method towards mobile edge calculations super-intensive network |
| CN109413724A (en) * | 2018-10-11 | 2019-03-01 | 重庆邮电大学 | A kind of task unloading and Resource Allocation Formula based on MEC |
Non-Patent Citations (3)
| Title |
|---|
| Siemens, Nokia, Huawei, ETRI.Editorial cleanup to clause 5 in TR 22.804.《3GPP TSG-SA WG1 Meeting #82 S1-181474》.2018, * |
| Yuyi Mao;Jun Zhang;Khaled B. Letaief.Joint Task Offloading Scheduling and Transmit Power Allocation for Mobile-Edge Computing Systems.《2017 IEEE Wireless Communications and Networking Conference (WCNC)》.2017, * |
| 超密集网络中基于移动边缘计算的卸载策略研究;郭俊;《中国优秀硕士学位论文库》;20181115;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109905888A (en) | 2019-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109905888B (en) | Joint optimization migration decision and resource allocation method in mobile edge calculation | |
| CN109684075B (en) | Method for unloading computing tasks based on edge computing and cloud computing cooperation | |
| CN110392079B (en) | Node computing task scheduling method and device for fog computing | |
| CN111372314A (en) | Task unloading method and task unloading device based on mobile edge computing scene | |
| CN107333267A (en) | A kind of edge calculations method for 5G super-intensive networking scenes | |
| CN113504999B (en) | Scheduling and resource allocation method for high-performance hierarchical federal edge learning | |
| CN110351760A (en) | A kind of mobile edge calculations system dynamic task unloading and resource allocation methods | |
| CN111182495B (en) | 5G internet of vehicles partial calculation unloading method | |
| CN110096362A (en) | A kind of multitask discharging method based on Edge Server cooperation | |
| CN115802389B (en) | Federal learning method trained by digital twin auxiliary model | |
| CN112860429A (en) | Cost-efficiency optimization system and method for task unloading in mobile edge computing system | |
| CN113115459B (en) | Multi-scale and multi-dimensional resource allocation method for power Internet of things mass terminal | |
| CN109548161A (en) | A kind of method, apparatus and terminal device of wireless resource scheduling | |
| CN106792919B (en) | Equivalent capacity-based resource allocation method in wireless virtual network | |
| CN111328146B (en) | Service scheduling method for optimizing transmission rate weight based on genetic algorithm | |
| CN113286317A (en) | Task scheduling method based on wireless energy supply edge network | |
| Chehri et al. | Real‐time multiuser scheduling based on end‐user requirement using big data analytics | |
| CN111935825A (en) | Depth value network-based cooperative resource allocation method in mobile edge computing system | |
| CN114885420A (en) | User grouping and resource allocation method and device in NOMA-MEC system | |
| CN114554495A (en) | Federal learning-oriented user scheduling and resource allocation method | |
| Wu et al. | Deep reinforcement learning-based video quality selection and radio bearer control for mobile edge computing supported short video applications | |
| Sundan et al. | Energy and spectral efficiency optimization using probabilistic based spectrum slicing (PBSS) in different zones of 5G wireless communication network | |
| CN106304307B (en) | A kind of resource allocation methods under heterogeneous network converged | |
| CN111542091A (en) | Wireless and computing resource joint allocation method for network slice | |
| CN113271221B (en) | Network capacity opening method and system and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |