CN110162894A - With the energy efficient mist computation migration method of delay guaranteed in industrial scenes of internet of things - Google Patents
With the energy efficient mist computation migration method of delay guaranteed in industrial scenes of internet of things Download PDFInfo
- Publication number
- CN110162894A CN110162894A CN201910445848.8A CN201910445848A CN110162894A CN 110162894 A CN110162894 A CN 110162894A CN 201910445848 A CN201910445848 A CN 201910445848A CN 110162894 A CN110162894 A CN 110162894A
- Authority
- CN
- China
- Prior art keywords
- energy consumption
- mist
- variable
- function
- equation
- 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
Links
Classifications
-
- 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
- G06F17/15—Correlation function computation including computation of convolution operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- 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
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
Abstract
The present invention relates to a kind of energy efficient mist computation migration methods with delay guaranteed in industrial scenes of internet of things.For efficient process computation-intensive task and meet harshness requirement of the industrial scenes of internet of things to delay, the present invention is first, it gives the comprehensive energy consumption for considering i.e. mist node for the structure of energy consumption of mist node and is made of the energy consumption of local computing, transmission and wait state, consider constructing the mist node energy consumption comprising energy expense and time delay constraint and minimize model based on this；Secondly, the accelerating gradient method for solving based on Duality Decomposition is proposed for solving, the method for solving can rapid solving go out optimal migration than achieving the purpose that minimize mist node energy consumption, significantly improve the convergence rate of traditional solution method, reduce the deadline of calculating task.Meanwhile the specific implementation of the method for the present invention demonstrates proposed mist computation migration method in the advantage of mist node energy consumption and calculating task deadline.
Description
Technical field
The invention belongs to cordless communication network, wireless sensor network, technical field of the computer network, and in particular to a kind of
With the energy efficient mist computation migration method of delay guaranteed in industrial scenes of internet of things.
Background technique
With the fast development of wireless sensor technology, sensing equipment will generate mass data, to this large-scale data
Being effectively treated in the case where low latency and low energy consumption is a huge challenge.
Remote cloud server is used to handle and excavate the data that a large amount of collections come, and still, it violates industrial Internet of Things
Low latency requires and leads to very big communication overhead.The shortcomings that development that mist calculates compensates for cloud computing, when closer data source
When, mist is calculated in processing locality calculating task, can be reduced communication overhead in this way and be provided at data in the case where low time delay
Reason service.Due to the limitation of energy, mist node computing capability and storage resource and the rapid growth of computation-intensive service, work
Mist node in industry Internet of Things can not handle all computation requests in time.
In view of the above-mentioned problems, computation migration technology grows up and the burden to mitigate processing mist node.Currently, about
The research work of computation migration is concentrated mainly on through optimization task immigration ratio and solves optimal delay and energy consumption problem.It is existing
Computation migration scheme be roughly divided into following three classes:
First kind scheme only considers computation migration problem, such as time delay or energy by one performance indicator of optimization
Consumption.Although the characteristics of program is to realize performance boost in terms of deadline and energy consumption cost, it can not be expected
The deadline is minimized under energy consumption or can not minimize energy consumption under expected delay.
The shortcomings that second class scheme overcomes first kind scheme is studied, and main thought is under given deferred constraint
It minimizes energy consumption or minimizes energy consumption under joint energy consumption and deferred constraint.The program improves the performance of first kind scheme,
But the combined optimization problem of deadline and energy consumption is not accounted for, thus the significant performance for improving whole system.
Third class scheme passes through while optimizing two or more performance indicators.But the solution side of such scheme optimization problem
Method is mainly based upon exhaustive search and traditional alternative manner, and the convergence rate for finding optimal value is slower, it is difficult to meet industrial object
Stringent delay requirement in networking scenario.
Summary of the invention
In view of the above technical problems, the present invention proposes the energy efficient with delay guaranteed in a kind of industrial scenes of internet of things
Mist computation migration method, it is therefore intended that the component part for comprehensively considering mist node energy consumption, the energy consumption including local computing pass
Defeated and wait state minimizes the energy consumption of mist node on the basis of guaranteeing tolerable delay and energy consumption.
With the energy efficient mist computation migration method of delay guaranteed in industrial scenes of internet of things, include the following steps:
Step 1, the structure of energy consumption for mist node give the comprehensive energy consumption for considering i.e. mist node by local computing,
The energy consumption of transmission and wait state forms, and the mist node energy constructed comprising energy expense and time delay constraint is considered based on this
Amount consumption minimizes model；
Step 2, the constraint condition for minimizing model is obtained；
Step 3, the constraint condition according to obtained in step 2 solves the minimum model using Dual Method, building
Lagrangian, and original optimization problem and corresponding dual problem are constructed, define optimized variable therein and dual variable；
Step 4, the Lagrangian that definition step 3 proposes corresponds to the gradient function of variable, by derived function and instead
Function obtains the calculation formula of optimized variable；
Step 5, it proposes a Speed gradient algorithm, defines temporary variable, and dual variable is carried out by temporary variable
It updates, solves the optimal solution for obtaining dual problem in step 3, then gone out according to the optimized variable formula proving in step 4
The optimal solution of optimization problem in step 3, the optimal solution of the optimization problem are to minimize the optimal solution of model.
Further, in the step 1, it is assumed that mist node number n=10 has 1 Cloud Server, the meter of these mist nodes
Calculation ability is uniformly set as fi l=2*106Cycles/sec, i ∈ { 1,2 ..., 10 }, and the computing capability of Cloud Server is fc=2*108
Cycles/sec；In a communication environment, wireless signal-path band width B=50Mb/s, the Carrier To Noise Power Density N of channel are defined0=10-10W with
And the wireless channel gain h between mist node and Cloud Serveri, andThe calculating task size w of different mist node isiPhase
Together, their size is between 100Kb to 1000Kb；Complete revolving speed c required for cpu when 1 bit calculating taskiEqual to 1000
Period/bit；The expectation energy consumption E of each mist nodeiFor 1.5J, it is expected that the T that is delayediFor 600ms；Power pi,cIt is 10-3W, energy disappear
Consume miIt is 10-8Joule/the period；The message transmission rate of mist node i is defined as Ri；Symbol aiIndicate the data of progress computation migration
The ratio of the total task data amount of Zhan is measured, and it can be expressed as di=wiai.Obviously, aiMeet 0≤a of conditioni≤1。
Energy consumption minimized model is as follows under tolerable time delay:
Further, in the step 2, the constraint condition, specifically:
First constraint condition is, the deadline of task no more than the expected time, when the deadline includes processing locality
Between, transmission time and Cloud Server handle the time；
Second constraint condition is to need to meet the energy consumption of mist node no more than desired energy consumption, energy consumption packet
Include local computing, transmission and waiting energy consumption；
Third constraint condition is channel width；
4th constraint condition is computing resource limitation.
Further, in the step 3, specifically, Lagrangian letter is constructed by first and second constraint condition
Number, is expressed as follows:
Wherein λiAnd μiIt is Lagrange multiplier, they are associated with task completion time and energy consumption respectively；
Secondly, the optimization problem for constructing following format is expressed as follows:
S.t.0≤a≤1, t > 0. (3)
To solve the above-mentioned problems, it is as follows to establish corresponding dual problem:
s.t.λ≥0,μ≥0. (4)
Further, in the step 4, the dual problem according to step 3, definition LagrangianL (a, t,
λ, μ) about variable ai, ti, λiAnd μiGradient function it is as follows:
According to equation (6), function β=y (x)-xy'(x is defined), wherein y'(x)=N02x/BLn2 is function y (x)
Derived function is obtained, the inverse function that the variable x in beta function is then defined according to lambert's function is as follows:
Wherein W0(x) be lambert's function W Main Branches, be expressed asSymbol x and e are natures pair
Several radixes；
When equation (6) is set to 0, beta function can be rewritten as:
According to equation (9), available following equation:
Equation (10), (11), which are combined, to be rewritten as following form for equation (11):
When equation (7) is set to 0, available aiCalculated result are as follows:
Therefore, according to equation (12), (13), tiIt can be calculated as follows: ti=wiai/Ri。
Further, in the step 5, the dual problem according to step 3, propose a Speed gradient algorithm with
Faster convergence rate solves the optimal value of dual variable, the temporary variable being defined as follows:
As k >=1, temporary variable εi(r) and δi(r) itself is updated using the information of first two steps.In equation (15),
(16) based on, dual variable λiAnd μiIt is updated according to following:
Finally, solving the optimal solution for obtaining dual problem (4)Then it according to equation (12), (13), derives excellent
The optimal solution of change problem (3)
Compared to the prior art, low energy consumption by the present invention, achieves significant energy-saving effect；Since deferred constraint being integrated into
In Optimized model, the computing capability of Cloud Server is rationally utilized, there is very big advantage on the deadline.In task size
In identical situation, heretofore described Speed gradient algorithm converges to the speed of stationary value faster than traditional gradient method.More
Precisely, speedup gradient method will converge to optimal value in 100 iteration, and traditional gradient method the number of iterations is 200 times.Tradition
Gradient method is only iterated using back information, and accelerating gradient rule is iterated using first two steps information.Therefore, accelerate
Convergence speed of the algorithm is significantly improved.
The method that the present invention designs makes energy minimizing overhead under the premise of guaranteeing delay, design method of the present invention into
The optimization of migration decision is gone, and during solving optimal value, the convergence rate of accelerating algorithm is faster than traditional algorithm.Finally,
Numerical result demonstrate proposed migration scheme energy consumption, the deadline and in terms of superior function.
Detailed description of the invention
Fig. 1 is the step flow chart that optimization problem is solved using Speed gradient algorithm.
Fig. 2 is influence of the calculating task size to mist node average energy consumption.
Fig. 3 is influence of the calculating task size to the mist node deadline.
When Fig. 4 is using traditional gradient method solution formulaization migration optimization problem, the evolution of mist node average mobility ratio
Journey.
Fig. 5 is that different method for solving mist the convergence effect of node average mobility ratio.
Specific embodiment
Technical solution of the present invention is described in further detail with reference to the accompanying drawings of the specification.
With the energy efficient mist computation migration method of delay guaranteed in industrial scenes of internet of things, include the following steps:
Step 1, the structure of energy consumption for mist node give the comprehensive energy consumption for considering i.e. mist node by local computing,
The energy consumption of transmission and wait state forms, and the mist node energy constructed comprising energy expense and time delay constraint is considered based on this
Amount consumption minimizes model.
Specifically, it is assumed that mist node number n=10 has 1 Cloud Server, and the computing capability of these mist nodes is uniformly set as
fi l=2*106Cycles/sec, i ∈ { 1,2 ..., 10 }, and the computing capability of Cloud Server is fc=2*108Cycles/sec；It is communicating
In environment, wireless signal-path band width B=50Mb/s, the Carrier To Noise Power Density N of channel are defined0=10-10W and mist node and cloud clothes
Wireless channel gain h between business devicei, andThe calculating task size w of different mist node isiIdentical, their size exists
Between 100Kb to 1000Kb；Complete revolving speed c required for cpu when 1 bit calculating taskiEqual to 1000 periods/bit；Each
The expectation energy consumption E of mist nodeiFor 1.5J, it is expected that the T that is delayediFor 600ms；Power pi,cIt is 10-3W, energy consumption miIt is 10-8Joule/
Period；The message transmission rate of mist node i is defined as Ri；Symbol aiIndicate the total number of tasks of the data volume Zhan for carrying out computation migration
According to the ratio of amount, and it can be expressed as di=wiai.Obviously, aiMeet 0≤a of conditioni≤1。
Energy consumption minimized model is as follows under tolerable time delay:
Step 2, the constraint condition for minimizing model is obtained.
The constraint condition, specifically:
First constraint condition is, the deadline of task no more than the expected time, when the deadline includes processing locality
Between, transmission time and Cloud Server handle the time.
Second constraint condition is to need to meet the energy consumption of mist node no more than desired energy consumption, energy consumption packet
Include local computing, transmission and waiting energy consumption.
Third constraint condition is channel width.
4th constraint condition is computing resource limitation.
Step 3, the constraint condition according to obtained in step 2 solves the minimum model using Dual Method, building
Lagrangian, and original optimization problem and corresponding dual problem are constructed, define optimized variable therein and dual variable.
Specifically, Lagrangian is constructed by first and second constraint condition, is expressed as follows:
Wherein λiAnd μiIt is Lagrange multiplier, they are associated with task completion time and energy consumption respectively.
Secondly, the optimization problem for constructing following format is expressed as follows:
S.t.0≤a≤1, t > 0. (3)
To solve the above-mentioned problems, it is as follows to establish corresponding dual problem:
s.t.λ≥0,μ≥0. (4)
Step 4, the Lagrangian that definition step 3 proposes corresponds to the gradient function of variable, by derived function and instead
Function obtains the calculation formula of optimized variable.
Specifically, the dual problem according to step 3 defines LagrangianL (a, t, λ, μ) about variable
ai, ti, λiAnd μiGradient function it is as follows:
According to equation (6), function β=y (x)-xy'(x is defined), wherein y'(x)=N02x/BLn2 is function y (x)
Derived function is obtained, the inverse function that the variable x in beta function is then defined according to lambert's function is as follows:
Wherein W0(x) be lambert's function W Main Branches, be expressed asSymbol x and e are natures pair
Several radixes.
When equation (6) is set to 0, beta function can be rewritten as:
According to equation (9), available following equation:
Equation (10), (11), which are combined, to be rewritten as following form for equation (11):
When equation (7) is set to 0, available aiCalculated result are as follows:
Therefore, according to equation (12), (13), tiIt can be calculated as follows: ti=wiai/Ri。
Step 5, it proposes a Speed gradient algorithm, defines temporary variable, and dual variable is carried out by temporary variable
It updates, solves the optimal solution for obtaining dual problem in step 3, then gone out according to the optimized variable formula proving in step 4
The optimal solution of optimization problem in step 3, the optimal solution of the optimization problem are to minimize the optimal solution of model.
Specifically, the dual problem according to step 3 proposes a Speed gradient algorithm with faster convergence rate
The optimal value of dual variable is solved, the temporary variable that is defined as follows:
As k >=1, temporary variable εi(r) and δi(r) itself is updated using the information of first two steps.In equation (15),
(16) based on, dual variable λiAnd μiIt is updated according to following:
Finally, solving the optimal solution for obtaining dual problem (4)Then it according to equation (12), (13), derives excellent
The optimal solution of change problem (3)
Figure it is seen that average energy consumption increases, moving method energy proposed by the present invention with the increase of task scale
It consumes minimum.By the solution to mist computation migration optimization problem, optimal migration ratio is obtained, significant energy-saving effect is achieved.
Ignore the availability of cloud computing resources, the average energy consumption highest of local calculation.The method all migrated improves local calculation
Energy consumption, but huge energy is consumed in transmission process.
From figure 3, it can be seen that the present invention has apparent advantage on the deadline compared with other two methods.Example
Such as, when task size is 800kb, the deadline of the invention is 0.244s, compared with local calculation and full unloading, is dropped respectively
Low 39.3% and 19.2%.Since deferred constraint to be integrated into Optimized model, the calculating energy of Cloud Server is rationally utilized
Power, it is proposed that scheme on the deadline have very big advantage.
In conjunction with Fig. 2 and Fig. 3, it can be deduced that, the present invention has reached the smallest at mist node in the case where guaranteeing delay
Energy consumption.
From fig. 4, it can be seen that the variation more violent than having occurred is migrated, then as the number of iterations in preceding 100 iteration
Increase and increase, gradually converge to certain value.With the increase of task quantity, mist node tends to for calculating task being transferred to
Cloud Server, to reduce local energy consumption.When average unloading ratio converges to optimal valueWhen, it is optimal to indicate that optimization problem Pl is converged to
Value, that is, reach lowest energy consumption.
From fig. 5, it can be seen that heretofore described Speed gradient algorithm is than tradition ladder in the identical situation of task size
Degree method converges to the speed of stationary value faster.More precisely, speedup gradient method will converge to optimal value in 100 iteration,
Traditional gradient method the number of iterations is 200 times.Traditional gradient method is only iterated using back information, and accelerating gradient rule benefit
It is iterated with first two steps information.Therefore, the convergence rate of accelerating algorithm is significantly improved.At the same time, these numerical results
Further demonstrate the validity and superiority of designed Speed gradient algorithm.
The method that the present invention designs makes energy minimizing overhead under the premise of guaranteeing delay, design method of the present invention into
The optimization of migration decision is gone, and during solving optimal value, the convergence rate of accelerating algorithm is faster than traditional algorithm.Finally,
Numerical result demonstrate proposed migration scheme energy consumption, the deadline and in terms of superior function.
The foregoing is merely better embodiment of the invention, protection scope of the present invention is not with above embodiment
Limit, as long as those of ordinary skill in the art's equivalent modification or variation made by disclosure according to the present invention, should all be included in power
In the protection scope recorded in sharp claim.
Claims (6)
1. the energy efficient mist computation migration method with delay guaranteed in industrial scenes of internet of things, it is characterised in that: including such as
Lower step:
Step 1, the structure of energy consumption for mist node gives the comprehensive energy consumption for considering i.e. mist node by local computing, transmission
It is formed with the energy consumption of wait state, the mist node energy for constructing and constraining comprising energy expense and time delay is considered based on this and is disappeared
Consumption minimizes model；
Step 2, the constraint condition for minimizing model is obtained；
Step 3, the constraint condition according to obtained in step 2 solves the minimum model using Dual Method, constructs glug
Bright day function, and original optimization problem and corresponding dual problem are constructed, define optimized variable therein and dual variable；
Step 4, the Lagrangian that definition step 3 proposes corresponds to the gradient function of variable, by derived function and inverse function,
Obtain the calculation formula of optimized variable；
Step 5, it proposes a Speed gradient algorithm, defines temporary variable, and carry out the update of dual variable by temporary variable,
The optimal solution for obtaining dual problem in step 3 is solved, step 3 is then gone out according to the optimized variable formula proving in step 4
The optimal solution of middle optimization problem, the optimal solution of the optimization problem are to minimize the optimal solution of model.
2. the energy efficient mist computation migration side with delay guaranteed in industry scenes of internet of things according to claim 1
Method, it is characterised in that: in the step 1, it is assumed that mist node number n=10 has 1 Cloud Server, the calculating of these mist nodes
Ability is uniformly set as fi l=2*106Cycles/sec, i ∈ { 1,2 ..., 10 }, and the computing capability of Cloud Server is fc=2*108Week
Phase/second；In a communication environment, wireless signal-path band width B=50Mb/s, the Carrier To Noise Power Density N of channel are defined0=10-10W and
Wireless channel gain h between mist node and Cloud Serveri, andThe calculating task size w of different mist node isiIt is identical,
Their size is between 100Kb to 1000Kb；Complete revolving speed c required for cpu when 1 bit calculating taskiEqual to 1000 weeks
Phase/bit；The expectation energy consumption E of each mist nodeiFor 1.5J, it is expected that the T that is delayediFor 600ms；Power pi,cIt is 10-3W, energy consumption
miIt is 10-8Joule/the period；The message transmission rate of mist node i is defined as Ri；Symbol aiIndicate the data volume of progress computation migration
The ratio of Zhan total task data amount, and it can be expressed as di=wiai.Obviously, aiMeet 0≤a of conditioni≤1。
Energy consumption minimized model is as follows under tolerable time delay:
3. the energy efficient mist computation migration side with delay guaranteed in industry scenes of internet of things according to claim 1
Method, it is characterised in that: in the step 2, the constraint condition, specifically:
First constraint condition is that the deadline of task, the deadline included the processing locality time no more than the expected time, is passed
Defeated time and Cloud Server handle the time；
Second constraint condition is to need to meet the energy consumption of mist node no more than desired energy consumption, and energy consumption includes this
Ground calculates, transmission and waiting energy consumption；
Third constraint condition is channel width；
4th constraint condition is computing resource limitation.
4. the energy efficient mist computation migration side with delay guaranteed in industry scenes of internet of things according to claim 1
Method, it is characterised in that: in the step 3, specifically, Lagrangian is constructed by first and second constraint condition,
It is expressed as follows:
Wherein λiAnd μiIt is Lagrange multiplier, they are associated with task completion time and energy consumption respectively；
Secondly, the optimization problem for constructing following format is expressed as follows:
S.t.0≤a≤1, t > 0. (3)
To solve the above-mentioned problems, it is as follows to establish corresponding dual problem:
s.t.λ≥0,μ≥0. (4) 。
5. the energy efficient mist computation migration side with delay guaranteed in industry scenes of internet of things according to claim 1
Method, it is characterised in that: in the step 4, the dual problem according to step 3, definition LagrangianL (a, t, λ,
μ) about variable ai, ti, λiAnd μiGradient function it is as follows:
According to equation (6), function β=y (x)-xy'(x is defined), wherein y'(x)=N02x/BLn2 is that function y (x) must lead letter
Number, the inverse function that the variable x in beta function is then defined according to lambert's function are as follows:
Wherein W0(x) be lambert's function W Main Branches, be expressed asSymbol x and e are natural logrithms
Radix；
When equation (6) is set to 0, beta function can be rewritten as:
According to equation (9), available following equation:
Equation (10), (11), which are combined, to be rewritten as following form for equation (11):
When equation (7) is set to 0, available aiCalculated result are as follows:
Therefore, according to equation (12), (13), tiIt can be calculated as follows: ti=wiai/Ri。
6. the energy efficient mist computation migration side with delay guaranteed in industry scenes of internet of things according to claim 1
Method, it is characterised in that: in the step 5, the dual problem according to step 3, propose a Speed gradient algorithm with compared with
Fast convergence rate solves the optimal value of dual variable, the temporary variable being defined as follows:
As k >=1, temporary variable εi(r) and δi(r) itself is updated using the information of first two steps.(15), (16) are in equation
Basis, dual variable λiAnd μiIt is updated according to following:
Finally, solving the optimal solution for obtaining dual problem (4)Then according to equation (12), (13), derive that optimization is asked
Inscribe the optimal solution of (3)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910445848.8A CN110162894B (en) | 2019-05-27 | 2019-05-27 | Energy-efficient fog calculation migration method with delay guarantee in industrial Internet of things scene |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910445848.8A CN110162894B (en) | 2019-05-27 | 2019-05-27 | Energy-efficient fog calculation migration method with delay guarantee in industrial Internet of things scene |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110162894A true CN110162894A (en) | 2019-08-23 |
| CN110162894B CN110162894B (en) | 2023-06-16 |
Family
ID=67629160
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910445848.8A Active CN110162894B (en) | 2019-05-27 | 2019-05-27 | Energy-efficient fog calculation migration method with delay guarantee in industrial Internet of things scene |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110162894B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113285988A (en) * | 2021-05-14 | 2021-08-20 | 南京邮电大学 | Energy consumption minimization fair calculation migration method based on fog calculation |
| CN113326127A (en) * | 2021-05-28 | 2021-08-31 | 南京邮电大学 | Chargeable fog calculation migration method integrating wireless energy-carrying communication |
| CN113835894A (en) * | 2021-09-28 | 2021-12-24 | 南京邮电大学 | Intelligent calculation migration method based on double-delay depth certainty strategy gradient |
-
2019
- 2019-05-27 CN CN201910445848.8A patent/CN110162894B/en active Active
Non-Patent Citations (1)
| Title |
|---|
| 史子新: "雾计算环境下基于多目标优化的计算迁移策略研究", 《西安文理学院学报（自然科学版）》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113285988A (en) * | 2021-05-14 | 2021-08-20 | 南京邮电大学 | Energy consumption minimization fair calculation migration method based on fog calculation |
| CN113285988B (en) * | 2021-05-14 | 2022-07-29 | 南京邮电大学 | Energy consumption minimization fair calculation migration method based on fog calculation |
| CN113326127A (en) * | 2021-05-28 | 2021-08-31 | 南京邮电大学 | Chargeable fog calculation migration method integrating wireless energy-carrying communication |
| CN113326127B (en) * | 2021-05-28 | 2022-09-06 | 南京邮电大学 | Chargeable fog calculation migration method integrating wireless energy-carrying communication |
| CN113835894A (en) * | 2021-09-28 | 2021-12-24 | 南京邮电大学 | Intelligent calculation migration method based on double-delay depth certainty strategy gradient |
| CN113835894B (en) * | 2021-09-28 | 2024-03-26 | 南京邮电大学 | Intelligent calculation migration method based on dual-delay depth deterministic strategy gradient |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110162894B (en) | 2023-06-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Luo et al. | HFEL: Joint edge association and resource allocation for cost-efficient hierarchical federated edge learning | |
| Zeng et al. | Energy-efficient resource management for federated edge learning with CPU-GPU heterogeneous computing | |
| CN111176929B (en) | Edge federal learning-oriented high-energy-efficiency calculation communication joint optimization method | |
| Wen et al. | Joint optimal software caching, computation offloading and communications resource allocation for mobile edge computing | |
| CN110162894A (en) | With the energy efficient mist computation migration method of delay guaranteed in industrial scenes of internet of things | |
| Zhao et al. | Energy-aware task offloading and resource allocation for time-sensitive services in mobile edge computing systems | |
| Chen et al. | Delay guaranteed energy-efficient computation offloading for industrial IoT in fog computing | |
| CN112105062B (en) | Mobile edge computing network energy consumption minimization strategy method under time-sensitive condition | |
| Zhang et al. | Energy-efficient federated learning with intelligent reflecting surface | |
| Xie et al. | Dynamic computation offloading in IoT fog systems with imperfect channel-state information: A POMDP approach | |
| CN110351754A (en) | Industry internet machinery equipment user data based on Q-learning calculates unloading decision-making technique | |
| Li et al. | Delay optimization strategy for service cache and task offloading in three-tier architecture mobile edge computing system | |
| CN109618399A (en) | Distributed energy management solutions optimization method in the mobile edge calculations system of multi-user | |
| CN110351145A (en) | A kind of radio network functions method of combination of the virtualization based on economic benefit | |
| Teng et al. | Mixed-timescale joint computational offloading and wireless resource allocation strategy in energy harvesting multi-MEC server systems | |
| Zhang et al. | Joint offloading and resource allocation using deep reinforcement learning in mobile edge computing | |
| Cang et al. | Online resource allocation for semantic-aware edge computing systems | |
| Chu et al. | Federated learning over wireless channels: Dynamic resource allocation and task scheduling | |
| Zhong et al. | POTAM: A parallel optimal task allocation mechanism for large-scale delay sensitive mobile edge computing | |
| Zhao et al. | Energy-Efficient Federated Learning Over Cell-Free IoT Networks: Modeling and Optimization | |
| Zhao et al. | Energy-aware offloading in time-sensitive networks with mobile edge computing | |
| Lin et al. | Joint gradient sparsification and device scheduling for federated learning | |
| Zhao et al. | Partial critical path based greedy offloading in small cell cloud | |
| Sun | Certificateless batch authentication scheme and intrusion detection model based on the mobile edge computing technology NDN-IoT environment | |
| Lin et al. | Aoi research on pmu cloud side cooperative system of active distribution network |
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 |