CN109829332B - Joint calculation unloading method and device based on energy collection technology - Google Patents

Joint calculation unloading method and device based on energy collection technology Download PDF

Info

Publication number
CN109829332B
CN109829332B CN201910015984.3A CN201910015984A CN109829332B CN 109829332 B CN109829332 B CN 109829332B CN 201910015984 A CN201910015984 A CN 201910015984A CN 109829332 B CN109829332 B CN 109829332B
Authority
CN
China
Prior art keywords
mobile terminal
energy
task
time slice
function
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
Application number
CN201910015984.3A
Other languages
Chinese (zh)
Other versions
CN109829332A (en
Inventor
杜薇
雷启旺
刘伟
雷涛
赵海亮
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.)
Wuhan University of Technology WUT
Original Assignee
Wuhan University of Technology WUT
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 Wuhan University of Technology WUT filed Critical Wuhan University of Technology WUT
Priority to CN201910015984.3A priority Critical patent/CN109829332B/en
Publication of CN109829332A publication Critical patent/CN109829332A/en
Application granted granted Critical
Publication of CN109829332B publication Critical patent/CN109829332B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Mobile Radio Communication Systems (AREA)

Abstract

The invention provides a combined computing unloading method and a combined computing unloading device based on an energy collection technology, wherein the method comprises the following steps: after real-time information of the mobile terminal and the edge server in each preset time slice is collected, a virtual energy queue of each mobile terminal is constructed according to the real-time information of the mobile terminal; then constructing a task execution cost model, a terminal energy consumption model and an energy collection model; then, according to the real-time information of the edge server, the real-time information of all the mobile terminals, the constructed task execution cost model, the constructed terminal energy consumption model and the constructed energy collection model, constructing a Lyapunov drift and penalty function, and further obtaining a target function; and then taking the virtual energy queue as a queue needing to be stabilized, and solving a calculation unloading decision in the current time slice through a minimized objective function. The invention realizes the technical effects of effectively reducing the average execution cost of the tasks and the rejection rate of the tasks and improving the performance of the edge computing system.

Description

Joint calculation unloading method and device based on energy collection technology
Technical Field
The invention relates to the technical field of mobile edge calculation, in particular to a joint calculation unloading method and device based on an energy collection technology.
Background
With the rapid iterative update of mobile communication technology and the explosive popularization of various mobile intelligent terminals, mobile internet services have entered a rapid development stage. According to the estimation of the cisco global cloud index, by 2019, 45% of data generated by the internet of things equipment is stored, processed and analyzed at the network edge, and the total data traffic of a global data center is expected to reach 10.4 Zettabyte (ZB). Meanwhile, according to the forecast of the Cisco Internet service solution group, nearly 500 hundred million wireless devices are connected to the network in 2020.
In the prior art, a traditional Mobile Cloud Computing (MCC) mode is generally adopted, i.e. a mode of delivering and using IT resources or (information) services to obtain required infrastructure, platform, software (or application) in an on-demand and easily expandable manner through a Mobile network.
In the process of implementing the present invention, the applicant of the present invention finds that the methods in the prior art have at least the following technical problems:
because the application service based on the internet of everything platform needs shorter response time, and meanwhile, a large amount of private data needs to be protected better, the traditional mobile cloud computing mode cannot support the application service program based on the internet of everything efficiently due to limited computing capability.
Therefore, the method in the prior art has the technical problem of poor performance due to limited computing capacity.
Disclosure of Invention
In view of the above, the present invention provides a joint computation offloading method and apparatus based on an energy collection technology, so as to solve or at least partially solve the technical problem of poor performance caused by limited computation capability of the method in the prior art.
The invention provides a joint computation unloading method based on an energy collection technology, which comprises the following steps:
step S1: when each preset time slice starts, collecting real-time information of an edge server and real-time information of all mobile terminals, wherein the real-time information of the mobile terminals comprises collected energy, calculation task arrival conditions and positions of the mobile terminals;
step S2: constructing a virtual energy queue of each mobile terminal according to the collected energy, the task achievement condition and the position of the mobile terminal;
step S3: constructing a task execution cost model, a terminal energy consumption model and an energy collection model;
step S4: constructing a Lyapunov drift plus penalty function according to the real-time information of the edge server, the real-time information of all the mobile terminals, the constructed task execution cost model, the constructed terminal energy consumption model and the constructed energy collection model, and obtaining a target function according to the Lyapunov drift plus penalty function;
step S5: and taking the virtual energy queue as a queue needing to be stabilized, and solving a calculation unloading decision in the current time slice by minimizing an objective function, wherein the calculation unloading decision is used for representing an execution mode of a calculation task generated by the mobile terminal.
In one embodiment, after step S5, the method further comprises:
step S6: based on the objective function, resource allocation is performed for different computational offload decisions.
In one embodiment, step S2 specifically includes:
step S2.1: calculating the energy consumed by the mobile terminal in each preset time slice according to the arrival condition of the calculation task of the mobile terminal and the position of the mobile terminal;
step S2.2: according to the energy collected in each preset time slice and the energy consumed by the mobile terminal, the residual energy of the mobile terminal in the next time slice is calculated, wherein the calculation formula of the residual energy of the mobile terminal is as follows:
Figure GDA0002864482760000021
wherein the content of the first and second substances,
Figure GDA0002864482760000022
representing the remaining energy of the ith mobile terminal within the t time slice,
Figure GDA0002864482760000023
representing the remaining energy of the ith mobile terminal in the t +1 time slice,
Figure GDA0002864482760000024
representing the energy consumed by the mobile terminal within each time slice,
Figure GDA0002864482760000025
respectively representing the collected energy;
step S2.3: constructing a virtual energy queue according to the residual energy of each mobile terminal, wherein the virtual energy queue of the mobile terminal is represented as:
Figure GDA0002864482760000026
wherein the content of the first and second substances,
Figure GDA0002864482760000027
represents the virtual energy queue of the ith mobile terminal in the time slice t, thetaiIs the perturbation parameter of the mobile terminal.
In one embodiment, in step S3,
the task execution cost model of a single user is calculated according to the formula (3):
Figure GDA0002864482760000028
wherein the content of the first and second substances,
Figure GDA0002864482760000031
represents the execution delay of the ith mobile terminal generating the calculation task in the t time slice, including the execution delay of the calculation task executed in the mobile terminal or unloaded to the edge server, psi represents the penalty delay when the task is discarded,
Figure GDA0002864482760000032
and
Figure GDA0002864482760000033
respectively indicating that the ith mobile terminal generates a calculation task and the calculation task is abandoned in the t time slice;
the terminal energy consumption model is used for representing the energy consumed by the mobile terminal and the edge server in each time slice, and the calculation mode is as follows:
Figure GDA0002864482760000034
wherein the content of the first and second substances,
Figure GDA0002864482760000035
and
Figure GDA0002864482760000036
respectively representing the calculation task and the mobile terminalThe power consumption when offloaded to an edge server implementation,
Figure GDA0002864482760000037
and
Figure GDA0002864482760000038
respectively representing that the task is executed at the mobile terminal and the task is unloaded to the edge server for execution;
(3) the energy harvesting model is shown in equation (5):
Figure GDA0002864482760000039
wherein the content of the first and second substances,
Figure GDA00028644827600000310
represents the maximum energy collected in a single time slice, and the energy actually collected in each time slice of the mobile terminal is evenly distributed from 0 to 0
Figure GDA00028644827600000311
And the energy collected per time slice is independent of each other.
In one embodiment, step S4 specifically includes:
step S4.1: constructing a Lyapunov function, which is specifically shown in formula (6):
Figure GDA00028644827600000312
wherein U represents a set of mobile terminals;
step S4.2: constructing a Lyapunov drift function, which is specifically shown in a formula (7):
Figure GDA00028644827600000313
wherein the content of the first and second substances,
Figure GDA00028644827600000314
expressing the expectation;
step S4.3: constructing a Lyapunov drift and penalty function, which is specifically shown in formula (8):
Figure GDA00028644827600000315
wherein, V represents a penalty factor, and N represents the number of elements in the mobile terminal set;
step S4.4: according to the form of Lyapunov drift plus penalty function, the upper bound of the function is obtained, and the specific formula is as follows:
Figure GDA0002864482760000041
wherein, C is a constant,
Figure GDA0002864482760000042
representing the consumed energy of the ith mobile terminal in the t time slice; step S4.5: obtaining a target function according to the Lyapunov drift plus a penalty function
Figure GDA0002864482760000043
Figure GDA0002864482760000044
In an embodiment, the execution manner of the computing task includes that the computing task is discarded, the computing task is executed at the edge server, and the computing task is executed at the mobile terminal, and step S6 specifically includes:
when the calculation task is executed on the mobile terminal, a first resource allocation result is solved by minimizing an objective function, wherein the first resource allocation result is a CPU frequency, and the calculation method of the CPU frequency comprises the following steps:
Figure GDA0002864482760000045
wherein f isi,URepresents the maximum CPU frequency, f, that the mobile terminal can reachi,LA minimum CPU frequency indicating that the mobile terminal satisfies a task execution deadline,
Figure GDA0002864482760000046
is between fi,UAnd fi,LThe CPU frequency between the two;
by minimizing the objective function when the computing task is executed at the edge server
Figure GDA0002864482760000047
Solving a second resource allocation result, wherein the second resource allocation result is a network bandwidth allocation scheme, and the network bandwidth allocation scheme is as follows:
Figure GDA0002864482760000048
wherein, wi t*Representing the network bandwidth resulting from solving the objective function.
In one embodiment, the method further comprises:
constructing a return function for the mobile terminal executed on the edge server for the execution mode of the computing task, and adjusting the execution mode of the computing task generated by the mobile terminal based on the return function, wherein the return function is expressed as follows:
Figure GDA0002864482760000049
Figure GDA00028644827600000410
and
Figure GDA00028644827600000411
respectively indicates that the calculation task of the ith mobile terminal in the time slice t is executed locally,The drift plus penalty function values that are offloaded to the server execution and discarded,
Figure GDA00028644827600000412
and the report function value of the ith mobile terminal in the time slice t is shown.
In one embodiment, the method further comprises:
and adjusting the calculation unloading decision and the resource allocation result by judging whether the value of the target function meets the preset condition or not.
Based on the same inventive concept, the second aspect of the present invention provides a joint computation offloading device based on energy harvesting technology, comprising:
the information collection module is used for collecting the real-time information of the edge server and the real-time information of all the mobile terminals when each preset time slice starts, wherein the real-time information of the mobile terminals comprises collected energy, calculation task arrival conditions and the positions of the mobile terminals;
the queue building module is used for building a virtual energy queue of each mobile terminal according to the collected energy, the achievement condition of the calculation task and the position of the mobile terminal;
the model construction module is used for constructing a task execution cost model, a terminal energy consumption model and an energy collection model;
the target function obtaining module is used for constructing a Lyapunov drift penalty function according to the real-time information of the edge server, the real-time information of all the mobile terminals, the constructed task execution cost model, the constructed terminal energy consumption model and the constructed energy collection model, and obtaining a target function according to the Lyapunov drift penalty function;
and the unloading decision module is used for solving a calculation unloading decision in the current time slice by taking the virtual energy queue as a queue needing to be stabilized through a minimized objective function, wherein the calculation unloading decision is used for representing an execution mode of a calculation task generated by the mobile terminal.
Based on the same inventive concept, a third aspect of the present invention provides a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method according to the first aspect when executing the program.
One or more technical solutions in the embodiments of the present application have at least one or more of the following technical effects:
the method provided by the invention comprises the steps of establishing a virtual energy queue, a task execution cost model, a terminal energy consumption model and an energy collection model by collecting real-time information of a mobile terminal and an edge server, and constructing a Lyapunov drift penalty function according to the real-time information of the edge server, the real-time information of all mobile terminals, the constructed task execution cost model, the constructed terminal energy consumption model and the constructed energy collection model, so as to obtain a target function; and finally, taking the virtual energy queue as a queue needing to be stabilized, and solving a calculation unloading decision in the current time slice through a minimized objective function.
Compared with the traditional mobile cloud computing mode in the prior art, the invention provides a multi-user joint computing unloading mobile edge computing method based on an energy collection technology.
Further, by adopting the Lyapunov-based optimization technology, resource allocation can be performed based on the current calculation unloading decision, so that the resource allocation result is optimized.
Furthermore, by constructing a return function, the computation and unloading decision of multiple users can be adjusted, so that the computation and unloading decision is optimal at the level of an edge computing system.
Further, through an iterative mode, the optimal computation unloading and resource allocation strategy is finally made by combining computation unloading and resource allocation, so that the average task execution cost of multiple users is reduced, the task abandon rate is reduced, and the user experience quality of a system level is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a flow chart of a joint computation offloading method based on energy harvesting techniques in an embodiment of the present invention;
FIG. 2 is a block diagram of a joint computation offload device based on energy harvesting technology according to an embodiment of the present invention;
FIG. 3 is a block diagram of computing offload and resource allocation in an embodiment of the present invention;
fig. 4 is a block diagram of a computer device in an embodiment of the present invention.
Detailed Description
The invention aims to provide a multi-user joint computing unloading and resource allocation method based on an energy collection technology, aiming at the technical problem that the performance is poor due to limited computing capability of the traditional mobile cloud computing method. By collecting real-time information of the mobile terminal and the edge server, a task execution cost model, a terminal energy consumption model, an energy collection model and a virtual energy queue are established, and an initial unloading decision set is obtained.
And then, based on a Lyapunov optimization technology and based on a current unloading decision set (namely, calculating an unloading decision), performing a multi-user optimal resource allocation strategy, and adjusting the multi-user unloading decision set in reverse through a return function to enable the unloading decision set to be optimal on the level of an edge computing system. In addition, an optimal computation unloading and resource allocation strategy is finally made by combining computation unloading and resource allocation in an iterative mode, so that the average task execution cost of multiple users is reduced, the task abandon rate is reduced, and the user experience quality of a system level is improved.
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example one
The embodiment provides a joint computation offloading method based on an energy harvesting technology, please refer to fig. 1, which includes:
step S1 is first executed: and when each preset time slice starts, collecting the real-time information of the edge server and the real-time information of all the mobile terminals, wherein the real-time information of the mobile terminals comprises the collected energy, the arrival condition of the calculation task and the positions of the mobile terminals.
Specifically, the preset time slice may be set according to actual conditions, for example, 1ms, 1s, 10s, and the like. The real-time information of the mobile terminal comprises the position of the mobile terminal, the arrival condition of a calculation task, the residual energy, the collected energy and the like, and the real-time information of the edge server comprises the network bandwidth, the calculation resources and the like of the edge server.
The mobile terminal based on the energy collection technology is not limited to the limit of the power grid any more, and the mobile terminal can collect energy even from light, vibration or temperature change by virtue of the information collection module. Due to the use of an energy collection technology, the development of technologies such as the Internet of things and wearable equipment enters a brand new stage.
With the proposal of Mobile Edge Computing (MEC), the disadvantages of the traditional Mobile cloud Computing mode are well solved. In the mobile edge computing model, the network edge device already has enough computing power to implement local processing of data and send the computation results to the mobile terminal. The edge calculation model not only reduces the data transmission bandwidth, but also can better protect the privacy data of the user and reduce the sharing of the privacy disclosure of the sensitive data of the terminal. The mobile device based on the energy collection technology can avoid the dependence of the mobile terminal on a power grid in an edge computing mode, and can capture, collect and then utilize small-batch energy through the electronic device through the information collection module, so that complex computing tasks can be completed without integrating a traditional power supply in system design. Therefore, the research on the problems of computing unloading and resource allocation of the mobile equipment based on the energy collection technology in the marginal computing environment has great significance to the world of everything interconnection.
Then, step S2 is executed: and constructing a virtual energy queue of each mobile terminal according to the collected energy, the task achievement condition and the position of the mobile terminal.
Specifically, the virtual energy queue is an initial queue constructed according to the energy situation of the mobile terminal.
In one embodiment, step S2 specifically includes:
step S2.1: calculating the energy consumed by the mobile terminal in each preset time slice according to the arrival condition of the calculation task of the mobile terminal and the position of the mobile terminal;
step S2.2: according to the energy collected in each preset time slice and the energy consumed by the mobile terminal, the residual energy of the mobile terminal in the next time slice is calculated, wherein the calculation formula of the residual energy of the mobile terminal is as follows:
Figure GDA0002864482760000081
wherein the content of the first and second substances,
Figure GDA0002864482760000082
representing the remaining energy of the ith mobile terminal within the t time slice,
Figure GDA0002864482760000083
representing the remaining energy of the ith mobile terminal in the t +1 time slice,
Figure GDA0002864482760000084
representing the energy consumed by the mobile terminal within each time slice,
Figure GDA0002864482760000085
respectively representing the collected energy;
step S2.3: constructing a virtual energy queue according to the residual energy of each mobile terminal, wherein the virtual energy queue of the mobile terminal is represented as:
Figure GDA0002864482760000086
wherein the content of the first and second substances,
Figure GDA0002864482760000087
represents the virtual energy queue of the ith mobile terminal in the time slice t, thetaiIs the perturbation parameter of the mobile terminal.
Specifically, by establishing the virtual energy queue, the residual energy of the mobile terminal can be stabilized near a value, so that the situation that the energy of the mobile terminal is too low to execute a calculation task is avoided, and the calculation task is abandoned. When the virtual energy queue of the mobile terminal is stable, the residual energy is thetaiThe vicinity fluctuates.
Step S3 is executed next: and constructing a task execution cost model, a terminal energy consumption model and an energy collection model.
Specifically, the corresponding model may be constructed according to a specific application, and in this embodiment, the execution cost, the energy consumption, and the energy are mainly considered.
In one embodiment, in step S3,
the task execution cost model of a single user is calculated according to the formula (3):
Figure GDA0002864482760000088
wherein the content of the first and second substances,
Figure GDA0002864482760000089
represents the execution delay of the ith mobile terminal generating the calculation task in the t time slice, including the execution delay of the calculation task executed in the mobile terminal or unloaded to the edge server, psi represents the penalty delay when the task is discarded,
Figure GDA00028644827600000810
and
Figure GDA00028644827600000811
respectively indicating that the ith mobile terminal generates a calculation task in the time slice t and the calculation task is abandoned.
In particular, task execution cost, which is defined as a weighted sum of the execution delay of a task and the penalty delay when discarded, is an important metric for measuring the performance of an edge computing system. By minimizing the task execution cost, penalty delays may be reduced, thereby reducing the task castout rate.
The terminal energy consumption model is used for representing the energy consumed by the mobile terminal and the edge server in each time slice, and the calculation mode is as follows:
Figure GDA0002864482760000091
wherein the content of the first and second substances,
Figure GDA0002864482760000092
and
Figure GDA0002864482760000093
respectively representing the energy consumption of the computing task when the mobile terminal and the computing task are offloaded to the edge server for execution,
Figure GDA0002864482760000094
and
Figure GDA0002864482760000095
respectively representing that a task is executed on a mobile terminal and the task is unloadedTo the edge server for execution;
specifically, energy consumption may be generated by a plurality of components in the mobile terminal, the energy consumption generated by data transmission between the mobile terminal CPU and the network interface is mainly considered in the present invention, the energy consumption generated by the terminal CPU is mainly generated when the computing task is executed locally (i.e. executed in the mobile terminal), and the energy consumption generated by the data transmission between the network interface is mainly generated when the computing task is unloaded to the edge server for execution.
(3) The energy harvesting model is shown in equation (5):
Figure GDA0002864482760000096
wherein the content of the first and second substances,
Figure GDA0002864482760000097
represents the maximum energy collected in a single time slice, and the energy actually collected in each time slice of the mobile terminal is evenly distributed from 0 to 0
Figure GDA0002864482760000098
And the energy collected per time slice is independent of each other.
Specifically, the characteristic of collecting energy is specific to a mobile terminal based on an energy collection technology, the mobile terminal collects energy in the nature through an information collection module to improve the residual energy of the mobile terminal and prolong the working time of the terminal, and an energy collection model is mainly used for simulating randomness and unpredictability of collecting energy in the nature.
Step S4 is executed again: and constructing a Lyapunov drift plus penalty function according to the real-time information of the edge server, the real-time information of all the mobile terminals, the constructed task execution cost model, the constructed terminal energy consumption model and the constructed energy collection model, and obtaining a target function according to the Lyapunov drift plus penalty function.
In particular, Lyapunov stability (Lyapunov stability, or Lyapunov stability) may be used to describe the stability of a powertrain system. If the trajectory of any initial condition of the system near the equilibrium state is maintained near the equilibrium state, it may be referred to as at-lyapunov stabilization.
In one embodiment, step S4 specifically includes:
step S4.1: constructing a Lyapunov function, which is specifically shown in formula (6):
Figure GDA0002864482760000101
wherein U represents a set of mobile terminals;
step S4.2: constructing a Lyapunov drift function, which is specifically shown in a formula (7):
Figure GDA0002864482760000102
wherein the content of the first and second substances,
Figure GDA0002864482760000103
expressing the expectation;
step S4.3: constructing a Lyapunov drift and penalty function, which is specifically shown in formula (8):
Figure GDA0002864482760000104
wherein, V represents a penalty factor, and N represents the number of elements in the mobile terminal set;
step S4.4: according to the form of Lyapunov drift plus penalty function, the upper bound of the function is obtained, and the specific formula is as follows:
Figure GDA0002864482760000105
wherein, C is a constant,
Figure GDA0002864482760000106
indicating the ith mobile terminal at time tEnergy consumed within the chip;
step S4.5: obtaining a target function according to the Lyapunov drift plus a penalty function
Figure GDA0002864482760000107
Figure GDA0002864482760000108
Specifically, minimizing the upper bound of lyapunov drift plus penalty functions may result in optimal computational offloading and resource allocation decisions. For the sake of clarity and conciseness of the expression, the specific optimization objective may be optimized in the form of equation (10).
Step S5 is executed again: and taking the virtual energy queue as a queue needing to be stabilized, and solving a calculation unloading decision in the current time slice by minimizing an objective function, wherein the calculation unloading decision is used for representing an execution mode of a calculation task generated by the mobile terminal.
Specifically, the task offloading scheme, which is a computation offloading decision in the current time slice, may specifically include that the computation task is discarded, the computation task is executed at the edge server, and the computation task is executed at the mobile terminal.
The virtual energy queue of the mobile terminal is used as a queue needing to be stable, the average task execution cost of the multiple mobile terminals is used as a penalty term of the Lyapunov function, and the weight of the penalty term can be controlled by adjusting the size of the penalty factor V. By constructing the Lyapunov drift plus penalty function, the residual energy of the mobile terminal can be stabilized, the calculation task is prevented from being abandoned, the average execution cost of the task can be reduced, and the performance of the edge calculation system is improved.
In one embodiment, after step S5, the method further comprises:
step S6: based on the objective function, resource allocation is performed for different computational offload decisions.
Specifically, after the current offload decision is obtained through the foregoing steps, resource allocation may be further performed through an objective function.
In an embodiment, the execution manner of the computing task includes that the computing task is discarded, the computing task is executed at the edge server, and the computing task is executed at the mobile terminal, and step S6 specifically includes:
when the calculation task is executed on the mobile terminal, a first resource allocation result is solved by minimizing an objective function, wherein the first resource allocation result is a CPU frequency, and the calculation method of the CPU frequency comprises the following steps:
Figure GDA0002864482760000111
wherein f isi,URepresents the maximum CPU frequency, f, that the mobile terminal can reachi,LA minimum CPU frequency indicating that the mobile terminal satisfies a task execution deadline,
Figure GDA0002864482760000112
is between fi,UAnd fi,LThe CPU frequency between the two.
In particular, the present invention relates to a method for producing,
Figure GDA0002864482760000113
is between fi,UAnd fi,LThe optimal CPU frequency between the two and the concrete expression form of the three are related to the task model and the energy consumption model in the foregoing.
Specifically, according to the initial state information, the mobile terminal with the calculation task is set as the task to be executed at the edge server, so that an initial unloading decision set is determined. The initial offload decision set is:
Figure GDA0002864482760000114
where A (t), B (t), and C (t) represent the set of offload decisions as local execution, edge server execution, and discard, respectively. It can be seen that the set of tasks in the initial offload decision set that are executed locally is an empty set, i.e., the set is empty
Figure GDA0002864482760000115
The calculation mode of the mobile terminal generated by the calculation task is executed by the edge server, and the calculation mode of the mobile terminal generated by the calculation task is abandoned.
By minimizing the objective function when the computing task is executed at the edge server
Figure GDA0002864482760000116
Solving a second resource allocation result, wherein the second resource allocation result is a network bandwidth allocation scheme, and the network bandwidth allocation scheme is as follows:
Figure GDA0002864482760000121
wherein, wi t*Representing the network bandwidth resulting from solving the objective function.
In particular, the objective function
Figure GDA0002864482760000122
Can be represented as follows
Figure GDA0002864482760000123
Wherein the content of the first and second substances,
Figure GDA0002864482760000124
and
Figure GDA0002864482760000125
a value representing a drift-plus-penalty function when a computing task of the mobile terminal is executed locally, is offloaded to an edge server for execution and is discarded,
Figure GDA0002864482760000126
and
Figure GDA0002864482760000127
are respectively the tasks ofAn indication of whether the function is executing locally, offloaded to an edge server, and discarded. By minimizing an objective function
Figure GDA0002864482760000128
An optimal network bandwidth resource allocation scheme can be obtained, thereby realizing optimal resource allocation.
In one embodiment, the method further comprises:
constructing a return function for the mobile terminal executed on the edge server for the execution mode of the computing task, and adjusting the execution mode of the computing task generated by the mobile terminal based on the return function, wherein the return function is expressed as follows:
Figure GDA0002864482760000129
Figure GDA00028644827600001210
and
Figure GDA00028644827600001211
respectively representing the values of the drift-plus-penalty sub-function when the computing task of the ith mobile terminal in the time slice t is executed locally, is unloaded to the server and is discarded,
Figure GDA00028644827600001212
and the report function value of the ith mobile terminal in the time slice t is shown.
Specifically, it is obvious that a larger reward function value indicates that a higher profit is brought by the mobile terminal to offload the computing task to the edge server, whereas a smaller reward function value indicates that a lower profit is brought by the mobile terminal to offload the computing task to the edge server. Therefore, the mobile terminal which should be adjusted in the calculation mode most can be found by traversing the reward function value. That is, only the mobile terminal with the smallest value of the reward function needs to be found, for example, if the value is smaller
Figure GDA00028644827600001213
The calculation mode is adjusted to be executed locally, if the value is
Figure GDA00028644827600001214
The calculation mode is adjusted to discard the task.
In one embodiment, the method further comprises:
and adjusting the calculation unloading decision and the resource allocation result by judging whether the value of the target function meets the preset condition or not.
Specifically, the current unloading decision and the objective function under the optimal resource allocation condition can be calculated
Figure GDA0002864482760000131
If any offload decision set is readjusted, the objective function is made
Figure GDA0002864482760000132
If the value of (a) is increased, the current unloading decision and the optimal resource allocation are optimal (namely, the preset conditions are met), and at the moment, the combined calculation unloading and resource allocation strategy is obtained; if the calculation mode of the mobile terminal with the maximum return function value is adjusted, the objective function is enabled
Figure GDA0002864482760000133
If the value of (a) is smaller, it indicates that the current offload decision is not optimal, and therefore, the calculation mode of the mobile terminal with the constructed reward function needs to be adjusted, and then resource allocation is continued for the mobile terminal.
In this embodiment, the objective function is determined
Figure GDA0002864482760000134
Is minimum, if not, the step of constructing the reward function is repeated, the process is an iterative process until the objective function is reached
Figure GDA0002864482760000135
And if the value is the minimum, skipping iteration at the moment, and outputting the optimal unloading decision and resource allocation result.
It should be noted that the formulas and the meanings of the characters involved in the present invention are shown in table 1.
TABLE 1
Figure GDA0002864482760000136
Figure GDA0002864482760000141
Based on the same inventive concept, the application also provides a device corresponding to the joint computation unloading method based on the energy collection technology in the first embodiment, which is detailed in the second embodiment.
Example two
The embodiment provides a joint computation offloading device based on energy harvesting technology, please refer to fig. 2, the device includes:
an information collecting module 201, configured to collect real-time information of the edge server and real-time information of all mobile terminals when each preset time slice starts, where the real-time information of the mobile terminals includes collected energy, a calculation task arrival condition, and a location of the mobile terminal;
a queue building module 202, configured to build a virtual energy queue of each mobile terminal according to the collected energy, the task achievement condition, and the position of the mobile terminal;
the model construction module 203 is used for constructing a task execution cost model, a terminal energy consumption model and an energy collection model;
an objective function obtaining module 204, configured to construct a lyapunov drift plus penalty function according to the real-time information of the edge server, the real-time information of all mobile terminals, and the constructed task execution cost model, the terminal energy consumption model, and the energy collection model, and obtain an objective function according to the lyapunov drift plus penalty function;
and the unloading decision module 205 is configured to take the virtual energy queue as a queue needing to be stabilized, and solve a calculation unloading decision in the current time slice by minimizing an objective function, where the calculation unloading decision is used to represent an execution manner of a calculation task generated by the mobile terminal.
In one embodiment, the apparatus further comprises a resource allocation module for, after solving for the computational offload decision within the current timeslice:
based on the objective function, resource allocation is performed for different computational offload decisions.
In one embodiment, the queue building module 202 is specifically configured to perform the following steps:
step S2.1: calculating the energy consumed by the mobile terminal in each preset time slice according to the arrival condition of the calculation task of the mobile terminal and the position of the mobile terminal;
step S2.2: according to the energy collected in each preset time slice and the energy consumed by the mobile terminal, the residual energy of the mobile terminal in the next time slice is calculated, wherein the calculation formula of the residual energy of the mobile terminal is as follows:
Figure GDA0002864482760000151
wherein the content of the first and second substances,
Figure GDA0002864482760000152
representing the remaining energy of the ith mobile terminal within the t time slice,
Figure GDA0002864482760000153
representing the remaining energy of the ith mobile terminal in the t +1 time slice,
Figure GDA0002864482760000154
representing the energy consumed by the mobile terminal within each time slice,
Figure GDA0002864482760000155
respectively representing the collected energy;
step S2.3: constructing a virtual energy queue according to the residual energy of each mobile terminal, wherein the virtual energy queue of the mobile terminal is represented as:
Figure GDA0002864482760000156
wherein the content of the first and second substances,
Figure GDA0002864482760000157
represents the virtual energy queue of the ith mobile terminal in the time slice t, thetaiIs the perturbation parameter of the mobile terminal.
In one embodiment, in model building module 203,
the task execution cost model of a single user is calculated according to the formula (3):
Figure GDA0002864482760000158
wherein the content of the first and second substances,
Figure GDA0002864482760000159
represents the execution delay of the ith mobile terminal generating the calculation task in the t time slice, including the execution delay of the calculation task executed in the mobile terminal or unloaded to the edge server, psi represents the penalty delay when the task is discarded,
Figure GDA00028644827600001510
and
Figure GDA00028644827600001511
respectively indicating that the ith mobile terminal generates a calculation task and the calculation task is abandoned in the t time slice;
the terminal energy consumption model is used for representing the energy consumed by the mobile terminal and the edge server in each time slice, and the calculation mode is as follows:
Figure GDA00028644827600001512
wherein the content of the first and second substances,
Figure GDA00028644827600001513
and
Figure GDA00028644827600001514
respectively representing the energy consumption of the computing task when the mobile terminal and the computing task are offloaded to the edge server for execution,
Figure GDA00028644827600001515
and
Figure GDA00028644827600001516
respectively representing that the task is executed at the mobile terminal and the task is unloaded to the edge server for execution;
(3) the energy harvesting model is shown in equation (5):
Figure GDA00028644827600001517
wherein the content of the first and second substances,
Figure GDA00028644827600001518
represents the maximum energy collected in a single time slice, and the energy actually collected in each time slice of the mobile terminal is evenly distributed from 0 to 0
Figure GDA00028644827600001519
And the energy collected per time slice is independent of each other.
In one embodiment, the objective function obtaining module 204 is specifically configured to perform the following steps:
step S4.1: constructing a Lyapunov function, which is specifically shown in formula (6):
Figure GDA0002864482760000161
wherein U represents a set of mobile terminals;
step S4.2: constructing a Lyapunov drift function, which is specifically shown in a formula (7):
Figure GDA0002864482760000162
wherein the content of the first and second substances,
Figure GDA0002864482760000163
expressing the expectation;
step S4.3: constructing a Lyapunov drift and penalty function, which is specifically shown in formula (8):
Figure GDA0002864482760000164
wherein, V represents a penalty factor, and N represents the number of elements in the mobile terminal set;
step S4.4: according to the form of Lyapunov drift plus penalty function, the upper bound of the function is obtained, and the specific formula is as follows:
Figure GDA0002864482760000165
wherein, C is a constant,
Figure GDA0002864482760000166
representing the consumed energy of the ith mobile terminal in the t time slice;
step S4.5: obtaining a target function according to the Lyapunov drift plus a penalty function
Figure GDA0002864482760000167
Figure GDA0002864482760000168
In an embodiment, the execution mode of the computing task includes that the computing task is discarded, the computing task is executed at the edge server, and the computing task is executed at the mobile terminal, and the resource allocation module is specifically configured to:
when the calculation task is executed on the mobile terminal, a first resource allocation result is solved by minimizing an objective function, wherein the first resource allocation result is a CPU frequency, and the calculation method of the CPU frequency comprises the following steps:
Figure GDA0002864482760000169
wherein f isi,URepresents the maximum CPU frequency, f, that the mobile terminal can reachi,LA minimum CPU frequency indicating that the mobile terminal satisfies a task execution deadline,
Figure GDA00028644827600001610
is between fi,UAnd fi,LThe CPU frequency between the two;
by minimizing the objective function when the computing task is executed at the edge server
Figure GDA0002864482760000171
Solving a second resource allocation result, wherein the second resource allocation result is a network bandwidth allocation scheme, and the network bandwidth allocation scheme is as follows:
Figure GDA0002864482760000172
wherein, wi t*Representing the network bandwidth resulting from solving the objective function.
In one embodiment, the method further comprises a first adjusting module for:
constructing a return function for the mobile terminal executed on the edge server for the execution mode of the computing task, and adjusting the execution mode of the computing task generated by the mobile terminal based on the return function, wherein the return function is expressed as follows:
Figure GDA0002864482760000173
Figure GDA0002864482760000174
and
Figure GDA0002864482760000175
respectively representing the values of the drift-plus-penalty sub-function when the computing task of the ith mobile terminal in the time slice t is executed locally, is unloaded to the server and is discarded,
Figure GDA0002864482760000176
and the report function value of the ith mobile terminal in the time slice t is shown.
In one embodiment, the apparatus further comprises a second adjustment module configured to:
and adjusting the calculation unloading decision and the resource allocation result by judging whether the value of the target function meets the preset condition or not.
In order to more clearly illustrate the embodiment of the apparatus of the present invention, it is described in detail below with reference to a model example, and in particular, fig. 3.
Generally, the invention mainly adopts the technical scheme that: the device specifically comprises a calculation unloading module and a resource allocation module.
The calculation unloading module includes an information collection module, a queue construction module, a model construction module, an objective function obtaining module, and an unloading decision module, and it should be noted that fig. 3 shows that the unloading decision module includes a queue construction module and a model construction module, and an objective function obtaining module. Thus, from this figure, the computation offload module includes an information collection module and an offload decision module. The calculation unloading module determines whether the calculation task generated by the mobile terminal is executed at the mobile terminal, executed by the edge server or abandoned, and the decision is given by the unloading decision module. The information collection module is responsible for collectingAnd the real-time information of the mobile terminal and the edge server is collected, such as the position of the mobile terminal, the residual electric quantity, the task arrival condition, the energy collected by the energy collection module, the computing resource and the network bandwidth resource of the edge server and the like. And the unloading decision module gives an unloading decision according to the established model and the information collected by the information collection module. In an initial state, the offload decision module classifies the computing mode of the mobile terminal into two categories according to whether the mobile terminal generates a computing task: the calculation mode of the mobile terminal with the task generation is unloaded execution, and the calculation mode of the mobile terminal without the task generation is abandoned. Late stage pass minimization objective function
Figure GDA0002864482760000181
And determines its value to determine whether to adjust the calculation mode to be an offloaded mobile terminal to be executed locally or discarded.
The resource allocation module is a successor module to the compute offload module. For certain computing offload decisions, the resource allocation module may allocate resources according to different computing modes. The computing mode is a mobile terminal executed locally, and the resource allocation module can allocate the optimal mobile terminal CPU frequency; the calculation mode is to unload the executed mobile terminal, and the resource allocation module can allocate the optimal network bandwidth resource. Then by judging the objective function
Figure GDA0002864482760000182
Judging whether the unloading decision is optimal or not by judging whether the value of (A) is minimum or not, if not, adjusting the unloading decision to achieve the optimal state, and if so, directly outputting the optimal calculation unloading and resource allocation decision. Specifically, the resource allocation module is responsible for allocating the computing resources of the edge server and the bandwidth resources of the wireless channel; the allocation of the computing resources is mainly to allocate the CPU core of the edge server, and the mobile terminal needs to occupy one CPU core when unloading the computing task to the edge server for execution. The wireless channel bandwidth resource allocation is mainly to allocate the network bandwidth of the edge server. The resource allocation module can determine all movements according to the unloading decision and the real-time information of the mobile terminal taskAnd (4) an optimal resource allocation scheme of the terminal.
Since the apparatus described in the second embodiment of the present invention is an apparatus used for implementing the joint computation offloading method based on the energy harvesting technology in the first embodiment of the present invention, a person skilled in the art can understand the specific structure and the deformation of the apparatus based on the method described in the first embodiment of the present invention, and thus the details are not described herein again. All the devices adopted in the method of the first embodiment of the present invention belong to the protection scope of the present invention.
EXAMPLE III
Based on the same inventive concept, the present application further provides a computer device, please refer to fig. 4, which includes a storage 401, a processor 402, and a computer program 403 stored in the storage and running on the processor, and when the processor 402 executes the above program, the method in the first embodiment is implemented.
Since the computer device introduced in the third embodiment of the present invention is a computer device used for implementing the joint computation offloading method based on the energy harvesting technology in the first embodiment of the present invention, based on the method introduced in the first embodiment of the present invention, those skilled in the art can understand the specific structure and deformation of the computer device, and thus, details are not described herein. All the computer devices used in the method in the first embodiment of the present invention are within the scope of the present invention.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various modifications and variations can be made in the embodiments of the present invention without departing from the spirit or scope of the embodiments of the invention. Thus, if such modifications and variations of the embodiments of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to encompass such modifications and variations.

Claims (8)

1. A joint computation unloading method based on an energy collection technology is characterized by comprising the following steps:
step S1: when each preset time slice starts, collecting real-time information of an edge server and real-time information of all mobile terminals, wherein the real-time information of the mobile terminals comprises collected energy, calculation task arrival conditions and positions of the mobile terminals;
step S2: constructing a virtual energy queue of each mobile terminal according to the collected energy, the arrival condition of the calculation task and the position of the mobile terminal;
step S3: constructing a task execution cost model, a terminal energy consumption model and an energy collection model;
step S4: constructing a Lyapunov drift plus penalty function according to the real-time information of the edge server, the real-time information of all the mobile terminals, the constructed task execution cost model, the constructed terminal energy consumption model and the constructed energy collection model, and obtaining a target function according to the Lyapunov drift plus penalty function;
step S5: taking the virtual energy queue as a queue needing to be stabilized, and solving a calculation unloading decision in the current time slice through a minimized objective function, wherein the calculation unloading decision is used for representing an execution mode of a calculation task generated by the mobile terminal;
after step S5, the method further includes:
step S6: based on the objective function, performing resource allocation aiming at different calculation unloading decisions;
wherein, step S4 specifically includes:
step S4.1: constructing a Lyapunov function, which is specifically shown in formula (6):
Figure FDA0002916223320000011
wherein U represents a set of mobile terminals;
step S4.2: constructing a Lyapunov drift function, which is specifically shown in a formula (7):
Figure FDA0002916223320000012
wherein the content of the first and second substances,
Figure FDA0002916223320000013
expressing the expectation;
step S4.3: constructing a Lyapunov drift and penalty function, which is specifically shown in formula (8):
Figure FDA0002916223320000014
wherein, V represents a penalty factor, and N represents the number of elements in the mobile terminal set;
step S4.4: according to the form of Lyapunov drift plus penalty function, the upper bound of the function is obtained, and the specific formula is as follows:
Figure FDA0002916223320000021
wherein, C is a constant,
Figure FDA0002916223320000022
representing the consumed energy of the ith mobile terminal in the t time slice;
step S4.5: obtaining a target function according to the Lyapunov drift plus a penalty function
Figure FDA0002916223320000023
Figure FDA0002916223320000024
Wherein the content of the first and second substances,
Figure FDA0002916223320000025
a virtual energy queue representing the ith mobile terminal at time t,
Figure FDA0002916223320000026
representing the collected energy of the ith mobile terminal within the t time slice,
Figure FDA0002916223320000027
indicating the cost of execution of the ith mobile terminal at time t,
Figure FDA0002916223320000028
weighting factor, I, being the execution cost in the drift plus penalty functiont,ft,wtRespectively representing the calculation mode, the CPU frequency and the bandwidth size of the mobile terminal in the time slice t,
Figure FDA0002916223320000029
respectively representing the calculation mode, the CPU frequency and the bandwidth size of the ith mobile terminal in the time slice t.
2. The method according to claim 1, wherein step S2 specifically comprises:
step S2.1: calculating the energy consumed by the mobile terminal in each preset time slice according to the arrival condition of the calculation task of the mobile terminal and the position of the mobile terminal;
step S2.2: according to the energy collected in each preset time slice and the energy consumed by the mobile terminal, the residual energy of the mobile terminal in the next time slice is calculated, wherein the calculation formula of the residual energy of the mobile terminal is as follows:
Figure FDA00029162233200000210
wherein the content of the first and second substances,
Figure FDA00029162233200000211
representing the remaining energy of the ith mobile terminal within the t time slice,
Figure FDA00029162233200000212
representing the remaining energy of the ith mobile terminal in the t +1 time slice,
Figure FDA00029162233200000213
representing the collected energy;
step S2.3: constructing a virtual energy queue according to the residual energy of each mobile terminal, wherein the virtual energy queue of the mobile terminal is represented as:
Figure FDA00029162233200000214
wherein the content of the first and second substances,
Figure FDA00029162233200000215
represents the virtual energy queue of the ith mobile terminal in the time slice t, thetaiIs the perturbation parameter of the mobile terminal.
3. The method of claim 1, wherein in step S3,
the task execution cost model of a single user is calculated according to the formula (3):
Figure FDA0002916223320000031
wherein the content of the first and second substances,
Figure FDA0002916223320000032
representing the execution delay of the ith mobile terminal to generate a computation task within a time slice t, including the execution delay of the computation task executed at the mobile terminal or offloaded to an edge server, It,ft,wtRespectively representing the calculation mode, the CPU frequency and the bandwidth size of the mobile terminal within a time slice t, psi represents the penalty delay when a task is discarded,
Figure FDA0002916223320000033
and
Figure FDA0002916223320000034
respectively indicating that the ith mobile terminal generates a calculation task and the calculation task is abandoned in the t time slice;
the terminal energy consumption model is used for representing the energy consumed by the mobile terminal and the edge server in each time slice, and the calculation mode is as follows:
Figure FDA0002916223320000035
wherein the content of the first and second substances,
Figure FDA0002916223320000036
and
Figure FDA0002916223320000037
respectively representing the energy consumption of the computing task when the mobile terminal and the computing task are offloaded to the edge server for execution,
Figure FDA0002916223320000038
and
Figure FDA0002916223320000039
respectively representing that the task is executed at the mobile terminal and the task is unloaded to the edge server for execution;
(3) the energy harvesting model is shown in equation (5):
Figure FDA00029162233200000310
wherein the content of the first and second substances,
Figure FDA00029162233200000311
represents the maximum energy collected in a single time slice, and the energy actually collected in each time slice of the mobile terminal is evenly distributed from 0 to 0
Figure FDA00029162233200000312
And the energy collected per time slice is independent of each other.
4. The method according to claim 1, wherein the execution manner of the computing task includes that the computing task is discarded, the computing task is executed at the edge server, and the computing task is executed at the mobile terminal, and the step S6 specifically includes:
when the calculation task is executed on the mobile terminal, a first resource allocation result is solved by minimizing an objective function, wherein the first resource allocation result is a CPU frequency, and the calculation method of the CPU frequency comprises the following steps:
Figure FDA00029162233200000313
wherein the content of the first and second substances,
Figure FDA00029162233200000314
the optimal CPU frequency of the ith mobile terminal in the time slice t is represented; f. ofi,URepresents the maximum CPU frequency, f, that the mobile terminal can reachi,LA minimum CPU frequency indicating that the mobile terminal satisfies a task execution deadline,
Figure FDA0002916223320000041
is between fi,UAnd fi,LThe CPU frequency between the two;
by minimizing the objective function when the computing task is executed at the edge server
Figure FDA0002916223320000047
Solving a second resource allocation result, wherein the second resource allocation result is a network bandwidth allocation scheme, and the network bandwidth allocation scheme is as follows:
Figure FDA0002916223320000042
wherein, wi t*And solving an objective function to obtain the network bandwidth of the ith mobile terminal in the t time slice.
5. The method of claim 4, wherein the method further comprises:
constructing a return function for the mobile terminal executed on the edge server for the execution mode of the computing task, and adjusting the execution mode of the computing task generated by the mobile terminal based on the return function, wherein the return function is expressed as follows:
Figure FDA0002916223320000043
Figure FDA0002916223320000044
and
Figure FDA0002916223320000045
respectively representing the values of the drift-plus-penalty sub-function when the computing task of the ith mobile terminal in the time slice t is executed locally, is unloaded to the server and is discarded,
Figure FDA0002916223320000046
and the report function value of the ith mobile terminal in the time slice t is shown.
6. The method of claim 1, wherein the method further comprises:
and adjusting the calculation unloading decision and the resource allocation result by judging whether the value of the target function meets the preset condition or not.
7. A joint computation offload device based on energy harvesting technology, comprising:
the information collection module is used for collecting the real-time information of the edge server and the real-time information of all the mobile terminals when each preset time slice starts, wherein the real-time information of the mobile terminals comprises collected energy, calculation task arrival conditions and the positions of the mobile terminals;
the queue building module is used for building a virtual energy queue of each mobile terminal according to the collected energy, the arrival condition of the calculation task and the position of the mobile terminal;
the model construction module is used for constructing a task execution cost model, a terminal energy consumption model and an energy collection model;
the target function obtaining module is used for constructing a Lyapunov drift penalty function according to the real-time information of the edge server, the real-time information of all the mobile terminals, the constructed task execution cost model, the constructed terminal energy consumption model and the constructed energy collection model, and obtaining a target function according to the Lyapunov drift penalty function;
the unloading decision module is used for solving a calculation unloading decision in the current time slice by taking the virtual energy queue as a queue needing to be stabilized through a minimized objective function, wherein the calculation unloading decision is used for representing an execution mode of a calculation task generated by the mobile terminal;
the resource allocation module is used for allocating resources for different calculation unloading decisions based on the objective function;
the objective function obtaining module is specifically configured to:
constructing a Lyapunov function, which is specifically shown in formula (6):
Figure FDA0002916223320000051
wherein U represents a set of mobile terminals;
constructing a Lyapunov drift function, which is specifically shown in a formula (7):
Figure FDA0002916223320000052
wherein the content of the first and second substances,
Figure FDA0002916223320000053
expressing the expectation;
constructing a Lyapunov drift and penalty function, which is specifically shown in formula (8):
Figure FDA0002916223320000054
wherein, V represents a penalty factor, and N represents the number of elements in the mobile terminal set;
according to the form of Lyapunov drift plus penalty function, the upper bound of the function is obtained, and the specific formula is as follows:
Figure FDA0002916223320000055
wherein, C is a constant,
Figure FDA0002916223320000056
representing the consumed energy of the ith mobile terminal in the t time slice;
obtaining a target function according to the Lyapunov drift plus a penalty function
Figure FDA0002916223320000057
Figure FDA0002916223320000058
Wherein the content of the first and second substances,
Figure FDA0002916223320000059
a virtual energy queue representing the ith mobile terminal at time t,
Figure FDA00029162233200000510
representing the collected energy of the ith mobile terminal within the t time slice,
Figure FDA00029162233200000511
indicating the cost of execution of the ith mobile terminal at time t,
Figure FDA00029162233200000512
weighting factor, I, being the execution cost in the drift plus penalty functiont,ft,wtRespectively representing the calculation mode, the CPU frequency and the bandwidth size of the mobile terminal in the time slice t,
Figure FDA00029162233200000513
respectively representing the calculation mode, the CPU frequency and the bandwidth size of the ith mobile terminal in the time slice t.
8. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method according to any of claims 1 to 6 when executing the program.
CN201910015984.3A 2019-01-03 2019-01-03 Joint calculation unloading method and device based on energy collection technology Active CN109829332B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910015984.3A CN109829332B (en) 2019-01-03 2019-01-03 Joint calculation unloading method and device based on energy collection technology

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910015984.3A CN109829332B (en) 2019-01-03 2019-01-03 Joint calculation unloading method and device based on energy collection technology

Publications (2)

Publication Number Publication Date
CN109829332A CN109829332A (en) 2019-05-31
CN109829332B true CN109829332B (en) 2021-05-04

Family

ID=66861605

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910015984.3A Active CN109829332B (en) 2019-01-03 2019-01-03 Joint calculation unloading method and device based on energy collection technology

Country Status (1)

Country Link
CN (1) CN109829332B (en)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110533332B (en) * 2019-09-02 2023-05-02 海南电网有限责任公司 Computing resource allocation method and server based on multiple initial point penalty functions
CN110650487B (en) * 2019-09-27 2022-10-28 常熟理工学院 Internet of things edge computing configuration method based on data privacy protection
CN111104211A (en) * 2019-12-05 2020-05-05 山东师范大学 Task dependency based computation offload method, system, device and medium
CN111124639B (en) * 2019-12-11 2023-05-23 安徽大学 Operation method and system of edge computing system and electronic equipment
CN111163521B (en) * 2020-01-16 2022-05-03 重庆邮电大学 Resource allocation method in distributed heterogeneous environment in mobile edge computing
CN111479242A (en) * 2020-04-01 2020-07-31 重庆邮电大学 Task unloading method for assisting vehicle formation through fog calculation
CN111338807B (en) * 2020-05-21 2020-08-14 中国人民解放军国防科技大学 QoE (quality of experience) perception service enhancement method for edge artificial intelligence application
CN112988345B (en) * 2021-02-09 2024-04-02 江南大学 Dependency task unloading method and device based on mobile edge calculation
CN113064665B (en) * 2021-03-18 2022-08-30 四川大学 Multi-server computing unloading method based on Lyapunov optimization
CN113377447B (en) * 2021-05-28 2023-03-21 四川大学 Multi-user computing unloading method based on Lyapunov optimization
CN113626107B (en) * 2021-08-20 2024-03-26 中南大学 Mobile computing unloading method, system and storage medium
CN114461299B (en) * 2022-01-26 2023-06-06 中国联合网络通信集团有限公司 Unloading decision determining method and device, electronic equipment and storage medium
CN115278728A (en) * 2022-06-23 2022-11-01 重庆邮电大学 On-line joint control method for transmission power, modulation order and coding code rate in energy collection wireless communication system
CN115843070B (en) * 2023-02-23 2023-06-16 山东省计算中心（国家超级计算济南中心） Ocean sensing network calculation unloading method and system based on energy collection technology

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9554239B2 (en) * 2015-04-21 2017-01-24 Apple Inc. Opportunistic offloading of tasks between nearby computing devices
CN107682443A (en) * 2017-10-19 2018-02-09 北京工业大学 Joint considers the efficient discharging method of the mobile edge calculations system-computed task of delay and energy expenditure
CN107911478B (en) * 2017-12-06 2020-09-22 武汉理工大学 Multi-user calculation unloading method and device based on chemical reaction optimization algorithm

Also Published As

Publication number Publication date
CN109829332A (en) 2019-05-31

Similar Documents

Publication Publication Date Title
CN109829332B (en) Joint calculation unloading method and device based on energy collection technology
CN111835827B (en) Internet of things edge computing task unloading method and system
CN103220337B (en) Based on the cloud computing resources Optimal Configuration Method of self adaptation controller perturbation
CN112291793B (en) Resource allocation method and device of network access equipment
CN110570075B (en) Power business edge calculation task allocation method and device
CN110717300A (en) Edge calculation task allocation method for real-time online monitoring service of power internet of things
CN108304256B (en) Task scheduling method and device with low overhead in edge computing
Li et al. Method of resource estimation based on QoS in edge computing
CN114567895A (en) Method for realizing intelligent cooperation strategy of MEC server cluster
CN110968366A (en) Task unloading method, device and equipment based on limited MEC resources
Tian et al. User preference-based hierarchical offloading for collaborative cloud-edge computing
CN115629865B (en) Deep learning inference task scheduling method based on edge calculation
CN109639833A (en) A kind of method for scheduling task based on wireless MAN thin cloud load balancing
CN115421930B (en) Task processing method, system, device, equipment and computer readable storage medium
CN115473896A (en) Electric power internet of things unloading strategy and resource configuration optimization method based on DQN algorithm
Zheng et al. Learning based task offloading in digital twin empowered internet of vehicles
CN116963182A (en) Time delay optimal task unloading method and device, electronic equipment and storage medium
CN113766037B (en) Task unloading control method and system for large-scale edge computing system
CN113747450A (en) Service deployment method and device in mobile network and electronic equipment
CN115952054A (en) Simulation task resource management method, device, equipment and medium
CN113176936B (en) QoE-aware distributed edge task scheduling and resource management method and system
CN113553188A (en) Mobile edge calculation unloading method based on improved longicorn whisker algorithm
CN110570136B (en) Distribution range determining method, distribution range determining device, electronic equipment and storage medium
CN113485718B (en) Context-aware AIoT application program deployment method in edge cloud cooperative system
CN115051999B (en) Energy consumption optimal task unloading method, device and system based on cloud edge cooperation

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