CN115767637A - Cloud computing network resource optimal allocation method based on opportunistic access - Google Patents

Cloud computing network resource optimal allocation method based on opportunistic access Download PDF

Info

Publication number
CN115767637A
CN115767637A CN202211647220.4A CN202211647220A CN115767637A CN 115767637 A CN115767637 A CN 115767637A CN 202211647220 A CN202211647220 A CN 202211647220A CN 115767637 A CN115767637 A CN 115767637A
Authority
CN
China
Prior art keywords
task
user
cloud
node
fog node
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.)
Pending
Application number
CN202211647220.4A
Other languages
Chinese (zh)
Inventor
孙文彬
赵磊
张兆林
王伶
杨欣
许茜
韩闯
宫延云
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.)
Northwestern Polytechnical University
Original Assignee
Northwestern Polytechnical University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Northwestern Polytechnical University filed Critical Northwestern Polytechnical University
Priority to CN202211647220.4A priority Critical patent/CN115767637A/en
Publication of CN115767637A publication Critical patent/CN115767637A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Mobile Radio Communication Systems (AREA)

Abstract

The invention provides a cloud and mist computing network resource optimal allocation method based on opportunistic access, which belongs to the technical field of wireless network computing and solves the problems of a large amount of delay and computation in the existing cloud and mist computing network; the method comprises the following steps: s1, a user requests a calculation task, and the user accesses to a fog node through opportunity and transmits data; s2, carrying out cloud and mist computing network processing and transmitting a result to a user; s3, determining the adaptive resource allocation and calculation unloading optimization problem of the cloud and fog calculation network; s4, dividing the optimization problem into a plurality of suboptimal problems, including: a fog node selection problem, a power allocation problem, a bandwidth allocation problem, and a task offloading problem; s5, solving the suboptimal problem, and approximating the optimal solution of the optimization problem through a joint iterative algorithm to obtain an optimal scheme; the opportunistic access scheme of the invention can achieve performance similar to that of the traditional optimal communication scheme, and the opportunistic access scheme can greatly reduce delay and energy consumption caused by communication process.

Description

Cloud computing network resource optimal allocation method based on opportunistic access
Technical Field
The invention belongs to the technical field of wireless network computing, is applied to a cloud and mist computing network, and particularly relates to an opportunistic access-based cloud and mist computing network resource optimal allocation method.
Background
With the rapid development of wireless communication networks, the service types of users have changed greatly in a short time, and the development from simple wireless data transmission to complex multimedia services has progressed. In order to meet the multimedia service requirements of users, wireless communication networks now need to provide not only transmission services, but also comprehensive multimedia services such as caching and computing. The necessary components for realizing multimedia services are a Central Processing Unit (CPU) mainly used for calculation and various memories; however, due to the limitations of size, power and mobility, these hardware are difficult to be integrated into users for practical application, and therefore, in order to provide multimedia services under the constraint condition of mobile devices, cloud computing networks have been developed.
Although the energy efficiency and the resource utilization rate of the cloud computing network are high, with the rapid increase of the number of users, the load and the operating pressure of a central computing unit of the cloud computing network are increased sharply, resulting in significant performance loss of the cloud computing network; therefore, in order to solve the problem of performance loss due to an increase in task load caused by an increase in the number of users, the idea of fog calculation has been created.
In the prior art, in order to complete complex and large-scale computing tasks, a cloud and mist computing network combining mist computing and cloud computing is adopted. In a cloud and mist computing network, there is one remote central cloud node and a plurality of edge mist nodes, which together provide computing services to users in a collaborative manner.
In the current research on cloud and mist computing networks, a mist-cloud computing resource allocation scheme based on an Analytic Hierarchy Process (AHP) is available, and network load and computing load are considered simultaneously in a decision process so as to reduce delay caused by computing tasks; there are also cloud computing networks based on full duplex (FD-) and non-orthogonal multiple access (NOMA-). However, FD-and NOMA-based communication schemes introduce a significant amount of additional delay and computation in the cloud computing network due to complex communication channel estimation, channel State Information (CSI) feedback, and complex interference cancellation algorithms.
Disclosure of Invention
The invention provides an opportunistic access-based cloud computing network resource optimization allocation method, wherein a cloud computing network adopting an opportunistic access mode is referred to as OFCN for short; the method aims to further reduce delay and energy consumption of the cloud and mist computing network so as to improve the resource utilization rate as much as possible; opportunistic access is a potential communication scheme that does not require channel estimation and interference cancellation algorithms, and whose feedback information only includes the signal-to-noise ratio; therefore, when the number of mobile users and the number of fog nodes are large enough, the opportunistic access scheme can achieve the performance similar to the optimal communication scheme in the traditional research, and the use of the opportunistic access scheme can greatly reduce the delay and energy consumption caused by the communication process.
In order to achieve the purpose, in the technical scheme of the invention, the resource optimization allocation method is embodied by solving an optimization problem of allocating resources and unloading calculation tasks under the constraint of user QoS requirements, wherein delay and energy consumption are jointly weighted by two variables; and dividing the target optimization problem into 4 sub-optimal problems, and after solving the 4 sub-optimal problems respectively, approximating the optimal solution of the original target optimization problem by adopting an iterative algorithm according to the solution of the sub-optimal problem.
The technical scheme is as follows:
an opportunistic access-based cloud computing network resource optimal allocation method comprises the following steps:
s1, when a user requests a calculation task, connecting to a fog node of a cloud and fog calculation network in an opportunistic access mode, and wirelessly transmitting the calculation task and the QoS requirement of the user to the fog node;
s2, the cloud and fog computing network processes the computing task according to the QoS requirement of the user and then wirelessly transmits the processing result to the user;
s3, determining the adaptive resource allocation and calculation unloading optimization problem of the cloud computing network under the constraint of the QoS requirement of the user according to the contents in the step S1 and the step S2;
s4, dividing the optimization problem determined in the step S3 into a plurality of suboptimal problems, wherein the suboptimal problems comprise: a fog node selection problem, a power allocation problem, a bandwidth allocation problem, and a task offloading problem;
and S5, solving the plurality of suboptimal problems in the step S4, and approaching the optimal solution of the optimization problem through a joint iterative algorithm, wherein the optimal solution is the optimal self-adaptive resource allocation and calculation unloading scheme of the cloud and mist calculation network during wireless communication transmission in an opportunistic access mode.
Further, the step S2 specifically includes:
s2-1, the fog node receives a calculation task requested by a user, if the calculation task is within the self capacity range of the fog node, the fog node independently completes the calculation task and returns a processing result, and if not, the step S2-2 is carried out;
s2-2, if the fog node can not complete part of the computing task, unloading part of the computing task to a remote cloud node; for the part of the calculation task which can be completed by the fog node, the fog node calculates and obtains a calculation result of the fog node;
s2-3, after receiving part of the computing tasks, the remote cloud nodes feed back computing results of the remote cloud nodes to corresponding fog nodes;
and S2-4, the fog node receives the calculation result fed back by the remote cloud node, integrates the calculation result, and wirelessly transmits the processing result to the user after the processing result is obtained.
Further, in step S1, the u-th user requests a calculation task, and the calculation task and the corresponding user QoS requirement are expressed as
Figure BDA0004010208020000031
In the formula, xi u Indicating the number of partitionable task pieces, δ u Data size, beta, representing each task slice u Indicating the number of turns to complete each task slice,
Figure BDA0004010208020000032
represents the maximum tolerable delay of the computational task; in the opportunistic access mode, before the calculation task is wirelessly transmitted to the fog node, the task data is multiplied by a random preprocessing vector
Figure BDA0004010208020000033
In the formula, n u The nth antenna (1 is more than or equal to n) of the u user u ≤N u ) Random variable
Figure BDA0004010208020000034
Satisfy the requirement of
Figure BDA0004010208020000035
In which
Figure BDA0004010208020000036
And
Figure BDA0004010208020000037
denotes amplitude and phase, and is set | w u ‖ 2 =1; thus, the form of a wireless channel matrix between the f-th fog node and the u-th user is obtained as follows:
Figure BDA0004010208020000038
in the formula (I), the compound is shown in the specification,
Figure BDA0004010208020000039
n-th representing the u-th user and the f-th fog node f A wireless channel vector between the root antennas.
Further, in step S1, the calculation task and the QoS requirement of the user are transmitted through a wireless channel, and the received antenna array signals are combined by using a receive beam to form a combined signal; rate of combined signal
Figure BDA00040102080200000310
Wherein
Figure BDA00040102080200000311
N-th representing the u-th user and the f-th fog node f A wireless channel rate between root antennas; radio channel rate
Figure BDA00040102080200000312
Wherein B is u The frequency band occupied for the u-th user,
Figure BDA00040102080200000313
for the u-th user and the n-th user f Signal-to-noise ratio of the wireless channel between the root antennas; n th f The received signal of the root antenna is:
Figure BDA00040102080200000314
wherein x is u The transmission data representing the u-th user,
Figure BDA00040102080200000315
additive white gaussian noise representing the u-th user; thus, the equivalent channel is represented as:
Figure BDA00040102080200000316
the received signal is changed to:
Figure BDA0004010208020000041
further, in step S2-1 and step S2-2, A is used f Representing the computing power of the f-th fog node, the computing delay of the fog node
Figure BDA0004010208020000042
Wherein
Figure BDA0004010208020000043
The number of task pieces calculated for the f-th fog node,
Figure BDA0004010208020000044
calculated energy consumption of fog nodes
Figure BDA0004010208020000045
Wherein eta f The energy efficiency coefficient of the f-th fog node; offloading latency for remote cloud nodes
Figure BDA0004010208020000046
Wherein
Figure BDA0004010208020000047
Refers to the rate of the backbone link and,
Figure BDA0004010208020000048
is the number of the task slice computed by the remote cloud node; computing latency of remote cloud nodes
Figure BDA0004010208020000049
Wherein A is c Representing the computing power of a remote cloud node; computing energy consumption of remote cloud node
Figure BDA00040102080200000410
Figure BDA00040102080200000411
Wherein
Figure BDA00040102080200000412
For offloading power of tasks, eta c Is the energy efficiency coefficient of the remote cloud node.
Further, according to the specific transmission and calculation process data, the adaptive resource allocation and calculation offloading optimization problem of the cloud and mist computing network determined in step S3 is as follows:
Figure BDA00040102080200000413
the constraint conditions of the optimization problem are as follows:
c 1 ：
Figure BDA00040102080200000414
c 2 ：
Figure BDA00040102080200000415
c 3 ：
Figure BDA00040102080200000416
c 4 ：
Figure BDA00040102080200000417
c 5 ：
Figure BDA00040102080200000418
c 6 ：
Figure BDA00040102080200000419
c 7 ：
Figure BDA00040102080200000420
c 8 ：κ 1 + 2 ＝1.
c 9 ：κ 1 ,2＝0or1.
wherein, γ C For the maximum computing power of the remote cloud node,
Figure BDA00040102080200000421
and
Figure BDA00040102080200000422
representing the weight of latency and energy consumption in an opportunistic access connected cloud computing network.
Further, in step S4, the process of dividing the optimization problem into the fog node selection problem is as follows:
given the
Figure BDA00040102080200000423
Defining:
Figure BDA0004010208020000051
thus, the optimization problem is written as a fog node selection problem as follows:
Figure BDA0004010208020000052
the constraint conditions of the fog node selection problem are as follows:
c 10 ：
Figure BDA0004010208020000053
c 11 ：
Figure BDA0004010208020000054
c 12 ：b u, ＝0or1.
and finishing the division of the fog node selection problem.
Further, in step S4, the process of dividing the optimization problem into the power allocation problem is as follows: given the
Figure BDA0004010208020000055
Writing the optimization problem as a power allocation problem as follows:
Figure BDA0004010208020000056
the constraints of the power allocation problem are:
c 13 ：
Figure BDA0004010208020000057
the partition of the power allocation problem is completed.
Further, in step S4, the process of dividing the optimization problem into the bandwidth allocation problem is as follows: given the
Figure BDA0004010208020000058
Writing the optimization problem as a bandwidth allocation problem as follows:
Figure BDA0004010208020000059
the constraints of the bandwidth allocation problem are:
c 14 ：
Figure BDA0004010208020000061
the partitioning of the bandwidth allocation problem is done.
Further, in step S4, the process of dividing the optimization problem into task offloading problems is as follows:
given { b u,f ,p u ,B u }, defining:
Figure BDA0004010208020000062
thus, the optimization problem is written as a task offload problem, as follows:
Figure BDA0004010208020000063
the constraints of the task offloading problem are:
c 15 ：
Figure BDA0004010208020000064
c 16 ：
Figure BDA0004010208020000065
c 17 ：
Figure BDA0004010208020000066
c 18 ：
Figure BDA0004010208020000067
c 19 ：κ 1 ,κ2＝0or1.
and finishing the division of the task unloading problem.
In summary, due to the adoption of the technical scheme, the invention has the following beneficial effects:
compared with the conventional OMA and NOMA cloud computing network, the OFCN of the invention adopts the cloud computing network of the opportunistic access mode, and the scheme of the invention is convergent in convergence performance, and the invention can realize the minimum system total cost.
Different from different preprocessing weight generation methods (optimal preprocessing OP, grassmann subspace filling GPS and genetic algorithm GA) of wireless access, the method of the invention is adopted to complete the solution of the optimization problem, and the obtained OFCN total system overhead is obtained.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention;
FIG. 2 is a schematic diagram of the OFCN architecture;
FIG. 3 is a diagram illustrating the transmission mechanism of OFCN;
FIG. 4 is a diagram illustrating a comparison of convergence performance for different access schemes;
fig. 5 is a schematic diagram showing a comparison of total system costs of methods for generating different preprocessing weights for wireless access.
Detailed Description
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. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the 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.
Referring to the method flow illustration of fig. 1, a cloud computing network resource optimal allocation method based on opportunistic access includes the following steps:
s1, when a user requests a calculation task, connecting to a fog node of a cloud and fog calculation network in an opportunistic access mode, and wirelessly transmitting the calculation task and the QoS requirement of the user to the fog node;
s2, the cloud and fog computing network processes the computing task according to the QoS requirement of the user and then wirelessly transmits the processing result to the user;
s3, determining the adaptive resource allocation and calculation unloading optimization problem of the cloud computing network under the constraint of the QoS requirement of the user according to the contents in the step S1 and the step S2;
s4, dividing the optimization problem determined in the step S3 into a plurality of suboptimal problems, wherein the suboptimal problems comprise: a fog node selection problem, a power allocation problem, a bandwidth allocation problem, and a task offloading problem;
and S5, solving the plurality of suboptimal problems in the step S4, and approaching the optimal solution of the optimization problem through a joint iterative algorithm, wherein the optimal solution is the optimal adaptive resource allocation and calculation unloading scheme of the cloud and mist computing network during wireless communication transmission in the opportunistic access mode.
In this example, use is made ofThe cloud computing network with the opportunistic access mode is abbreviated as OFCN, and the architecture of the OFCN please refer to the illustration in fig. 2; the OFCN has three layers, namely a user layer, a fog layer and a remote cloud layer. And (3) a user layer: u (U is more than or equal to 1) users are randomly distributed in different places, and each user is configured with N u (N u Not less than 1) antenna. It is assumed that the user is only transmitting and receiving wireless signals and has no ability to perform computational tasks. The users use different frequency bands to transmit data of tasks, so the OFCN ignores interference between users. User layer radius is denoted as r u . Fog layer: in order to realize the multi-node selection gain of opportunistic access, F (F is more than or equal to U) fog nodes are assumed in the fog layer. The fog node positions obey two-dimensional Poisson distribution, and the influence of the fog node heights is not considered. The radius of the haze is denoted as r f Wherein r is f Far greater than r u . Each fog node has an antenna array and a small computational processing unit. Configuring N in the F (1 ≤ F) th fog node antenna array f (N f ≧ 1) an antenna that transmits and receives wireless signals. The remote cloud includes a high performance computing center and a mass storage center. The high-performance computing center can complete computing tasks sent by the fog nodes in real time, and computing results are fed back to the fog nodes through backbone links.
Under the above OFCN architecture, the OFCN transmission mechanism is shown in fig. 3, and step S2 specifically includes:
s2-1, the fog node receives a calculation task requested by a user, if the calculation task is within the self capacity range of the fog node, the fog node independently completes the calculation task and returns a processing result, and if not, the step S2-2 is carried out;
s2-2, if the fog node can not complete part of the computing task, unloading part of the computing task to a remote cloud node; for the part which can be completed by the fog node in the calculation task, the fog node calculates and obtains the calculation result of the fog node;
s2-3, after receiving part of the computing tasks, the remote cloud nodes feed back computing results of the remote cloud nodes to corresponding fog nodes;
and S2-4, the fog node receives the calculation result fed back by the remote cloud node, integrates the calculation result, and wirelessly transmits the processing result to the user after the processing result is obtained.
The following detailed description of the specific method and optimization problem process adopted in the present embodiment is provided, and before that, some basic expressions and definitions need to be explained; setting X, X and X to respectively represent variables, vectors and matrixes; x is a radical of a fluorine atom * Represents the conjugate of variable x; x is the number of T ，X T Respectively representing the transpose of vector X and matrix X; x is the number of H ，X H Respectively representing the conjugate transposes of vector X and matrix X; ii denotes the Frobenius norm of the matrix or vector, | represents the absolute value of the variable,
Figure BDA0004010208020000081
denotes a minimum integer greater than x, [ x]Representing the smallest integer less than x.
In step S1, the u-th user requests a computing task and is connected to the fog node in an opportunistic access mode. The computational tasks for the u-th user and the corresponding user QoS requirements are expressed as
Figure BDA0004010208020000091
In the formula, xi u Indicating the number of partitionable task pieces, δ u Data size, beta, representing each task slice u Indicating the number of turns to complete each task slice,
Figure BDA0004010208020000092
representing the maximum tolerable delay of a computational task, given xi u ,δ u ,β u ,
Figure BDA0004010208020000093
There are no error values on both the user node and the fog node; in an opportunistic access scheme, the user increases the transmission rate of the radio channel using opportunistic preprocessing in which the data of a task is multiplied by a random preprocessing vector W prior to transmission u Thereby reducing the delay and the energy consumption,
Figure BDA0004010208020000094
in the formula, n u The nth antenna (1 is more than or equal to n) of the u user u ≤N u ) Random variable
Figure BDA0004010208020000095
Satisfy the requirement of
Figure BDA0004010208020000096
In which
Figure BDA0004010208020000097
And
Figure BDA0004010208020000098
representing amplitude and phase, and is set | w u ‖ 2 =1; thus, the form of a wireless channel matrix between the f-th fog node and the u-th user is obtained as follows:
Figure BDA0004010208020000099
in the formula (I), the compound is shown in the specification,
Figure BDA00040102080200000910
n-th representing the u-th user and the f-th fog node f A wireless channel vector between the root antennas.
The computational tasks and the QoS requirements of the user are transmitted over a wireless channel to the connected fog node. Combining received antenna array signals using receive beamforming to combine the rate R of the signals u,f Is composed of
Figure BDA00040102080200000911
In the formula (I), the compound is shown in the specification,
Figure BDA00040102080200000912
n-th representing the u-th user and the f-th fog node f The radio channel rate between the root antennas. The radio channel rate is
Figure BDA00040102080200000913
B u The frequency band occupied for the u-th user,
Figure BDA00040102080200000914
for the u-th user and the n-th user f Signal-to-noise ratio of the wireless channel between the root antennas. N th f The received signal of the root antenna is:
Figure BDA00040102080200000915
x u and
Figure BDA00040102080200000916
respectively, the transmission data of the u-th user and Additive White Gaussian Noise (AWGN). The equivalent channel can be expressed as:
Figure BDA00040102080200000917
the received signal is changed into:
Figure BDA00040102080200000918
in the specific step S2-1 and the step S2-2, the fog node receives task data of the user and completes the task within the capability range of the fog node. If the fog node cannot independently complete the task, the task will be offloaded to the remote cloud node. With A f Representing the computational capability of the f-th fog node, the delay of the fog computation being represented by
Figure BDA0004010208020000101
It is determined that the user is to be,
Figure BDA0004010208020000102
the number of task pieces calculated for the f-th fog node is satisfied
Figure BDA0004010208020000103
The energy consumption for the mist calculation was:
Figure BDA0004010208020000104
Figure BDA0004010208020000105
η f the energy efficiency coefficient of the f-th fog node. The latency of cloud computing includes offload latency and processing latency. Offload delay
Figure BDA0004010208020000106
By backbone links
Figure BDA0004010208020000107
Is determined by the volume of the unloading task, the unloading delay is:
Figure BDA0004010208020000108
Figure BDA0004010208020000109
is the number of task slices computed by the remote cloud node. The computing latency of cloud computing is:
Figure BDA00040102080200001010
A c representing the computing power of the remote cloud node. The energy consumption of cloud computing is:
Figure BDA00040102080200001011
Figure BDA00040102080200001012
to off-load the power of the task, η c Is the energy efficiency coefficient of the remote cloud node.
The cloud node completes the received partial tasks and feeds the result back to the fog node through the backbone link. The backhaul links in the OFCN architecture include links from the cloud to the fog nodes and from the fog nodes to the users. And the backhaul link is used for transmitting computing task data and results between the cloud node and the fog node. The transmission rate and energy efficiency of the backbone link are much greater than those of the radio access link. Furthermore, the latency and energy consumption of the backbone link are difficult to adjust compared to the radio access link. Therefore, the network scheme provided by the embodiment ignores the time delay and the energy consumption of cloud and mist communication.
And the fog node integrates the results of the original calculation tasks and transmits the integrated results to the u-th user.
In this embodiment, according to the above specific contents, a specific optimization problem of accessing the cloud and mist computing network at an opportunity can be proposed to realize adaptive resource allocation and computation offloading under the user QoS constraint, that is, the specific contents of step S3. The optimization problem is as follows:
Figure BDA00040102080200001013
the constraint conditions of the optimization problem are as follows:
c 1 ：
Figure BDA00040102080200001014
c 2 ：
Figure BDA00040102080200001015
c 3 ：
Figure BDA00040102080200001016
c 4 ：
Figure BDA00040102080200001017
c 5 ：
Figure BDA00040102080200001018
c 6 ：
Figure BDA0004010208020000111
c 7 ：
Figure BDA0004010208020000112
c 8 ：κ 1 + 2 ＝1.
c 9 ：κ 1 ,2＝0or1.
wherein, γ C For the maximum computing power of the remote cloud node,
Figure BDA0004010208020000113
and
Figure BDA0004010208020000114
weights representing delay and energy consumption in an opportunistic access connected cloud computing network.
Next, the optimization problem in step S3 is divided into 4 sub-optimal problems, first the fog node selection problem, given
Figure BDA0004010208020000115
Defining:
Figure BDA0004010208020000116
thus, the optimization problem is written as a fog node selection problem, as follows:
Figure BDA0004010208020000117
the constraint conditions of the fog node selection problem are as follows:
c 10 ：
Figure BDA0004010208020000118
c 11 ：
Figure BDA0004010208020000119
c 12 ：b u, ＝0or1.
and finishing the division of the fog node selection problem.
And continuing to divide, wherein the process of dividing the optimization problem into the power distribution problem comprises the following steps:
given a
Figure BDA00040102080200001110
Writing the optimization problem as a power allocation problem as follows:
Figure BDA00040102080200001111
the constraints of the power allocation problem are:
c 13 ：
Figure BDA0004010208020000121
the partition of the power allocation problem is completed.
And continuing to divide, wherein the process of dividing the optimization problem into the bandwidth allocation problem comprises the following steps:
given a
Figure BDA0004010208020000122
Writing the optimization problem as a bandwidth allocation problem as follows:
Figure BDA0004010208020000123
the constraints of the bandwidth allocation problem are:
c 14 ：
Figure BDA0004010208020000124
the partitioning of the bandwidth allocation problem is done.
Finally, the process of dividing the optimization problem into task offloading problems is:
given { b u,f ,p u ,B u Define:
Figure BDA0004010208020000125
thus, the optimization problem is written as a task offload problem, as follows:
Figure BDA0004010208020000126
the constraint conditions of the task uninstalling problem are as follows:
c 15 ：
Figure BDA0004010208020000127
c 16 ：
Figure BDA0004010208020000128
c 17 ：
Figure BDA0004010208020000129
c 18 ：
Figure BDA0004010208020000132
c 19 ：κ 1 ,κ 2 ＝0or1.
and finishing the division of the task unloading problem.
Through the mode, the division of four suboptimal problems is completed. The solution is then performed by a joint iterative algorithm. The joint iteration algorithm comprises the following five contents:
(1) Inputting and initializing algorithm parameters;
(2) Respectively searching a global solution by using an outer layer, an adjacent inner layer and an inner layer four-layer iterative algorithm;
(3) Calculating the solution of the adaptive resource allocation and calculation unloading optimization problem in the step S3;
(4) The objective function values of the current iteration and the previous iteration are compared. Recording the value of the objective function and the corresponding parameters
Figure BDA0004010208020000131
p u ，B u ；
(5) And (4) repeating the steps (1) to (4) until the maximum iteration times is reached, and outputting the minimum objective function value and corresponding parameters to obtain an optimal solution, so that an optimal self-adaptive resource allocation and calculation unloading scheme is found when the cloud computing network carries out wireless communication transmission in an opportunistic access mode.
Referring to fig. 4, in comparison of convergence performance between the method of the present embodiment and other conventional methods, it can be seen that the total system loss in the method of the present embodiment is significantly smaller than that in the other conventional methods under the same iteration number.
Referring to fig. 5, the method of the present embodiment is also significantly better than the other conventional methods in comparison with the total system cost of the method for generating different preprocessing weights for wireless access.

Claims (10)

1. An opportunistic access-based cloud computing network resource optimal allocation method is characterized by comprising the following steps:
s1, when a user requests a calculation task, connecting to a fog node of a cloud and fog calculation network in an opportunistic access mode, and wirelessly transmitting the calculation task and the QoS requirement of the user to the fog node;
s2, the cloud and fog computing network processes the computing task according to the QoS requirement of the user and then wirelessly transmits the processing result to the user;
s3, determining the adaptive resource allocation and calculation unloading optimization problem of the cloud and mist calculation network under the constraint of the QoS requirement of the user according to the contents in the step S1 and the step S2;
s4, dividing the optimization problem determined in the step S3 into a plurality of suboptimal problems, wherein the suboptimal problems comprise: a fog node selection problem, a power allocation problem, a bandwidth allocation problem, and a task offloading problem;
and S5, solving the plurality of suboptimal problems in the step S4, and approaching the optimal solution of the optimization problem through a joint iterative algorithm, wherein the optimal solution is the optimal self-adaptive resource allocation and calculation unloading scheme of the cloud and mist calculation network during wireless communication transmission in an opportunistic access mode.
2. The method for optimized allocation of cloud computing network resources based on opportunistic access according to claim 1, wherein the step S2 specifically comprises:
s2-1, the fog node receives a calculation task requested by a user, if the calculation task is within the self capacity range of the fog node, the fog node independently completes the calculation task and returns a processing result, and if not, the step S2-2 is carried out;
s2-2, if the fog node cannot complete part of the computing task, unloading part of the computing task to a remote cloud node; for the part of the calculation task which can be completed by the fog node, the fog node calculates and obtains a calculation result of the fog node;
s2-3, the remote cloud node receives part of the computing task, and after computing is completed, the computing result of the remote cloud node is fed back to the corresponding fog node;
and S2-4, the fog node receives the calculation result fed back by the remote cloud node, integrates the calculation result, and wirelessly transmits the processing result to the user after the processing result is obtained.
3. The method for optimizing and allocating the resources of the cloud computing network based on the opportunistic access as claimed in claim 2, wherein: in step S1, the u-th user requests a calculation task, the calculation task and the corresponding user QoS requirement are expressed as
Figure FDA0004010208010000011
In the formula, xi u Indicating the number of partitionable task pieces, δ u Data size, beta, representing each task slice u Indicating the number of turns to complete each task slice,
Figure FDA0004010208010000012
represents the maximum tolerable delay of the computational task; in the opportunistic access mode, before a calculation task is wirelessly transmitted to a fog node, task data is multiplied by a random preprocessing vector
Figure FDA0004010208010000021
In the formula, n u The nth antenna (1 is more than or equal to n) of the u user u ≤N u ) Random variable
Figure FDA0004010208010000022
Satisfy the requirement of
Figure FDA0004010208010000023
In which
Figure FDA0004010208010000024
And
Figure FDA0004010208010000025
representing amplitude and phase, and is set | w u ‖ 2 =1; thus, the form of a wireless channel matrix between the f-th fog node and the u-th user is obtained as follows:
Figure FDA0004010208010000026
in the formula (I), the compound is shown in the specification,
Figure FDA0004010208010000027
n-th representing the u-th user and the f-th fog node f A wireless channel vector between the root antennas.
4. The method for optimizing and allocating the resources of the cloud computing network based on the opportunistic access as claimed in claim 3, wherein: in the step S1, a calculation task and a user QoS requirement are transmitted through a wireless channel, and received antenna array signals are combined by adopting a receiving beam to form a combined signal; rate of combined signal
Figure FDA0004010208010000028
Wherein
Figure FDA0004010208010000029
N-th representing the u-th user and the f-th fog node f A wireless channel rate between root antennas; radio channel rate
Figure FDA00040102080100000210
Wherein B is u The frequency band occupied for the u-th user,
Figure FDA00040102080100000211
for the u-th user and the n-th user f Signal-to-noise ratio of the wireless channel between the root antennas; n th f The received signal of the root antenna is:
Figure FDA00040102080100000212
wherein x u The transmission data representing the u-th user,
Figure FDA00040102080100000213
additive white gaussian noise representing the u-th user; thus, the equivalent channel is represented as:
Figure FDA00040102080100000214
the received signal is changed to:
Figure FDA00040102080100000215
5. the method for optimizing and allocating the resources of the cloud computing network based on the opportunistic access as claimed in claim 4, wherein: in step S2-1 and step S2-2, A is used f Indicating the computing power of the f-th fog node, the computing delay of the fog node
Figure FDA00040102080100000216
Wherein
Figure FDA00040102080100000217
The number of task pieces calculated for the f-th fog node,
Figure FDA00040102080100000218
computational energy consumption of fog nodes
Figure FDA00040102080100000219
Wherein eta f The energy efficiency coefficient of the f-th fog node; offloading latency for remote cloud nodes
Figure FDA00040102080100000220
Wherein
Figure FDA00040102080100000221
Refers to the rate of the backbone link and,
Figure FDA00040102080100000222
is the number of the task piece computed by the remote cloud node; computing latency of remote cloud nodes
Figure FDA00040102080100000223
Wherein A is c Representing the computing power of a remote cloud node; computing energy consumption of remote cloud node
Figure FDA0004010208010000031
Figure FDA0004010208010000032
Wherein
Figure FDA0004010208010000033
To off-load the power of the task, η c Is the energy efficiency coefficient of the remote cloud node.
6. The method for optimal allocation of cloud computing network resources based on opportunistic access as claimed in claim 5, wherein the cloud computing network adaptive resource allocation and computation offload optimization problem determined in step S3 is as follows, according to specific transmission and computation process data:
Figure FDA0004010208010000034
the constraint conditions of the optimization problem are as follows:
c 1 ：
Figure FDA0004010208010000035
c 2 ：
Figure FDA0004010208010000036
c 3 ：
Figure FDA0004010208010000037
c 4 ：
Figure FDA0004010208010000038
c 5 ：
Figure FDA0004010208010000039
c 6 ：
Figure FDA00040102080100000310
c 7 ：
Figure FDA00040102080100000311
c 8 ：k 1 +k 2 ＝1.
c 9 ：κ 1 ,k 2 ＝0or1.
wherein, γ C For the maximum computing power of the remote cloud node,
Figure FDA00040102080100000312
and
Figure FDA00040102080100000313
weights representing delay and energy consumption in an opportunistic access connected cloud computing network.
7. The method for optimal allocation of cloud computing network resources based on opportunistic access as claimed in claim 6, wherein in step S4, the process of dividing the optimization problem into the cloud node selection problem is:
given the
Figure FDA00040102080100000314
Defining:
Figure FDA00040102080100000315
thus, the optimization problem is written as a fog node selection problem as follows:
Figure FDA00040102080100000316
the constraint conditions of the fog node selection problem are as follows:
c 10 ：
Figure FDA0004010208010000041
c 11 ：
Figure FDA0004010208010000042
c 12 ：b u,f ＝0or1.
and finishing the division of the fog node selection problem.
8. The method for cloud computing network resource optimal allocation based on opportunistic access as claimed in claim 6, wherein in step S4, the process of dividing the optimization problem into power allocation problems is as follows:
given the
Figure FDA0004010208010000043
Writing the optimization problem as a power allocation problem as follows:
Figure FDA0004010208010000044
the constraints of the power allocation problem are:
c 13 ：
Figure FDA0004010208010000045
the partition of the power allocation problem is completed.
9. The method for cloud computing network resource optimal allocation based on opportunistic access as claimed in claim 6, wherein in step S4, the process of dividing the optimization problem into bandwidth allocation problems is as follows:
given the
Figure FDA0004010208010000046
Writing the optimization problem as a bandwidth allocation problem as follows:
Figure FDA0004010208010000047
the constraints of the bandwidth allocation problem are:
c 14 ：
Figure FDA0004010208010000048
the partitioning of the bandwidth allocation problem is done.
10. The method for cloud computing network resource optimal allocation based on opportunistic access according to claim 6, wherein in step S4, the process of dividing the optimization problem into task offloading problems is:
given { b u,f ,p u ,B u Define:
Figure FDA0004010208010000051
thus, the optimization problem is written as a task offload problem, as follows:
Figure FDA0004010208010000052
the constraint conditions of the task uninstalling problem are as follows:
c 15 ：
Figure FDA0004010208010000053
c 16 ：
Figure FDA0004010208010000054
c 17 ：
Figure FDA0004010208010000055
c 18 ：
Figure FDA0004010208010000056
c 19 ：κ 1 ，κ 2 ＝0or1.
and finishing the division of the task unloading problem.
CN202211647220.4A 2022-12-21 2022-12-21 Cloud computing network resource optimal allocation method based on opportunistic access Pending CN115767637A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211647220.4A CN115767637A (en) 2022-12-21 2022-12-21 Cloud computing network resource optimal allocation method based on opportunistic access

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211647220.4A CN115767637A (en) 2022-12-21 2022-12-21 Cloud computing network resource optimal allocation method based on opportunistic access

Publications (1)

Publication Number Publication Date
CN115767637A true CN115767637A (en) 2023-03-07

Family

ID=85346925

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211647220.4A Pending CN115767637A (en) 2022-12-21 2022-12-21 Cloud computing network resource optimal allocation method based on opportunistic access

Country Status (1)

Country Link
CN (1) CN115767637A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116166341A (en) * 2023-04-25 2023-05-26 中国人民解放军军事科学院系统工程研究院 Static cloud edge collaborative architecture function calculation unloading method based on deep learning

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116166341A (en) * 2023-04-25 2023-05-26 中国人民解放军军事科学院系统工程研究院 Static cloud edge collaborative architecture function calculation unloading method based on deep learning

Similar Documents

Publication Publication Date Title
CN109756874B (en) Ultra-dense millimeter wave D2D communication interference management method
CN102457951B (en) Method for forming link combined wave beam in multi-cell collaborative communication, and base station
US7983710B2 (en) Method of coordinated wireless downlink transmission
CN114051254B (en) Green cloud edge collaborative computing unloading method based on star-ground fusion network
CN113644964A (en) Multi-dimensional resource joint allocation method of multi-beam satellite same-frequency networking system
CN111555795A (en) Network resource allocation method and device based on user grouping scheduling
KR102072285B1 (en) Method and apparatus for controlling interference for uplink and downlink in multi-tier heterogeneous network with in-band wireless backhaul
CN115767637A (en) Cloud computing network resource optimal allocation method based on opportunistic access
CN109687896A (en) The sparse Beamforming Method of distributing antenna system down channel and device
CN106231665A (en) Resource allocation methods based on the switching of RRH dynamic mode in number energy integrated network
CN113973204A (en) Nonlinear precoding method and system of MIMO-NOMA system
CN116600344A (en) Multi-layer MEC resource unloading method with power cost difference
CN114465643B (en) Mixed precoding method of millimeter wave large-scale MIMO antenna system based on gradient descent method
CN113541768A (en) NOMA-based LEO satellite communication system frequency point allocation method
CN117336790A (en) Task time delay optimization method in NOMA auxiliary millimeter wave mobile edge computing system based on intelligent reflecting surface
CN112770398A (en) Far-end radio frequency end power control method based on convolutional neural network
CN109831759B (en) Three-dimensional D2D matching algorithm based on software defined wireless network
CN115988536A (en) Dual-reconfigurable intelligent surface assisted mobile edge computing system optimization method
CN113258985B (en) Energy efficiency optimization method for single-station multi-satellite MIMO (multiple input multiple output) upper injection system
CN115361052A (en) Satellite beam forming method and device considering energy efficiency and spectral efficiency
CN113242602B (en) Millimeter wave large-scale MIMO-NOMA system resource allocation method and system
CN112584478B (en) Uplink energy efficiency minimization power distribution method based on NOMA-MEC
CN115065384A (en) Multi-beam satellite communication system resource allocation method considering user association, sub-channel allocation and beam association
CN105162505B (en) Cellular network down link self adaption method for precoding based on interference alignment
Al-Habob et al. Latency minimization in phase-coupled STAR-RIS assisted multi-MEC server systems

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