CN110659034B - Combined optimization deployment method, system and storage medium of cloud-edge hybrid computing service - Google Patents
Combined optimization deployment method, system and storage medium of cloud-edge hybrid computing service Download PDFInfo
- Publication number
- CN110659034B CN110659034B CN201910904428.1A CN201910904428A CN110659034B CN 110659034 B CN110659034 B CN 110659034B CN 201910904428 A CN201910904428 A CN 201910904428A CN 110659034 B CN110659034 B CN 110659034B
- Authority
- CN
- China
- Prior art keywords
- solution
- task
- gene
- value
- candidate
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/60—Software deployment
Abstract
The embodiment of the invention provides a combination optimization deployment method, a combination optimization deployment system and a storage medium of a cloud-side hybrid computing service, and belongs to the technical field of optimization deployment of the cloud-side hybrid computing service. The combined optimization deployment method, the system and the storage medium of the cloud-side hybrid computing service determine the optimal deployment scheme of the cloud-side hybrid computing service by combining various factors including cloud service satisfaction and service consistency except service quality expression and adopting an improved differential evolution algorithm, solve the technical problem that the deployment scheme cannot reach the optimal degree due to the fact that only single service quality factor is combined when the scheme is determined in the prior art, and enable the finally formed deployment scheme to meet the requirements of the cloud-side hybrid computing service to the maximum degree.
Description
Technical Field
The invention relates to the technical field of optimized deployment of cloud-side hybrid computing services, in particular to a combined optimized deployment method, a combined optimized deployment system and a storage medium of the cloud-side hybrid computing services.
Background
Under the support of a 5G network, a large number of devices can be accessed to the Internet of things to realize the interconnection of everything, wherein the edge computing service is a computing service closer to the object and the data end, so that the requirement of connecting massive heterogeneous devices can be met, and the digital agile connection, real-time service and application intelligence can be realized. In the implementation process, the computing services which need global, non-real-time and long-period big data processing and analysis to support complex decisions need to be supported by a cloud computing service center. Meanwhile, candidate computing services based on the characteristics of the enterprises are possibly built on the task that some enterprises are in safe privacy and special computing. Therefore, based on the above consideration, how to mix the computing services of the above feature types in the implementation process, the combined optimization deployment is an important basis for realizing future production and living computing services of everything interconnection, and meanwhile, the computing services with cloud edge cooperation can better fulfill target requirements, amplify data intelligent application values and reduce computing service cost.
Certain problems are also caused in the environment of mixing edge computing, cloud computing and enterprise private computing services, and the following aspects are particularly shown:
(1) each computing service needs to meet the requirement of basic QoS of a task, and the combination of the whole computing services needs to meet the requirement of basic QoS in the whole implementation process and can provide better computing service quality;
(2) in addition to a large number of edge calculations needing strong real-time agility in the overall implementation process, some tasks with special requirements need cloud computing with higher computing capacity, storage capacity and data analysis capacity to provide services, so that the special requirements need to be effectively considered in the combined optimization;
(3) compatibility problems are caused by various service combinations, particularly, compatibility is complicated due to different operating systems, service interfaces and computing modes of various services such as edge computing, Cloud computing and private computing, and TOSCA (Kim D, Muhammad H, Kim E, et al. TOSCA-Based and Federation-Aware Cloud architecture for Kuberneters Platform [ J ]. Applied Sciences-base, 2019,9(1)) and Container technologies such as Docker and Kubernetes are often adopted in the Cloud service combinations to process the technical problems of compatibility between different services. However, as the cloud-edge hybrid computing is still in a preliminary stage, and the corresponding technical method is not mature, a more general measurement method needs to be designed for service compatibility. (4) For the overall implementation process, according to different connection modes (S.Deng, H.Wu, et al.service selection for composition with QoS correlation [ J ]. IEEE Transactions on Services Computing,2016,9(2): 291-;
(5) with the increase of edge service computing devices in the market and the addition of different cloud computing service providers, more factors such as service types and service brand characteristics need to be considered for optimal selection of computing services;
(6) as the model of the computing service combination is a combination optimization model and is an NP-hard problem in algorithm, a heuristic evolutionary algorithm is adopted for iterative solution. In the research Of cloud service combination, many calculation methods have been implemented by using genetic algorithm (Lin Y-K, Chong CS. fast GA-based project scheduling for calculating resources allocation in a closed Manufacturing system [ J ]. Journal Of Intelligent Manufacturing,2017,28(5): 1189-.
Disclosure of Invention
The embodiment of the invention aims to provide a combination optimization deployment method, a combination optimization deployment system and a storage medium for cloud-side hybrid computing services, which can overcome the defects in the prior art and efficiently and reasonably plan the optimal deployment scheme of the cloud-side hybrid computing services.
In order to achieve the above object, an embodiment of the present invention provides a combinatorial optimization deployment method for a cloud-edge hybrid computing service, where the combinatorial optimization deployment method includes:
acquiring a task package of the cloud-edge hybrid computing service, wherein the task package comprises a common task set which needs to be completed by adopting an edge computing service and a special task set which needs to adopt a cloud computing service, the common task set and the special task set both comprise at least one task, and each task is connected with each other through a Petri network structure;
determining a set of candidate computing services capable of completing each of the tasks, wherein each task corresponds to a set of candidate computing services for completing the task, each set of candidate computing services comprises at least one candidate computing service, each candidate computing service comprises a first feature attribute and a second feature attribute, the first feature attribute comprises a type of the candidate computing service and a brand of a device for completing the candidate computing service, and the second feature attribute comprises a quality of service for completing the candidate computing service;
adopting the formula (1) as a preset objective function,
Fit(X)＝ω 1 ·TQoS+ω 2 ·TSR+ω 3 ·TSC，(1)
wherein Fit (X) is the objective function, ω, of deployment scenario X 1 、ω 2 And ω 3 For the preset weight, TQoS represents the total QoS, and TSR represents the public service qualityThe special computation service calculated by equation (2) satisfies the degree value,
wherein TSR satisfies a degree value, N, for the special computing service SRST Number of tasks, X, included for said set of special tasks SRST Is the number of tasks in the set of special tasks that the deployment scenario for completing the task package can complete, γ SRST Is a preset accumulated intensity value, and gamma SRST Not less than 1; TSC is the service consistency degree value calculated using equation (3),
wherein TSC is the service consistency degree value X B＝b Selecting the number of candidate computing services of brand b in a deployment scheme for completing the task package, wherein n is the number of tasks included in the task package;
and determining an optimal deployment scheme for completing the task package according to the task package, the candidate computing service set and the objective function by adopting a differential evolution algorithm.
Optionally, the combinatorial-optimization deployment method further includes:
presetting boundary conditions;
the step of determining the optimal deployment scheme for completing the task package according to the task package, the candidate computing service set and the objective function by adopting a differential evolution algorithm comprises the following steps:
and determining an optimal deployment scheme for completing the task package according to the task package, the candidate computing service set, the objective function and the boundary condition.
Optionally, the boundary condition comprises formula (4) and formula (5),
wherein, the first and the second end of the pipe are connected with each other,
to complete task T i A normalized index value of the time index Ind,
to complete task T i The minimum required normalized index value of the time index Ind, SQ Ind,P To complete the normalized index value of the index Ind at the time of the task package P,
the minimum required standardized index value of the index Ind when the task packet P is completed.
Optionally, the combinatorial-optimization deployment method further includes:
calculating the total quality of service (QoS) performance value according to formula (6),
TQoS＝∑ Ind (ω Ind /sum(ω Ind ))SQ Ind,P ，(6)
wherein, ω is Ind Weight, SQ, of an indicator Ind of the quality of service of the candidate calculation service Ind,P The index value is a standardized index value of the index Ind when the task packet P is completed.
Optionally, the determining an optimal deployment scenario for completing the task package according to the task package, the candidate computing service set, the objective function, and the boundary condition includes:
traversing each task in the task package, and eliminating the candidate computing services which do not meet the boundary condition in the candidate computing service set corresponding to the task;
randomly generating a plurality of solution genes X by adopting an integer coding mode to form an initial population, wherein,
to complete task T i The selected candidate computing service;
calculating an adaptation value of each solution gene of the initial population according to formula (1);
setting a current solution gene population, a historical optimal solution gene and a candidate optimal solution set, wherein the initial current solution gene population is the initial population, the historical optimal solution gene is the solution gene with the largest adaptive value in the current solution gene population, and the candidate optimal solution set is initially an empty set;
judging whether the iteration times are greater than or equal to a preset time threshold value or not;
under the condition that the iteration times are judged to be larger than or equal to the time threshold, judging whether the absolute value of the difference value of the adaptive values is smaller than or equal to a preset error or not;
under the condition that the absolute value is judged to be larger than the preset error and/or the iteration number is judged to be smaller than the number threshold, solution genes are randomly selected from the current solution gene population, the randomly generated solution genes, the candidate optimal solution set and the historical optimal solution genes respectively to create a target evolution solution gene set;
performing mutation operation on the target evolutionary solution gene set by adopting the formula (7) to the formula (10),
wherein, the first and the second end of the pipe are connected with each other,
for resolving genes after mutation manipulation, X rand For a solution gene selected from said current solution gene population and randomly generated solution genes, X best For randomly selected solution genes from the candidate optimal solution set and the historical optimal solution genes, X r1 、X r2 、X r3 、X r4 Randomly selecting a solution gene from a current solution gene population;
performing cross operation on the target evolutionary solution gene set before mutation and the target evolutionary solution gene set after mutation calculation by adopting formulas (11) and (12),
wherein the content of the first and second substances,
for solving gene V obtained after the cross operation rand Ith task T of i Is the candidate computing service of (a),
is gene solving before cross operation
Ith task T i Rand (0,1) is a number randomly generated from between 0 and 1, CR is a predetermined crossover operator,
serving candidate calculations of solution genes before mutation and crossover,
for solving gene V obtained after the cross operation best Ith task T of i Is the candidate computing service of (a),
for resolving genes
Task T of i Is the candidate computing service of (a),
candidate computing services for solution genes before mutation and crossover;
performing boundary correction operation on the solution genes subjected to the cross operation by adopting a formula (13),
wherein, bmv i For task T after boundary correction operation i Candidate computing service of, n i For task T i V number of candidate computing services of j Task T for solving Gene V before boundary correction operation i ，
Represents 1- j And n i -the remainder of the division by 1 is rounded down,
the remainder of the division of vi-ni and ni-1 is rounded;
creating a comparative solution gene set, wherein the comparative solution gene set comprises the target evolution solution gene set and all solution genes subjected to boundary correction operation;
calculating an adaptive value of each solution gene in the comparative solution gene set by adopting a formula (1), and calculating a corrected adaptive value of each solution gene by adopting a formula (14),
wherein bestfit is the maximum value of the adaptive values of the solution genes in the comparative solution gene set before correction, vio (x) is the default degree calculated by adopting the formula (15), and maxVio is the maximum default degree of the solution genes in the comparative solution gene set.
Judging whether the adaptive value of the solution gene corresponding to the maximum correction adaptive value of the comparative solution gene set is greater than the minimum adaptive value of the current solution gene population;
replacing the solution genes corresponding to the minimum adaptation value in the current solution gene population with the solution genes corresponding to the maximum correction adaptation value of the comparison solution gene set to update the current solution gene population under the condition that the judged adaptation value of the solution genes corresponding to the maximum correction adaptation value of the comparison solution gene set is greater than the minimum adaptation value of the current solution gene population;
judging whether the adaptive value of the solution gene corresponding to the maximum correction adaptive value is larger than the adaptive value of the historical optimal solution gene or not;
under the condition that the adaptive value of the solution gene corresponding to the maximum corrected adaptive value is judged to be larger than the adaptive value of the historical best solution gene, replacing the historical best solution gene with the solution gene corresponding to the maximum corrected adaptive value to update the historical best solution gene, and adding the replaced historical best solution gene into the candidate optimal solution set;
under the condition that the adaptive value of the solution gene corresponding to the maximum corrected adaptive value is judged to be smaller than or equal to the adaptive value of the historical optimal solution gene, adding the solution gene corresponding to the maximum corrected adaptive value into the candidate optimal solution set, and calculating the absolute value of the difference between the adaptive value of the solution gene corresponding to the maximum corrected adaptive value and the adaptive value of the historical optimal solution gene before updating;
and under the condition that the absolute value is judged to be larger than the preset error, outputting the historical best solution gene as the best deployment scheme.
Optionally, the determining an optimal deployment scenario for completing the task package according to the task package, the candidate computing service set, the objective function, and the boundary condition includes:
determining the calculation time of the combined optimization deployment method;
judging whether the calculation time is greater than a preset time threshold value;
and under the condition that the calculation time is judged to be greater than the time threshold, outputting the current historical best solution gene as a best deployment scheme.
Optionally, the outputting the historical best solution gene as the best deployment scenario includes:
and outputting the candidate optimal solution set.
Optionally, the adding the solution gene corresponding to the maximum modified adaptive value into the candidate optimal solution set includes:
judging the number of solution genes in the candidate optimal solution set;
judging whether the number of the solution genes is larger than a preset number value or not;
and deleting the solution genes with the minimum adaptive value in the optimal solution set under the condition that the number of the solution genes is judged to be larger than the number value.
In another aspect, the present invention further provides a combinatorial optimization deployment system of a cloud-edge hybrid computing service, where the combinatorial optimization deployment system includes a processor, and the processor is configured to execute any one of the combinatorial optimization deployment methods described above.
In yet another aspect, the present invention also provides a storage medium storing instructions for reading by a machine to cause the machine to perform any one of the combinatorial-optimized deployment methods described above.
Through the technical scheme, the method, the system and the storage medium for combined optimized deployment of the cloud-edge hybrid computing service provided by the invention determine the optimal deployment scheme of the cloud-edge hybrid computing service by adopting the improved differential evolution algorithm through combining various factors except the service quality, solve the technical problem that the deployment scheme cannot reach the optimal value due to only combining a single service quality factor when determining the scheme in the prior art, and enable the finally formed deployment scheme to meet the requirements of the cloud-edge hybrid computing service to the maximum extent.
Additional features and advantages of embodiments of the invention will be set forth in the detailed description which follows.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the embodiments of the invention without limiting the embodiments of the invention. In the drawings:
FIG. 1 is a flow diagram of a method for combinatorial-optimized deployment of cloud-edge hybrid computing services according to one embodiment of the invention;
FIG. 2 is a flow diagram of a method for combinatorial optimized deployment of cloud-edge hybrid computing services according to one embodiment of the invention;
FIG. 3 is a flow diagram of a differential evolution algorithm according to an embodiment of the present invention; and
FIG. 4 is a flow diagram of a differential evolution algorithm according to one embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating embodiments of the invention, are given by way of illustration and explanation only, not limitation.
In the embodiments of the present invention, unless otherwise specified, the use of directional terms such as "upper, lower, top, and bottom" is generally used with respect to the orientation shown in the drawings or the positional relationship of the components with respect to each other in the vertical, or gravitational direction.
In addition, if there is a description of "first", "second", etc. in the embodiments of the present invention, the description of "first", "second", etc. is for descriptive purposes only and is not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In addition, technical solutions between the various embodiments can be combined with each other, but must be realized by a person skilled in the art, and when the technical solutions are contradictory or cannot be realized, the combination of the technical solutions should be considered to be absent and not be within the protection scope of the present invention.
Fig. 1 is a flowchart illustrating a combined optimized deployment method of a cloud-edge hybrid computing service according to an embodiment of the present invention. In fig. 1, the combinatorial optimization deployment method may include:
in step S10, a task package of the cloud-edge hybrid computing service is acquired. The task package can comprise a common task set which needs to be completed by adopting the edge computing service and a special task set which needs to be completed by adopting the cloud computing service. The common task set and the special task set can respectively comprise at least one task, and each task can be connected through a Petri network structure.
In step S11, a set of candidate computing services that can complete each task is determined. Wherein each task corresponds to a set of candidate computing services for completing the task. Each set of candidate computing services includes at least one candidate computing service. Each candidate computing service includes a first feature attribute and a second feature attribute. The first feature attributes may include a type of candidate computing service and a brand of device used to complete the candidate computing service, a second feature attributeThe characteristic attribute may include a quality of service to complete the candidate computing service. Of this type, for example, a generic edge computing service GE, a cloud computing service SC, and an enterprise private computing service PC may be included. The quality of service may be, for example, various metrics for evaluating candidate computing services, including, for example, for representing the candidate computing service S ij Availability of probability of being awakenable within service execution time
For representing the candidate computing service S ij Reliability of probability of successful operation without false interruption within service execution time
For representing the candidate computing service S ij Run time of time used in execution
For representing the candidate computing service S ij Product cost of price of
For representing the candidate computing service S ij Overall quality of the product and integrated high and low degree of brand factor
(to distinguish the degree of trust in the product itself). Thus, for each candidate computing service, its quality of service can be expressed as an index set QoSset ═ { ava, rel, tim, cos, rep }. In this embodiment, for any index, Q may be used Ind To indicate. Accordingly, for any one task T i Any one of the corresponding candidate computing services S ij Can then adopt
Represents, and Ind ∈ QoSset. In addition, because of the clothesThe corresponding relationship between the quality of each index of the quality of service and the value size may be different, so as to run the time
For example, the larger the corresponding numerical value is, the lower the service quality is; at the product reputation level
For example, the larger the corresponding value, the higher the quality of service. Therefore, in order to facilitate the uniform calculation of the quality of service, in this embodiment, each index Ind may be divided into forward calculation quality of service Ind + And negative calculation of the quality of service Ind - . Wherein Ind + ＝{ava,rel,rep}，Ind - ＝{tim,cos}。
In step S12, using equation (1) as the preset objective function,
Fit(X)＝ω 1 ·TQoS+ω 2 ·TSR+ω 3 ·TSC，(1)
where Fit (X) is the objective function, ω, of the deployment scenario X 1 、ω 2 And ω 3 For the preset weight, TQoS represents the total quality of service, TSR represents the degree value of the special calculation service calculated by formula (2),
wherein TSR satisfies a degree value for a particular computing service, N SRST Number of tasks, X, included for a particular set of tasks SRST Is the number of tasks in the set of special tasks that the deployment scenario for completing the task package can complete, gamma SRST Is a preset accumulated intensity value, and gamma SRST Not less than 1; TSC is the service consistency degree value calculated using equation (3),
wherein TSC isService consistency degree value, X B＝b To select the number of candidate computing services of brand b in the deployment scenario for completing the task package, n is the number of tasks included in the task package. In addition, the calculation of the total QoS performance value may be in various forms known to those skilled in the art. In one example of the invention, the total quality of service QoS performance value may be calculated according to equation (4),
TQoS＝∑ Ind (ω Ind /sum(ω Ind ))SQ Ind,P ，(4)
wherein, ω is Ind Weight of index Ind of quality of service for candidate computing service, SQ Ind,P The index value is a standardized index value of the index Ind when the task packet P is completed. In this example, four relationships of a serial mode, a parallel mode, a selection mode and a loop mode may exist among tasks in a task package, where the serial mode indicates that the tasks need to be completed in a sequential order, the parallel mode indicates that the tasks need to be completed simultaneously, the selection mode indicates that only one or more tasks can be selected simultaneously to be completed, and the loop mode indicates that the tasks need to be completed circularly. Therefore, before calculating the objective function, the tasks in the task package can be divided into:
serial mode task group aggregation
Parallel mode task group aggregation
Selection mode task group aggregation
And
set of looping mode task groups
Wherein the content of the first and second substances,
set of serial mode task groups SEQ kth seq A serial task group, n seq The number of serial task groups which are serial mode task group sets SEQ;
aggregating the kth of the PAR for parallel mode task groups par A parallel task group, n par The number of parallel task groups for collecting PAR for the parallel tasks;
for selecting the kth of the set SEL of mode task groups sel A selection task group, n sel Selecting the number of task groups for the selection mode task group set SEL;
k-th of CIR for set of cyclic mode task groups cir A cyclic task group, n cir The number of the cyclic task groups in the cyclic mode task group set is shown. According to the knowledge of those skilled in the art (H.Liang, Y.Du.dynamic service selection with QoS constraints and inter-service correlation using coherent evaluation J]Future Generation Computer systems.2017,76:119- "135)), the parallel mode task group set, the selection mode task group set, and the loop mode task group set can be finally converted into a part of the serial mode task group set according to the preset rules. Therefore, when calculating the total QoS performance value, the aggregate calculation may be performed on each task set first. Specifically, the tasks of the serial mode task group set may be subjected to the aggregation calculation according to equations (5) to (9),
wherein the content of the first and second substances,
(Serial) task group SEQ for completing a set SEQ of task groups in serial mode k A normalized index value of availability, task group seq k Including multiple tasks in the same serial mode,
to complete task T i A normalized index value of the availability of,
task group SEQ for completing a set SEQ of task groups in serial mode k The normalized index value of the reliability of (a),
task group SEQ for completing a set SEQ of task groups in serial mode k The normalized index value of the running time of (c),
to complete task T i The normalized index value of the running time of (c),
task group SEQ for completing a set SEQ of task groups in serial mode k Product cost ofThe normalized index value of (a) is,
to complete task T i A normalized index value of the product cost of (a),
task group SEQ for completing a set SEQ of task groups in serial mode k The normalized index value of the product reputation level of (a),
to complete task T i The normalized index value of the product reputation level of (a),
is a task group seq k Number of tasks in, seq k ∈SEQ；
Performing aggregation calculation on the tasks of the parallel mode task group set according to the formula (10) to the formula (14),
wherein the content of the first and second substances,
(parallel) task group PAR for assembling PAR for completing parallel mode task group k A normalized index value of availability, task group par k Comprising a plurality of tasks in the same parallel mode,
to complete task T i A normalized index value of the availability of,
task group PAR for assembling PAR for task group in task parallel mode k The normalized index value of the reliability of (a),
to complete task T i The normalized index value of the reliability of (a),
task group PAR for assembling PAR for completing parallel mode task group k The normalized index value of the running time of (c),
to complete task T i The normalized index value of the operation time of (c),
task group PAR for assembling PAR for completing parallel mode task group k A normalized index value of the product cost of (a),
to complete task T i A normalized index value of the product cost of (a),
task group PAR for assembling PAR for completing parallel mode task group k A normalized index value of the reputation level of (c),
to complete task T i A normalized index value of the reputation level of (c),
is task group par k Number of tasks of, par k ∈PAR；
Performing an aggregation calculation on the tasks of the selection mode task group set according to equations (15) to (19),
wherein the content of the first and second substances,
(SELECT) task GROUP SEL FOR PERFORMING A SELECT MODE task GROUP SEL k A normalized index value of the availability of,
to complete task T i Normalized index value of availability, theta i For task T i In task group sel k Is selected to be performed, and the probability of being selected to be performed,
task group SEL for completing selection mode task group set SEL k The normalized index value of the reliability of (a),
to complete task T i The normalized index value of the reliability of (a),
task set SEL for completing selection mode task set SEL k The normalized index value of the operation time of (c),
to complete task T i The normalized index value of the running time of (c),
task set SEL for completing selection mode task set SEL k A normalized index value of the product cost of (a),
to complete task T i A normalized index value of the product cost of (a),
task set SEL for completing selection mode task set SEL k A normalized index value of the reputation level of (c),
to complete task T i A normalized index value of the reputation level of (c),
for task group sel k Number of tasks of sel k ∈SEL；
Performing an aggregation calculation on the task loops of the loop mode task group set according to the formula (20) to the formula (24),
wherein the content of the first and second substances,
(Cyclic) task group CIR for assembling CIR for completing cyclic mode task group k A normalized index value of availability, a task group cir k Comprising a plurality of tasks in the same cyclic pattern,
to complete task T i A normalized index value of the availability of,
task group CIR for assembling CIR for completing cycle mode task group k The normalized index value of the reliability of (a),
to complete task T i The normalized index value of the reliability of (a),
task group CIR for assembling CIR for completing cycle mode task group k Gamma is the task group cir k The number of cycles of (a) to (b),
to complete task T i The normalized index value of the running time of (c),
task group CIR for assembling CIR for completing cycle mode task group k A normalized index value of the product cost of (a),
to complete task T i A normalized index value of the product cost of (a),
task group CIR for assembling CIR for completing cycle mode task group k The normalized index value of the reputation level of (a),
to complete task T i A normalized index value of the reputation level of (c),
as task groups cir k Number of tasks of, cir k ∈CIR。
And performing aggregation calculation on the whole ALLMODE formed by the result of the aggregation calculation. Specifically, the aggregation calculation may be performed according to the formula (25) to the formula (29),
SQ ava,P ＝∏SQ ava,ALLMODE ，(25)
SQ rel,P ＝∏SQ rel,ALLMODE ，(26)
SQ tim,P ＝∑SQ tim,ALLMODE ，(27)
SQ cos,P ＝∑SQ cos,ALLMODE ，(28)
SQ rep,P ＝∑SQ rep,ALLMODE /n ALLMODE ，(29)
wherein SQ ava,P Standardized indicator value of availability of completion task Package P, SQ ava,ALLMODE Standardized indicator value for the availability of one element (task) in the complete ALLMODE, SQ rel,P Standardized indicator value for reliability of completion task packet P, SQ rel,ALLMODE To achieve a normalized index value of the reliability of an element in the overall ALLMODE, SQ tim,P To achieve a normalized index value for the running time of the task Package P, SQ tim,E To complete a normalized index value of runtime of an element in the overall ALLMODE, SQ cos,P Standardized indicator value of product cost for completing task Package P, SQ cos,ALLMODE To achieve a normalized index value of product cost for an element in the overall ALLMODE, SQ rep,P To complete a standardized indicator value of the reputation level of the task Package P, SQ rep,ALLMODE To achieve a normalized index value for the reputation level of an element in the overall ALLMODE, n ALLMODE The number of elements in the entire ALLMODE.
Considering each index Ind as a forward calculated quality of service Ind + And negative calculation of the quality of service Ind - . Wherein, Ind + ＝{ava,rel,rep}，Ind - Time, cos. Then, after calculating each normalized index value, the calculated normalized index value can be further normalized by using the formula (30) and the formula (31),
wherein, the first and the second end of the pipe are connected with each other,
for after standardization processAffair T i Forward to calculate a normalized index value of the quality of service,
for task T before standardization processing i Forward to calculate a normalized index value of the quality of service,
and
for task T i Maximum quality of service Q in the candidate computing service of Ina A value of (positive or negative computation quality of service),
and
for task T i Maximum quality of service Q in the candidate computing service of Ind To the value of (1), to availability
For example, its quality of service Q ava The quality of service is calculated for the forward direction, i.e.:
through the method, the finally calculated total QoS expression values are all positioned between [0,1], so that the numerical dimension and unit inconsistency of the five QoS expression values does not occur, and the calculation of a differential evolution algorithm is facilitated.
In step S13, an optimal deployment scenario for completing the task package is determined according to the task package, the candidate computing service set, and the objective function by using a differential evolution algorithm.
In one embodiment of the present invention, to improve the effect of the optimal deployment scenario determined by the differential evolution algorithm, the combinatorial optimization deployment method may include the steps as shown in fig. 2. The difference from the combinatorial-optimized deployment method shown in fig. 1 is that, in fig. 2, the combinatorial-optimized deployment method further includes:
in step S13, a boundary condition is preset. The boundary condition can be set by those skilled in the art in combination with actual user requirements. In one example of the present invention, the boundary condition may include formula (32) and formula (33),
wherein the content of the first and second substances,
to complete task T i A normalized index value of the time index Ind,
to complete task T i The minimum required normalized index value of the time index Ind, SQ Ind,P To complete the normalized index value of the index Ind for the task package P,
the minimum required standardized index value of the index Ind when the task packet P is completed.
Accordingly, the step S14 may determine the optimal deployment scenario for completing the task package according to the task package, the candidate computing service set, the objective function and the boundary condition. In this step S14, the specific details of determining the optimal deployment scenario from the task package, the set of candidate computing services, the objective function, and the boundary conditions using a differential evolution algorithm may be steps well known to those skilled in the art. In one example of the present invention, the step S14 may specifically include the steps as shown in fig. 3. In fig. 3, the step S14 may specifically include:
in step S1401, each parameter is input to the differential evolution algorithm.
In step S1402, each task in the task package is traversed, and candidate computing services that do not satisfy the boundary condition in the candidate computing service set corresponding to the task are removed. In this embodiment, each candidate computing service is computed in view of the differential evolution algorithm. Therefore, the number of candidate computing services directly affects the computing speed of the entire differential evolution algorithm. In step S1402, before executing the differential evolution algorithm, the candidate computing services that do not satisfy the condition in the candidate computing service set corresponding to each task are removed according to the preset boundary conditions (formula (31) and formula (32)), so that the number of the candidate computing services can be reduced, the operation speed of the overall algorithm is increased, and the load of the device executing the computation is also reduced.
In step S1403, a plurality of solution genes X are randomly generated by means of integer coding to form an initial population. Wherein the content of the first and second substances,
to complete task T i The selected candidate computing service;
in step S1404, an adaptation value of each solution gene of the initial population is calculated according to formula (1);
in step S1405, a current solution gene population, a historical best solution gene, and a candidate best solution set are set. The initial current solution gene population is an initial population, the historical best solution gene is a solution gene with the largest adaptive value in the current solution gene population, and the candidate optimal solution set is initially an empty set;
in step S1406, it is determined whether the number of iterations is greater than or equal to a preset number threshold. Wherein the initial number of iterations may be 0;
in step S1407, in a case where it is determined that the number of iterations is greater than or equal to the number threshold, it is determined whether an absolute value of a difference between the adaptation values is less than or equal to a preset error;
in step S1408, in the case that the absolute value is judged to be greater than the preset error and/or the number of iterations is judged to be less than the number threshold, solution genes are randomly selected from the current solution gene population, the randomly generated solution genes, the candidate optimal solution set, and the historical optimal solution genes, respectively, to create a target evolution solution gene set. Compared with the traditional differential evolution algorithm which only randomly selects the genes in the current solution gene population and the historical optimal genes, the differential evolution algorithm provided by the invention provides richer solution genes for subsequent screening by setting the target evolution solution gene set, and the randomly selected genes in the candidate optimal solution set have better adaptive values relative to the genes in other sets, so that the operation time of the algorithm is greatly reduced. In addition, the target evolutionary solution gene set is created in a random selection mode, so that the target evolutionary solution gene set is more general, and the accuracy of algorithm calculation is improved.
In step S1409, the target evolution solution gene set is mutated by using the formula (34) to the formula (37),
wherein, the first and the second end of the pipe are connected with each other,
for gene solving after mutation manipulation, X rand For a solution gene selected from the current solution gene population and randomly generated solution genes, X best For randomly selected solution genes from the candidate optimal solution set and the historical optimal solution genes, X r1 、X r2 、X r3 、X r4 Randomly selecting a solution gene from a current solution gene population;
in step S1410, the target evolution and solution gene set before mutation and the target evolution and solution gene set after mutation calculation are cross-operated using the formulas (38) and (39),
wherein the content of the first and second substances,
is a gene V obtained by cross operation rand Ith task T i The candidate computing services of (a) are,
is gene solving before cross operation
Ith task T i Rand (0,1) is a number randomly generated from between 0 and 1, CR is a predetermined crossover operator,
serving candidate calculations of solution genes before mutation and crossover,
is a gene V obtained by cross operation best Ith task T of i Is the candidate computing service of (a),
for resolving genes
Task T of i Is the candidate computing service of (a),
candidate computing services for variant and pre-crossover solution genes. Compared with the crossing mode of a single random number provided by the traditional differential evolution algorithm, the crossing mode of a random point method (formula (38) and formula (39)) is further adopted in the differential evolution algorithm provided by the invention, so that the types of solution genes subjected to crossing operation are richer;
in step S1411, a boundary correction operation is performed on the solution gene subjected to the crossover operation using formula (40),
wherein, bmv i For task T after boundary correction operation i Candidate computing service of, n i For task T i V number of candidate computing services of j Task T for solving genes V before performing boundary correction operation i ，
Expression 1- j And n i -the remainder of the division by 1 is rounded down,
the remainder of the vi-ni divided by ni-1 is rounded down. In this embodiment, it is considered that the integer codes at certain positions of the solution gene may exceed the range of the candidate calculation service numbers after the mutation operation. Therefore, in this embodiment, the solution gene after the mutation operation is subjected to boundary correction by using the loop coding (formula (40)).
In step S1412, a comparative solution gene set is created, where the comparative solution gene set includes the target evolution solution gene set and all solution genes after the boundary correction operation;
in step S1413, an adaptive value for each solution gene in the comparative solution gene set is calculated using formula (1), and a corrected adaptive value for each solution gene is calculated using formula (41),
wherein bestfit is the maximum value of the adaptive values of the solution genes in the comparative solution gene set before correction, vio (x) is the default degree calculated by adopting a formula (42), and maxVio is the maximum default degree of the solution genes in the comparative solution gene set.
In step S1414, it is determined whether the adaptive value of the solution gene corresponding to the maximum corrected adaptive value of the comparative solution gene set is greater than the minimum adaptive value of the current solution gene population;
in step S1415, in a case that the adaptive value of the solution gene corresponding to the determined maximum corrected adaptive value of the comparative solution gene set is greater than the minimum adaptive value of the current solution gene population, replacing the solution gene corresponding to the minimum adaptive value in the current solution gene population with the solution gene corresponding to the maximum corrected adaptive value of the comparative solution gene set to update the current solution gene population, and performing step S1416;
in step S1416, it is determined whether the adaptive value of the solution gene corresponding to the maximum corrected adaptive value is greater than the adaptive value of the historical best solution gene;
in step S1417, in a case where it is determined that the adaptation value of the solution gene corresponding to the maximum modified adaptation value is greater than the adaptation value of the historical best solution gene, replacing the historical best solution gene with the solution gene corresponding to the maximum modified adaptation value to update the historical best solution gene, adding the replaced historical best solution gene to the candidate best solution set, and performing step S1418;
in step S1418, in a case where it is determined that the adaptation value of the solution gene corresponding to the maximum corrected adaptation value is less than or equal to the adaptation value of the historical best solution gene, adding the solution gene corresponding to the maximum corrected adaptation value to the candidate optimal solution set, calculating an absolute value of a difference between the adaptation value of the solution gene corresponding to the maximum corrected adaptation value and the adaptation value of the historical best solution gene before update, and performing step S1406;
in step S1419, in the case where it is determined that the absolute value is larger than the preset error, the historical best solution gene is output as the best deployment scenario.
Further, to avoid the calculation time of the algorithm being too long, the differential evolution algorithm may also include the steps as shown in fig. 4. The difference from the differential evolution algorithm shown in fig. 3 is that, in fig. 4, the differential evolution algorithm further includes:
in step S1406, determining the calculation time of the combinatorial optimization deployment method, and determining whether the calculation time is greater than a preset time threshold;
in step S1420, in the case where it is judged that the calculation time is greater than the time threshold, the current historical best solution gene is output as the best deployment scenario.
In addition, in order to facilitate the selection of the optimal deployment scenario by the staff, in step S1420, while outputting the historical optimal solution genes as the optimal deployment scenario, the candidate optimal solution set may also be output for the staff to select. Furthermore, when the combinatorial optimization deployment method shown in any one of fig. 1 to fig. 3 is executed, considering that the number of the generated candidate optimal solution sets may be large, in order to avoid an excessive memory occupation, when adding the solution genes corresponding to the maximum modified adaptive value into the candidate optimal solution sets, the following steps may also be executed:
step 1, judging the number of solution genes in a candidate optimal solution set;
step 2, judging whether the number of the solution genes is larger than a preset number value or not;
and 3, deleting the solution genes with the minimum adaptive value in the optimal solution set under the condition that the number of the solution genes is judged to be larger than the number value.
In another aspect, the present invention further provides a combinatorial-optimized deployment system for cloud-edge hybrid computing services, where the combinatorial-optimized deployment system may include a processor, and the processor may be configured to execute any one of the combinatorial-optimized deployment methods described above.
In yet another aspect, the present invention also provides a storage medium that may store instructions that are readable by a machine to cause the machine to perform any of the combinatorial-optimized deployment methods described above.
According to the technical scheme, the optimal deployment scheme of the cloud-side hybrid computing service is determined by adopting the improved differential evolution algorithm by combining various factors except the service quality, so that the technical problem that the optimal deployment scheme cannot be achieved due to the fact that only a single service quality factor is combined when the scheme is determined in the prior art is solved, and the finally formed deployment scheme can meet the requirements of the cloud-side hybrid computing service to the greatest extent.
Although the embodiments of the present invention have been described in detail with reference to the accompanying drawings, the embodiments of the present invention are not limited to the details of the above embodiments, and various simple modifications can be made to the technical solution of the embodiments of the present invention within the technical idea of the embodiments of the present invention, and the simple modifications all belong to the protection scope of the embodiments of the present invention.
It should be noted that the various features described in the above embodiments may be combined in any suitable manner without departing from the scope of the invention. In order to avoid unnecessary repetition, the embodiments of the present invention will not be described separately for the various possible combinations.
Those skilled in the art can understand that all or part of the steps in the method for implementing the above embodiments may be implemented by a program to instruct related hardware, where the program is stored in a storage medium and includes several instructions to enable a (may be a single chip, a chip, etc.) or a processor (processor) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk, and various media capable of storing program codes.
In addition, any combination of various different embodiments of the present invention may be made, and the same should be considered as what is disclosed in the embodiments of the present invention as long as it does not depart from the spirit of the embodiments of the present invention.
Claims (10)
1. A combined optimization deployment method of a cloud-edge hybrid computing service is characterized by comprising the following steps:
acquiring a task package of the cloud-edge hybrid computing service, wherein the task package comprises a common task set which needs to be completed by adopting an edge computing service and a special task set which needs to adopt a cloud computing service, the common task set and the special task set both comprise at least one task, and each task is connected with each other through a Petri network structure;
determining a set of candidate computing services capable of completing each of the tasks, wherein each task corresponds to a set of candidate computing services for completing the task, each set of candidate computing services comprises at least one candidate computing service, each candidate computing service comprises a first feature attribute and a second feature attribute, the first feature attribute comprises a type of the candidate computing service and a brand of a device for completing the candidate computing service, and the second feature attribute comprises a quality of service for completing the candidate computing service;
adopting the formula (1) as a preset objective function,
Fit(X)＝ω 1 ·TQoS+ω 2 ·TSR+ω 3 ·TSC， (1)
wherein Fit (X) is the objective function, ω, of deployment scenario X 1 、ω 2 And omega 3 For the preset weight, TQoS represents the total quality of service, TSR represents the degree value of the special calculation service calculated by formula (2),
wherein TSR satisfies a degree value, N, for the special computing service SRST Number of tasks, X, included for said set of special tasks SRST Is the number of tasks in the set of special tasks that the deployment scenario for completing the task package can complete, γ SRST Is a preset accumulated intensity value, and gamma SRST Not less than 1; TSC is the service consistency degree value calculated using equation (3),
wherein TSC is the service consistency degree value X B＝b Selecting the number of candidate computing services of brand b in a deployment scheme for completing the task package, wherein n is the number of tasks included in the task package;
and determining an optimal deployment scheme for completing the task package according to the task package, the candidate computing service set and the objective function by adopting a differential evolution algorithm.
2. The combinatorial-optimized deployment method of claim 1, further comprising:
presetting boundary conditions;
the step of determining the optimal deployment scheme for completing the task package according to the task package, the candidate computing service set and the objective function by adopting a differential evolution algorithm comprises the following steps:
and determining an optimal deployment scheme for completing the task package according to the task package, the candidate computing service set, the objective function and the boundary condition.
3. The combinatorial-optimization deployment method of claim 2, wherein the boundary conditions include formula (4) and formula (5),
wherein the content of the first and second substances,
to complete task T i The normalized index value of the time index Ind,
to complete task T i Minimum required normalized index value, SQ, of the time index Ind Ind，P To complete the normalized index value of the index Ind for the task package P,
the minimum required standardized index value of the index Ind when the task packet P is completed.
4. The combinatorial-optimized deployment method of claim 1, further comprising:
calculating the total quality of service (QoS) performance value according to formula (6),
TQoS＝∑ Ind (ω Ind /sum(ω Ind ))SQ Ind，P ， (6)
wherein, ω is Ind Weight, SQ, of an indicator Ind of the quality of service of the candidate calculation service Ind，P The index value is a standardized index value of the index Ind when the task packet P is completed.
5. The combinatorial-optimization deployment method of claim 2, wherein the determining an optimal deployment scenario for completing the task package according to the task package, the set of candidate computing services, an objective function, and the boundary condition comprises:
traversing each task in the task package, and eliminating the candidate computing services which do not meet the boundary condition in the candidate computing service set corresponding to the task;
randomly generating a plurality of solution genes X by adopting an integer coding mode to form an initial population, wherein,
to complete task T i The selected candidate computing service;
calculating an adaptation value of each solution gene of the initial population according to formula (1);
setting a current solution gene population, a historical best solution gene and a candidate best solution set, wherein the initial current solution gene population is the initial population, the historical best solution gene is the solution gene with the largest adaptive value in the current solution gene population, and the candidate best solution set is initially an empty set;
judging whether the iteration times are greater than or equal to a preset time threshold value or not;
under the condition that the iteration times are judged to be larger than or equal to the time threshold, judging whether the absolute value of the difference value of the adaptive values is smaller than or equal to a preset error or not;
under the condition that the absolute value is judged to be larger than the preset error and/or the iteration number is judged to be smaller than the number threshold, solution genes are randomly selected from the current solution gene population, the randomly generated solution genes, the candidate optimal solution set and the historical optimal solution genes respectively to create a target evolution solution gene set;
performing mutation operation on the target evolutionary solution gene set by adopting the formula (7) to the formula (10),
wherein the content of the first and second substances,
for gene solving after mutation manipulation, X rand For a solution gene selected from said current solution gene population and randomly generated solution genes, X best For randomly selected solution genes from the candidate optimal solution set and the historical optimal solution genes, X r1 、X r2 、X r3 、X r4 Randomly selecting a solution gene from a current solution gene population;
performing cross operation on the target evolution solution gene set before mutation and the target evolution solution gene set after mutation calculation by adopting formulas (11) and (12),
wherein the content of the first and second substances,
for solving gene V obtained after the cross operation rand Ith task T i Is the candidate computing service of (a),
is gene solving before cross operation
Ith task T i Rand (0,1) is a number randomly generated from between 0 and 1, CR is a predetermined crossover operator,
serving candidate computing services for the solution genes before mutation and crossover,
for solving gene V obtained after the cross operation best Ith task T i Is the candidate computing service of (a),
for resolving genes
Task T of i Is the candidate computing service of (a),
candidate computing services for solution genes before mutation and crossover;
performing boundary correction operation on the solution gene subjected to the cross operation by using a formula (13),
wherein, bmv i For task T after boundary correction operation i Candidate computing service of, n i For task T i V number of candidate computing services of j Task T for solving genes V before performing boundary correction operation i ，
Represents 1-v j And n i -the remainder of the division by 1 is rounded down,
denotes v i -n i And n i -1 division by the remainder of the rounding;
creating a comparative solution gene set, wherein the comparative solution gene set comprises the target evolution solution gene set and all solution genes subjected to boundary correction operation;
calculating an adaptive value of each solution gene in the comparative solution gene set by adopting a formula (1), and calculating a corrected adaptive value of each solution gene by adopting a formula (14),
wherein bestfit is the maximum value of the adaptive values of the solution genes in the comparative solution gene set before correction, vio (x) is the default degree calculated by adopting the formula (15), and maxVio is the maximum default degree of the solution genes in the comparative solution gene set;
judging whether the adaptive value of the solution gene corresponding to the maximum correction adaptive value of the comparative solution gene set is greater than the minimum adaptive value of the current solution gene population;
replacing the solution genes corresponding to the minimum adaptation value in the current solution gene population with the solution genes corresponding to the maximum correction adaptation value of the comparison solution gene set to update the current solution gene population under the condition that the judged adaptation value of the solution genes corresponding to the maximum correction adaptation value of the comparison solution gene set is greater than the minimum adaptation value of the current solution gene population;
judging whether the adaptive value of the solution gene corresponding to the maximum correction adaptive value is larger than the adaptive value of the historical optimal solution gene or not;
under the condition that the adaptive value of the solution gene corresponding to the maximum corrected adaptive value is judged to be larger than the adaptive value of the historical best solution gene, replacing the historical best solution gene with the solution gene corresponding to the maximum corrected adaptive value to update the historical best solution gene, and adding the replaced historical best solution gene into the candidate optimal solution set;
under the condition that the adaptive value of the solution gene corresponding to the maximum corrected adaptive value is judged to be smaller than or equal to the adaptive value of the historical optimal solution gene, adding the solution gene corresponding to the maximum corrected adaptive value into the candidate optimal solution set, and calculating the absolute value of the difference between the adaptive value of the solution gene corresponding to the maximum corrected adaptive value and the adaptive value of the historical optimal solution gene before updating;
and under the condition that the absolute value is judged to be larger than the preset error, outputting the historical best solution gene as the best deployment scheme.
6. The combinatorial-optimization deployment method of claim 5, wherein the determining an optimal deployment scenario for completing the task package according to the task package, the set of candidate computing services, an objective function, and the boundary condition comprises:
determining the calculation time of the combined optimization deployment method;
judging whether the calculation time is greater than a preset time threshold value;
and under the condition that the calculation time is judged to be greater than the time threshold, outputting the current historical best solution gene as a best deployment scheme.
7. The combinatorial-optimized deployment method of claim 5, wherein the outputting the historical best solution gene as the best deployment scenario comprises:
and outputting the candidate optimal solution set.
8. The combinatorial optimization deployment method of claim 7, wherein the adding the solution gene corresponding to the maximum modified adaptive value to the candidate optimal solution set comprises:
judging the number of solution genes in the candidate optimal solution set;
judging whether the number of the solution genes is larger than a preset number value or not;
and deleting the solution genes with the minimum adaptive value in the candidate optimal solution set under the condition that the number of the solution genes is judged to be larger than the number value.
9. A combinatorial-optimization deployment system of a cloud-edge hybrid computing service, the combinatorial-optimization deployment system comprising a processor configured to perform the combinatorial-optimization deployment method of any one of claims 1 to 8.
10. A storage medium storing instructions for reading by a machine to cause the machine to perform the combinatorial-optimized deployment method of any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910904428.1A CN110659034B (en) | 2019-09-24 | 2019-09-24 | Combined optimization deployment method, system and storage medium of cloud-edge hybrid computing service |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910904428.1A CN110659034B (en) | 2019-09-24 | 2019-09-24 | Combined optimization deployment method, system and storage medium of cloud-edge hybrid computing service |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110659034A CN110659034A (en) | 2020-01-07 |
| CN110659034B true CN110659034B (en) | 2022-09-20 |
Family
ID=69038824
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910904428.1A Active CN110659034B (en) | 2019-09-24 | 2019-09-24 | Combined optimization deployment method, system and storage medium of cloud-edge hybrid computing service |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110659034B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111913723B (en) * | 2020-06-15 | 2022-09-23 | 合肥工业大学 | Cloud-edge-end cooperative unloading method and system based on assembly line |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108304168A (en) * | 2018-02-27 | 2018-07-20 | 苏州商合仕物联科技有限公司 | A kind of edge calculations machine operating system |
| CN109240768A (en) * | 2018-09-18 | 2019-01-18 | 广东微云科技股份有限公司 | The operation method and system of virtual machine under a kind of mixed computing model |
| CN109684075A (en) * | 2018-11-28 | 2019-04-26 | 深圳供电局有限公司 | A method of calculating task unloading is carried out based on edge calculations and cloud computing collaboration |
| CN110210806A (en) * | 2019-05-27 | 2019-09-06 | 大连理工大学 | A kind of the cloud base unmanned vehicle framework and its control evaluation method of 5G edge calculations |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010102084A2 (en) * | 2009-03-05 | 2010-09-10 | Coach Wei | System and method for performance acceleration, data protection, disaster recovery and on-demand scaling of computer applications |
| US10326766B2 (en) * | 2017-07-13 | 2019-06-18 | Dell Products, Lp | Method and apparatus for optimizing mobile edge computing for nomadic computing capabilities as a service |
-
2019
- 2019-09-24 CN CN201910904428.1A patent/CN110659034B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108304168A (en) * | 2018-02-27 | 2018-07-20 | 苏州商合仕物联科技有限公司 | A kind of edge calculations machine operating system |
| CN109240768A (en) * | 2018-09-18 | 2019-01-18 | 广东微云科技股份有限公司 | The operation method and system of virtual machine under a kind of mixed computing model |
| CN109684075A (en) * | 2018-11-28 | 2019-04-26 | 深圳供电局有限公司 | A method of calculating task unloading is carried out based on edge calculations and cloud computing collaboration |
| CN110210806A (en) * | 2019-05-27 | 2019-09-06 | 大连理工大学 | A kind of the cloud base unmanned vehicle framework and its control evaluation method of 5G edge calculations |
Non-Patent Citations (3)
| Title |
|---|
| A Time-Driven Data Placement Strategy for a Scientific Workflow Combining Edge Computing and Cloud Computing;Bing Lin;《IEEE Transactions on Industrial Informatics 》;20190317;4254-4265 * |
| 基于信任的云服务系统多目标任务分配模型;束柬等;《计算机研究与发展》;20180615(第06期);全文 * |
| 面向服务的云-端动态协作方法;曹云梦等;《计算机工程与科学》;20190415(第04期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110659034A (en) | 2020-01-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111047563B (en) | Neural network construction method applied to medical ultrasonic image | |
| CN108287666B (en) | Data storage method and device for cloud storage environment | |
| CN109710374A (en) | The VM migration strategy of task unloading expense is minimized under mobile edge calculations environment | |
| Yao et al. | Improved artificial bee colony algorithm for vehicle routing problem with time windows | |
| CN106202092A (en) | The method and system that data process | |
| CN115858168B (en) | Earth application model arrangement system and method based on importance ranking | |
| CN110659034B (en) | Combined optimization deployment method, system and storage medium of cloud-edge hybrid computing service | |
| CN110119399B (en) | Business process optimization method based on machine learning | |
| CN110929218A (en) | Difference minimization random grouping method and system | |
| TWI740895B (en) | Distribution method and device for application attribution service cluster | |
| CN109245948B (en) | Security-aware virtual network mapping method and device | |
| CN110275895A (en) | It is a kind of to lack the filling equipment of traffic data, device and method | |
| Gao et al. | The multi-objective cloud tasks scheduling based on hybrid particle swarm optimization | |
| CN113220437B (en) | Workflow multi-target scheduling method and device | |
| CN110309948A (en) | Complete vehicle logistics order forecast method and device, logistics system and computer-readable medium | |
| WO2022160752A1 (en) | Operations research optimization method and apparatus, and computing device | |
| CN110059880B (en) | Service discovery method and device | |
| CN112036794A (en) | Engineering material warehousing method, system and device for cost control | |
| CN114678069B (en) | Immune rejection prediction and signal path determination device for organ transplantation | |
| CN114826951B (en) | Service automatic degradation method, device, computer equipment and storage medium | |
| WO2022178917A1 (en) | Parallel flexible skyline service discovery method based on quality-of-service perception | |
| CN113162780B (en) | Real-time network congestion analysis method, device, computer equipment and storage medium | |
| US11886352B2 (en) | Access friendly memory architecture of graph neural network sampling | |
| CN116614392B (en) | Routing overhead determining method, routing overhead monitoring method and device | |
| US20230125509A1 (en) | Bayesian adaptable data gathering for edge node performance prediction |
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 |