CN109710372A - A kind of computation-intensive cloud workflow schedule method based on cat owl searching algorithm - Google Patents
A kind of computation-intensive cloud workflow schedule method based on cat owl searching algorithm Download PDFInfo
- Publication number
- CN109710372A CN109710372A CN201811336040.8A CN201811336040A CN109710372A CN 109710372 A CN109710372 A CN 109710372A CN 201811336040 A CN201811336040 A CN 201811336040A CN 109710372 A CN109710372 A CN 109710372A
- Authority
- CN
- China
- Prior art keywords
- task
- indicate
- virtual machine
- scheduling scheme
- cloud
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The computation-intensive cloud workflow schedule method based on cat owl searching algorithm that the invention proposes a kind of, belongs to field of cloud computer technology.By modifying population iteration more new formula in cat owl searching algorithm, so that each scheduling scheme is influenced size according to optimal scheduling scheme on it to update, optimizing is made to have more specific aim；In population iteration update mechanism, by utilizing hereditary variation thought, randomness is introduced, search process is avoided to fall into local optimum, optimal scheduling scheme can be obtained within the shorter time, realization reasonably distributes virtual machine, efficiently dispatched to task.The present invention can effectively overcome that optimal solution search randomness in existing method is big, is easy to fall into local optimum and disadvantage that convergence rate is slow, it promotes search efficiency, shorten search time, more preferably scheduling scheme can be searched out within the shorter time, reduce the time overhead of workflow schedule.
Description
Technical field
The present invention relates to a kind of cloud workflow schedule methods, and in particular to a kind of calculating based on cat owl searching algorithm is close
Collection type cloud workflow schedule method, belongs to field of cloud computer technology.
Background technique
Cloud computing calculates mode as a kind of business, using virtualization technology, by the storage, calculating and net of data center
The resource consolidations such as network communication are a shared, dynamically configurable computing resource pool, provide the meter of pay-per-use for user
Calculate service.User can be accessed by available, convenient and fast network without purchasing the hardware resources such as any server, into can
The shared computing resource pool (such as server, storage, application software and network etc.) of configuration obtains computing capability, storage on demand
Space and information service.
With the continuous development of cloud computing, large-scale complex work flow becomes the new model of cloud computing application.Cloud work
The execution of stream mainly includes two stages of task schedule and resource provision.In task scheduling process, need according to tune appropriate
Degree strategy, for the suitable virtual machine of task choosing of user's request, and the constraint such as meet its service quality (QoS), to complete
Entire scheduling process.Computation-intensive cloud workflow is made of multiple subtasks with relation of interdependence, therefore, whole
During a workflow schedule, the execution time of task is not only considered, it is also necessary to meet the dependence constraint between task
And make the execution span time (makespan) of entire workflow most short.Different Task Assigned Policies will have a direct impact on entirely
How the execution time of cloud workflow and cost for cloud workflow task distribute most suitable computing resource, and are meeting task
Between logic the problem of realizing its regulation goal, becoming each cloud service provider urgent need to resolve while rely on constraint.
Cloud workflow schedule is a kind of typical NP-hard problem, currently, mainly using heuritic approach and random search
Algorithm solves.Heuritic approach is broadly divided into list scheduling, Task Duplication, set of tasks cluster etc., such as HEFT, MIN-MIN,
MIN-MAX is not easy to find near-optimum solution.Random search algorithm mainly includes genetic algorithm, particle swarm algorithm, Immune Evolutionary Algorithm
Deng essence is to design a kind of efficient search strategy.Wherein, the evolution algorithms such as genetic algorithm have global search advantage and keep away
Exempt to fall into the ability of local optimum, but search time is too long, influences the real-time of algorithm；Colony intelligence optimization algorithm has convergence speed
Degree is fast, and adaptation is wide, but is a lack of effective local search mechanism.
Summary of the invention
The purpose of the present invention is to solve computation-intensive cloud scheduling problems, propose a kind of based on improvement cat owl
The computation-intensive cloud workflow schedule method of searching algorithm.Its basic thought is: using cat owl searching algorithm, works cloud
All dependence task carry out traversal search to the different scheduling schemes that resources of virtual machine maps in stream, and finding has minimum workflow
Execute the scheduling scheme of span time.Meanwhile the characteristics of according to cloud workflow, existing cat owl searching algorithm is changed
Into first is that Strength Changes amount is defined as optimal solution (i.e. optimal scheduling scheme) to other differences by sound intensity inverse square law
The influence size of individual (or scheduling scheme), and automatic adjusument is carried out to the optimizing step-length of Different Individual with this, thus significantly
Improve the search efficiency of optimal solution；Second is that the characteristics of being directed to cloud workflow schedule problem, has modified the search direction of individual, with
It avoids generating excessive trivial solution, and approaches all individuals directly gradually to optimal solution according to different step-lengths, to improve
The stability of individual solution, the speed of searching optimization for improving entire algorithm；Third is that being easily trapped into part for colony intelligence optimization algorithm
Optimal problem is introduced using the variation thought of evolutionary computation by increasing Mutation Strategy in population iteration update mechanism
Randomness, and when optimal solution iteration l times not yet updates changes reflecting for individual tasks and virtual machine in some scheduling schemes at random
Relationship is penetrated, to jump out local optimum, finds global better scheduling scheme.
The method of the present invention the following steps are included:
Step 1: the computation-intensive cloud Work flow model to be dispatched that input user submits and its dependence times for being included
Business set, for the virtual machine set of lease；
Step 2: the process that will be executed on each cloud workflow subtask scheduling to most suitable virtual machine, is modeled as standard
Minimum value Solve problems.Its regulation goal are as follows: the execution span time makespan for optimizing entire cloud workflow makes all Yun works
It is most short that work stream task execution finishes the time it takes.
Step 3: solving appointing under cloud computing environment using the cat owl searching algorithm based on sound intensity inverse square law
Business-scheduling virtual machine problem.Iterative process the following steps are included:
The basic parameter of step 1, initialization algorithm, including step parameter β, scheduling scheme number M and greatest iteration time
Number Iteration, optimizing the number of iterations bestNum when occurring optimal solution for the first time；
Step 2 initializes each scheduling scheme using equally distributed random number；
During step 3, algorithm iteration, when the number of iterations t is less than maximum number of iterations Iteration, t=t+1 turns
Step 4；When the number of iterations t is greater than or equal to maximum number of iterations Iteration, 8 are gone to step；
Step 4, according to cloud Work flow model, i.e. dependence between subtask, calculate when all dispatching parties in former generation
The workflow execution span time makespan of case；
Step 5 is found when scheduling scheme optimal in former generation.If optimal solution has update, bestNum=t.It updates every
A scheduling scheme and the range information of current optimal solution and the Strength Changes amount of each scheduling scheme；
Step 6 judges whether that iteration l Dai Erqi optimal solution does not still update, if t-bestNum > l, uses
The thought of hereditary variation changes the virtual machine mapping relations of any position of any individual at random, goes to step 7；Otherwise, directly turn
Step 7；
Step 7, according to Strength Changes amount, all scheduling schemes when former generation are updated, and return step 3；
Step 8 finds optimal scheduling scheme, the task and virtual machine mapping relations provided according to scheduling scheme, by work
It is bound with virtual machine stream subtask.
Beneficial effect
The present invention can effectively overcome in existing method that optimal solution search randomness is big, is easy to fall into local optimum and convergence
Slow-footed disadvantage promotes search efficiency, shortens search time, can search out more preferably scheduling scheme within the shorter time,
The overall time expense for reducing workflow schedule, specifically includes following three points:
1, cat owl searching algorithm is applied to scheduling problem for the first time, provides one for computation-intensive cloud workflow schedule
The new solution route of item.
2, by being improved to existing cat owl searching algorithm, that is, population recruitment iterative formula is modified, is effectively reduced
The randomness of optimal solution search, improves search efficiency, searching process is made to have more goal orientation.
3, by the population iteration update mechanism of existing cat owl searching algorithm apply hereditary variation thought, introduce with
Machine effectively prevents the case where search falls into local optimum, and can find the overall situation within the shorter time and more preferably dispatch
Scheme improves convergence speed of the algorithm, improves cloud workflow schedule performance.
Detailed description of the invention
Fig. 1 is the cloud workflow schedule method flow of the present invention based on cat owl searching algorithm.
Fig. 2 is the Montage workflow of simple 17 task.
Fig. 3 is algorithms of different for Montage_25 execution span time minimum value change procedure.
Fig. 4 is algorithms of different for Montage_50 execution span time minimum value change procedure.
Fig. 5 is algorithms of different for Montage_100 execution span time minimum value change procedure.
Fig. 6 is that algorithms of different is directed to Montage_25 searching process.
Fig. 7 is that algorithms of different is directed to Montage_50 searching process.
Fig. 8 is that algorithms of different is directed to Montage_100 searching process.
Fig. 9 is the program runtime of algorithms of different.
Figure 10 is the minimum value of the execution span time of algorithms of different optimal scheduling scheme.
Specific embodiment
With reference to the accompanying drawings and examples, the method for the present invention is described in detail.
A kind of computation-intensive cloud workflow schedule method based on cat owl searching algorithm, as shown in Figure 1, including following
Step:
Step 1: computation-intensive Work flow model to be dispatched and its corresponding dependence subtask collection that input user submits
It closes, for the resources of virtual machine set of lease；
For computation-intensive cloud workflow schedule problem, by cloud workflow be described as a directed acyclic graph G=(T,
E), in which: T is the set of directed acyclic graph interior joint, indicates n task in cloud workflow, i.e. T={ T1,T2,……,Ti,
Tj,……,Tn, wherein i, j=1,2 ... ..., n；TentryFor entrance task, TexitFor export task；E is in directed acyclic graph
The set of directed edge, E={ < Ti,Tj> | Ti,Tj∈ T }, directed edge Ti→TjIndicate father's task TiWith its subtask TjBetween
Dependence, TjOnly in his father's task TiIt can just start to execute after the completion.As shown in Fig. 2, a letter with 17 tasks
Single Montage workflow, i.e. task quantity n=17, the dependence between directed arc (arrow) expression task in cloud work flow diagram
Relationship E, the file size for needing to transmit between corresponding digital representation father, subtask on directed arc.
Indicate that virtual machine, m indicate that the virtual machine total number leased for user, resources of virtual machine set can indicate with VM
Are as follows: VM={ VM1,VM2,……,VMk,……,VMm, wherein k=1,2 ..., m.Assuming that MIPS indicate calculate equipment is per second can
Million grades of machine language instruction numbers of processing, then virtual machine VMkProcessing speed can use MIPS (VMk) indicate.
Step 2: the process that will be executed on each cloud workflow subtask scheduling to most suitable virtual machine, is modeled as standard
Minimum value Solve problems.It is specific as follows:
Assuming that: (1) all subtasks in set of tasks be all atomic task, i.e., each task can not be split as again more
The task of small grain size；(2) each virtual machine can only handle a task in the same time, i.e., only when virtual machine is finished currently
When handling for task, the request of new task just can receive；(3) task execution can not interrupt, i.e., each subtask is at it
When executing or calculated on the virtual machine leased, do not allow to be interrupted by other task requests.
Regulation goal is the execution span time expense for optimizing entire cloud workflow task, that is, makes all cloud workflows
The total time makespan that subtask is finished spent is most short；
Constraint condition is the number m, i.e. n > m that the number n of cloud workflow subtask is greater than the virtual machine for lease；
Definition task TiIn virtual machine VMkOn execution time ETC (Ti,VMk) and father's task TiWith subtask TjBetween
Transmission time TT (Ti,Tj) it is as follows:
Wherein, Length (Ti) indicate task TiCommand length, MIPS (VMk) indicate virtual machine VMkProcessing speed；
transferSize(Ti,Tj) indicate father's task TiWith subtask TjBetween transmission file size, bandwidth indicates virtual
The bandwidth of communication line between machine.
Define task T in cloud workflowiAt the beginning of ST (Ti) and deadline FT (Ti) it is as follows:
FT(Ti)=ST (Ti)+ETC(Ti,VMk) (4)
Wherein, ST (Tentry) indicate entrance task TentryAt the beginning of, ST (Ti) indicate task TiAt the beginning of, FT
(Tp)、FT(Ti) respectively indicate task TpAnd its subtask TiDeadline, avail (VMk) indicate virtual machine VMkIt is available when
Between, predr (Ti) indicate task TiAll father's tasks constitute set, TT (Tp,Ti) indicate task TpAnd its subtask TiIt
Between transmission time.
Total execution span time expense makespan of cloud workflow, with subtasks all in the cloud workflow deadline
Maximum value indicates, it may be assumed that
Step 3: solving cloud workflow task-virtual machine with the cat owl searching algorithm based on sound intensity inverse square law
Scheduling problem, iterative process the following steps are included:
The basic parameter of step 1, initialization algorithm, including step parameter β, scheduling scheme number M and greatest iteration time
Number Iteration, optimizing the number of iterations bestNum when occurring optimal solution for the first time；
Step 2, M initial schedule scheme for generating for the 0th generation using equally distributed random numberWherein s=1,
2 ..., M:
Wherein, OLAnd OUIt is the 0th s-th of scheduling scheme of generation respectivelyMiddle task TiThe lower and upper limit of number, U (0,1) are
Equally distributed random number in section [0,1] range.
During step 3, algorithm iteration, when the number of iterations t is less than maximum number of iterations Iteration, t=t+1 turns
Step 4；When the number of iterations t is greater than or equal to maximum number of iterations Iteration, 8 are gone to step；
Step 4, according to the dependence between subtask each in cloud Work flow model, to each of being generated as former generation t
Scheduling schemeCalculate its corresponding workflow execution span time expenseAnd it is completed with all subtasks
The maximum value of time indicates, as shown in formula (7):
Wherein,For task TiIn scheduling schemeUnder deadline.
Step 5 is found when scheduling scheme optimal in former generation.If optimal solution has update, bestNum=t.To each
Scheduling schemeIt updatesWith the range information of current optimal solutionIt updates simultaneouslyStrength Changes amount
In formula, V indicates that global optimum's scheduling scheme, makespan (V) indicate that the corresponding workflow of optimal scheduling scheme is held
Row span time, random indicate [0,1) random number；
Step 6, judge whether iteration l generation and also optimal solution still do not update.If t-bestNum > l is used
The thought of hereditary variation changes task-virtual machine mapping relations of any position of any individual at random, then goes to step 7；It is no
Then, 7 are directly gone to step；
Step 7, according to Strength Changes amount, to each scheduling scheme for working as former generation tIt is updated, and returns by formula (10)
Step 3:
Wherein,Indicate s-th of scheduling scheme in t generation.
Step 8 finds optimal scheduling scheme, the task and virtual machine mapping relations provided according to scheduling scheme, by work
It is bound with virtual machine stream subtask.
Embodiment
In order to examine the effect proposed by the present invention for carrying out cloud workflow schedule using improved cat owl searching algorithm (OSA)
Fruit, present invention uses cloud computing simulation modelling tools WorkflowSim, to simulate a cloud computation data center, and lead to
It crosses and the algorithm for estimating of workflow execution span time makespan is optimized, improve the computational efficiency of makespan.It is real
It tests and has chosen most common intelligent optimization algorithm and compare, such as ant group algorithm (ACO), particle swarm algorithm (PSO), genetic algorithm
(GA)。
It for the Montage Work flow model of different scales, is tested, is chosen respectively using 10 identical virtual machines
The average execution span time that workflow minimum executes the corresponding workflow of individual of span time and every generation is scheduling performance
Index, comes the generalization ability and performance of measure algorithm, and comparing result is as shown in Figures 3 to 10.
By Fig. 3, Fig. 4, Fig. 5 it is found that being directed to the Work flow model of different scales, the dispatching algorithm based on cat owl search is equal
Preferable solution can be found, and only needs the less number of iteration that can find approximate optimal solution.By Fig. 6, Fig. 7, Fig. 8 it is found that OSA is calculated
Method speed of searching optimization is very fast, and is not easy to fall into local optimum.As shown in Figure 9, the Algorithms T-cbmplexity of OSA is lower.It can by Figure 10
Know, for small-sized workflow, OSA algorithm can find preferable solution；For medium-sized and large-scale workflow, the improved OSA of the present invention
Algorithm can find a preferably scheduling scheme, and executing span time is respectively 95.03,174.15, the optimizing with GA algorithm
As a result identical.But as seen from Figure 9, the execution time overhead of GA algorithm is larger, is the 2 times or more of OSA algorithm.
Claims (6)
1. a kind of computation-intensive cloud workflow schedule method based on cat owl searching algorithm, which is characterized in that including following
Step:
Step 1: the computation-intensive cloud Work flow model to be dispatched that input user submits and its dependence subtask collection for being included
It closes, for the virtual machine set of lease；
Step 2: the process that will be executed on each cloud workflow subtask scheduling to most suitable virtual machine, is modeled as the minimum of standard
It is worth Solve problems, regulation goal are as follows: the execution span time makespan for optimizing entire cloud workflow makes all cloud workflows
It is most short that task execution finishes the time it takes；
Step 3: solving task-void under cloud computing environment using the cat owl searching algorithm based on sound intensity inverse square law
Quasi- machine scheduling problem, comprising the following steps:
The basic parameter of step 1, initialization algorithm, including step parameter β, scheduling scheme number M and maximum number of iterations
Iteration, optimizing the number of iterations bestNum when occurring optimal solution for the first time；
Step 2 initializes each scheduling scheme using equally distributed random number；
During step 3, algorithm iteration, when the number of iterations t is less than maximum number of iterations Iteration, t=t+1 is gone to step
4；When the number of iterations t is greater than or equal to maximum number of iterations Iteration, 8 are gone to step；
Step 4, according to cloud Work flow model, i.e. dependence between subtask, calculate when all scheduling schemes in former generation
Workflow execution span time makespan；
Step 5 is found when scheduling scheme optimal in former generation；If optimal solution has update, bestNum=t updates each tune
Degree scheme and the range information of current optimal solution and the Strength Changes amount of each scheduling scheme；
Step 6 judges whether that iteration l Dai Erqi optimal solution does not still update, if t-bestNum > l, uses heredity
The thought of variation changes the virtual machine mapping relations of any position of any individual at random, goes to step 7；Otherwise, it directly goes to step
7；
Step 7, according to Strength Changes amount, all scheduling schemes when former generation are updated, and return step 3；
Step 8 finds optimal scheduling scheme, the task and virtual machine mapping relations provided according to scheduling scheme, by workflow
Task is bound with virtual machine.
2. a kind of computation-intensive cloud workflow schedule method based on cat owl searching algorithm as described in claim 1,
It is characterized in that, the step 1, cloud workflow is specifically described as a directed acyclic graph G=(T, E), in which: T is oriented nothing
The set of ring figure interior joint indicates n task in cloud workflow, T={ T1,T2,……,Tn, TentryFor entrance task,
TexitFor export task；E is the set of directed edge in cloud Work flow model, E={ (Ti,Tj)|Ti,Tj∈ T }, directed edge Ti→Tj
Indicate father's task TiWith subtask TjBetween dependence, TjOnly in TiIt just can be performed after the completion；Virtual machine, m are indicated with VM
Indicate the virtual machine total number leased for user, resources of virtual machine set expression is VM={ VM1,VM2,……,VMk,……,
VMm, wherein k=1,2 ..., m；MIPS indicates calculating equipment accessible million grades of machine language instruction numbers per second, then MIPS
(VMk) indicate virtual machine VMkProcessing speed.
3. a kind of computation-intensive cloud workflow schedule method based on cat owl searching algorithm as described in claim 1,
It is characterized in that, the modeling method of the step 2 are as follows:
If all subtasks in set of tasks are all atomic tasks, i.e., each task can not be split as appointing for smaller particle size again
Business；Each virtual machine can only handle a task in the same time, i.e., only when virtual machine is finished appointing of being presently processing
When business, the request of new task just can receive；Task execution can not interrupt, i.e., each subtask is on the virtual machine that it is leased
When executing or being calculated, do not allow to be interrupted by other task requests；
Constraint condition is the number m, i.e. n > m that the number n of cloud workflow subtask is greater than the virtual machine for lease；
Definition task TiIn virtual machine VMkOn execution time ETC (Ti,VMk) and father's task TiWith subtask TjBetween biography
Defeated time TT (Ti,Tj) it is as follows:
Wherein, Length (Ti) indicate task TiCommand length, MIPS (VMk) indicate virtual machine VMkProcessing speed；
transferSize(Ti,Tj) indicate father's task TiWith subtask TjBetween transmission file size, bandwidth indicates virtual
The bandwidth of communication line between machine；
Define task T in cloud workflowiAt the beginning of ST (Ti) and deadline FT (Ti) it is as follows:
FT(Ti)=ST (Ti)+ETC(Ti,VMk) (4)
Wherein, ST (Tentry) indicate entrance task TentryAt the beginning of, ST (Ti) indicate task TiAt the beginning of, FT (Tp)、
FT(Ti) respectively indicate task TpAnd its subtask TiDeadline, avail (VMk) indicate virtual machine VMkPot life,
predr(Ti) indicate task TiAll father's tasks constitute set, TT (Tp,Ti) indicate task TpAnd its subtask TiBetween
Transmission time；
Total execution span time expense makespan of cloud workflow, with the maximum of subtasks all in cloud workflow deadline
Value indicates, it may be assumed that
4. a kind of computation-intensive cloud workflow schedule method based on cat owl searching algorithm as described in claim 1,
It is characterized in that, step 2 described in step 3 generates each initial schedule scheme method are as follows: is produced using equally distributed random number
The M initial schedule scheme in raw 0th generationWherein s=1,2 ..., M:
Wherein, OLAnd OUIt is the 0th s-th of scheduling scheme of generation respectivelyMiddle task TiThe lower and upper limit of number, U (0,1) is section
Equally distributed random number in [0,1] range.
5. a kind of computation-intensive cloud workflow schedule method based on cat owl searching algorithm as described in claim 1,
It is characterized in that, the method that step 5 described in step 3 updates the range information of each scheduling scheme and current optimal solution are as follows:
If optimal solution has update, bestNum=t；To each scheduling schemeIt updatesBelieve at a distance from current optimal solution
BreathIt updates simultaneouslyStrength Changes amount
Wherein, V indicates global optimum's scheduling scheme, makespan (V) indicate the corresponding workflow execution of optimal scheduling scheme across
Spend the time, random indicate [0,1) random number.
6. a kind of computation-intensive cloud workflow schedule method based on cat owl searching algorithm as described in claim 1,
It is characterized in that, the update method of step 7 described in step 3 are as follows:
Wherein,Indicate that i-th of scheduling scheme in t generation, β indicate step parameter, IciIndicate that the intensity of each scheduling scheme becomes
Change amount, V indicate optimal scheduling scheme.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811336040.8A CN109710372B (en) | 2018-11-12 | 2018-11-12 | Calculation intensive cloud workflow scheduling method based on owl search algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811336040.8A CN109710372B (en) | 2018-11-12 | 2018-11-12 | Calculation intensive cloud workflow scheduling method based on owl search algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109710372A true CN109710372A (en) | 2019-05-03 |
| CN109710372B CN109710372B (en) | 2020-10-13 |
Family
ID=66254212
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811336040.8A Active CN109710372B (en) | 2018-11-12 | 2018-11-12 | Calculation intensive cloud workflow scheduling method based on owl search algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109710372B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112685138A (en) * | 2021-01-08 | 2021-04-20 | 北京理工大学 | Multi-workflow scheduling method based on multi-population hybrid intelligent optimization in cloud environment |
| CN112905307A (en) * | 2021-04-22 | 2021-06-04 | 北京理工大学 | Multi-target cloud workflow scheduling method based on lean and rich optimization algorithm |
| CN113157381A (en) * | 2020-05-14 | 2021-07-23 | 北京理工大学 | Multi-workflow scheduling method based on non-dominated sorting and owl search |
| CN113220414A (en) * | 2021-04-22 | 2021-08-06 | 北京理工大学 | Cloud workflow scheduling method based on improved lean-rich optimization algorithm |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8296765B2 (en) * | 2010-07-27 | 2012-10-23 | Kurdi Heba A | Method of forming a personal mobile grid system and resource scheduling thereon |
| CN103067524A (en) * | 2013-01-18 | 2013-04-24 | 浪潮电子信息产业股份有限公司 | Ant colony optimization computing resource distribution method based on cloud computing environment |
| CN103412792A (en) * | 2013-07-18 | 2013-11-27 | 成都国科海博计算机系统有限公司 | Dynamic task scheduling method and device under cloud computing platform environment |
| CN106648831A (en) * | 2016-12-08 | 2017-05-10 | 杭州电子科技大学 | Cloud workflow scheduling method based on firefly algorithm and dynamic priority algorithm |
| CN108133260A (en) * | 2018-01-17 | 2018-06-08 | 浙江理工大学 | The workflow schedule method of multi-objective particle swarm optimization based on real-time status monitoring |
-
2018
- 2018-11-12 CN CN201811336040.8A patent/CN109710372B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8296765B2 (en) * | 2010-07-27 | 2012-10-23 | Kurdi Heba A | Method of forming a personal mobile grid system and resource scheduling thereon |
| CN103067524A (en) * | 2013-01-18 | 2013-04-24 | 浪潮电子信息产业股份有限公司 | Ant colony optimization computing resource distribution method based on cloud computing environment |
| CN103412792A (en) * | 2013-07-18 | 2013-11-27 | 成都国科海博计算机系统有限公司 | Dynamic task scheduling method and device under cloud computing platform environment |
| CN106648831A (en) * | 2016-12-08 | 2017-05-10 | 杭州电子科技大学 | Cloud workflow scheduling method based on firefly algorithm and dynamic priority algorithm |
| CN108133260A (en) * | 2018-01-17 | 2018-06-08 | 浙江理工大学 | The workflow schedule method of multi-objective particle swarm optimization based on real-time status monitoring |
Non-Patent Citations (2)
| Title |
|---|
| SAEID ABRISHAMI等: "Deadline-constrained workflow scheduling algorithms for Infrastructure as a", 《FUTURE GENERATION COMPUTER SYSTEMS》 * |
| 罗干: "基于启发式算法的云工作流任务调度方法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113157381A (en) * | 2020-05-14 | 2021-07-23 | 北京理工大学 | Multi-workflow scheduling method based on non-dominated sorting and owl search |
| CN113157381B (en) * | 2020-05-14 | 2022-07-05 | 北京理工大学 | Multi-workflow scheduling method based on non-dominated sorting and owl search |
| CN112685138A (en) * | 2021-01-08 | 2021-04-20 | 北京理工大学 | Multi-workflow scheduling method based on multi-population hybrid intelligent optimization in cloud environment |
| CN112685138B (en) * | 2021-01-08 | 2022-08-23 | 北京理工大学 | Multi-workflow scheduling method based on multi-population hybrid intelligent optimization in cloud environment |
| CN112905307A (en) * | 2021-04-22 | 2021-06-04 | 北京理工大学 | Multi-target cloud workflow scheduling method based on lean and rich optimization algorithm |
| CN113220414A (en) * | 2021-04-22 | 2021-08-06 | 北京理工大学 | Cloud workflow scheduling method based on improved lean-rich optimization algorithm |
| CN112905307B (en) * | 2021-04-22 | 2022-07-22 | 北京理工大学 | Multi-target cloud workflow scheduling method based on lean and rich optimization algorithm |
| CN113220414B (en) * | 2021-04-22 | 2022-09-09 | 北京理工大学 | Cloud workflow scheduling method based on improved lean-rich optimization algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109710372B (en) | 2020-10-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110737529B (en) | Short-time multi-variable-size data job cluster scheduling adaptive configuration method | |
| Cheng et al. | Cost-aware job scheduling for cloud instances using deep reinforcement learning | |
| CN108009023B (en) | Task scheduling method based on BP neural network time prediction in hybrid cloud | |
| CN106055395B (en) | Deadline constrains workflow schedule method in a kind of cloud environment based on ant colony optimization algorithm | |
| Zuo et al. | A multi-objective hybrid cloud resource scheduling method based on deadline and cost constraints | |
| CN109710372A (en) | A kind of computation-intensive cloud workflow schedule method based on cat owl searching algorithm | |
| Liu et al. | Resource preprocessing and optimal task scheduling in cloud computing environments | |
| CN110096349A (en) | A kind of job scheduling method based on the prediction of clustered node load condition | |
| Kamalinia et al. | Hybrid task scheduling method for cloud computing by genetic and DE algorithms | |
| CN112416585B (en) | Deep learning-oriented GPU resource management and intelligent scheduling method | |
| CN109634742B (en) | Time constraint scientific workflow optimization method based on ant colony algorithm | |
| CN105373432B (en) | A kind of cloud computing resource scheduling method based on virtual resource status predication | |
| Chakravarthi et al. | TOPSIS inspired budget and deadline aware multi-workflow scheduling for cloud computing | |
| Malik et al. | Comparison of task scheduling algorithms in cloud environment | |
| CN112231081B (en) | PSO-AHP-based monotonic rate resource scheduling method and system in cloud environment | |
| CN111309472A (en) | Online virtual resource allocation method based on virtual machine pre-deployment | |
| CN116467076A (en) | Multi-cluster scheduling method and system based on cluster available resources | |
| CN108845886A (en) | Cloud computing energy consumption optimization method and system based on phase space | |
| Yang et al. | A fully hybrid algorithm for deadline constrained workflow scheduling in clouds | |
| Yang et al. | Budget and SLA aware dynamic workflow scheduling in cloud computing with heterogeneous resources | |
| CN112306642B (en) | Workflow scheduling method based on stable matching game theory | |
| Li et al. | A QoS-based scheduling algorithm for instance-intensive workflows in cloud environment | |
| Ye et al. | Dynamic scheduling stochastic multiworkflows with deadline constraints in clouds | |
| CN105872109A (en) | Load running method of cloud platform | |
| Filippini et al. | Hierarchical Scheduling in on-demand GPU-as-a-Service 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |