CN104503833A - Task scheduling optimization method and device - Google Patents
Task scheduling optimization method and device Download PDFInfo
- Publication number
- CN104503833A CN104503833A CN201410820409.8A CN201410820409A CN104503833A CN 104503833 A CN104503833 A CN 104503833A CN 201410820409 A CN201410820409 A CN 201410820409A CN 104503833 A CN104503833 A CN 104503833A
- Authority
- CN
- China
- Prior art keywords
- stream
- work
- finishing
- input parameter
- consuming time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a task scheduling optimization method and device. The method comprises the following steps: receiving an input parameter; judging the times of recording the input parameters in a training module; if the times is smaller than a preset threshold value, finishing a workflow by adopting a random task execution sequence according to the dependency relationship among preset tasks, forming an incidence relation among the input parameter, the task execution sequence of finishing the workflow, and data on time consumed to finish the workflow, recording the incidence relationship in the training module in a newly increased manner; otherwise, selecting the task execution sequence with the shortest time consumed on average in the training module to finish the workflow, forming an incidence relation among the input parameter, the task execution sequence of finishing the workflow, and the data on time consumed to finish the workflow, and recording the incidence relation in the training module in a newly increased manner. A developer does not need to care about task execution states and complex dependency relationships, learns the task processing efficiency by himself or herself, and optimizes the execution path.
Description
Technical field
The present invention relates to computer program, be specifically related to task scheduling optimization method and device.
Background technology
The executive mode of current task scheduling has following 3 kinds:
1, order performs, and all task serializations perform, and this theoretical model is very simple, only has a task to complete, and then unconditionally performs next task.
2, concurrence performance, in workflow, multiple task can carry out asynchronous process simultaneously, to improve the concurrency of process.
3, preposition execution, waits for that all branches are completed merging task, and just can carry out next step task, this pattern is with preposition task-driven, and each task has preposition dependence.
In ordering system or commending system, need the Various types of data used from various system, because all data processings are all asynchronous, and each data is the need of asking to be the execution of taking over according to some request of data above, because involved request task has tens kinds, become very complicated and poor efficiency according to the how reasonable scheduler task efficiently of difference request.
Summary of the invention
An object of the present invention is to propose a kind of task scheduling optimization method, and it can solve the problem of the very complicated and poor efficiency of current task scheduling.
One of in order to achieve the above object, the technical solution adopted in the present invention is as follows:
Task scheduling optimization method, it comprises the following steps:
Step 1, receive an input parameter;
Step 2, judge the number of times that described input parameter records in training module, if described number of times is less than predetermined threshold value, then perform step 3, otherwise, perform step 4; Described training module records input parameter, the incidence relation of the tasks carrying order of stream of finishing the work and the data consuming time of stream of finishing the work;
Dependence between each task that step 3, basis are preset, random tasks carrying order is adopted to finish the work stream, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module;
Step 4, in training module, choose on average the shortest tasks carrying consuming time order to finish the work stream, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
Preferably, described step 4 comprises following sub-step:
Step 41, according to described input parameter in conjunction with data consuming time, what calculate corresponding to each tasks carrying order in training module is on average consuming time;
Step 42, choose on average the shortest tasks carrying consuming time order and to finish the work stream;
Step 43, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
Preferably, described dependence comprise adjustable order relation, can one or more in concurrency relation and alternative relation.
The present invention also proposes a kind of task scheduling optimization device corresponding with one of above-mentioned purpose, and it comprises with lower module:
Receiver module, for receiving an input parameter;
Judge module, for judging the number of times that described input parameter records in training module, if described number of times is less than predetermined threshold value, then performs randomized blocks, otherwise, perform and optimize module; Described training module records input parameter, the incidence relation of the tasks carrying order of stream of finishing the work and the data consuming time of stream of finishing the work;
Randomized blocks, for the dependence between each task that basis is preset, random tasks carrying order is adopted to finish the work stream, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module;
Optimize module, to finish the work stream for choosing on average the shortest tasks carrying consuming time order in training module, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
Preferably, described optimization module comprises following submodule:
Computing module, for according to described input parameter in conjunction with data consuming time, what calculate corresponding to each tasks carrying order in training module is on average consuming time;
Choosing module, to finish the work stream for choosing on average the shortest tasks carrying consuming time order;
Logging modle, for the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and is recorded to described incidence relation in described training module in newly-increased mode.
Two of object of the present invention is to propose another kind of task scheduling optimization method, and it can solve the problem of the very complicated and poor efficiency of current task scheduling.
In order to achieve the above object two, the technical solution adopted in the present invention is as follows:
Task scheduling optimization method, it comprises the following steps:
Step 1, receive an input parameter;
The rule that step 2, basis are preset is sorted out described input parameter, to obtain the classification belonging to described input parameter;
Step 3, judge the number of times that the classification belonging to described input parameter records in training module, if described number of times is less than predetermined threshold value, then perform step 4, otherwise, perform step 5; Described training module records the incidence relation of the classification belonging to input parameter, the tasks carrying order of stream of finishing the work and the data consuming time of stream of finishing the work;
Dependence between each task that step 4, basis are preset, random tasks carrying order is adopted to finish the work stream, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module;
Step 5, in training module, choose on average the shortest tasks carrying consuming time order to finish the work stream, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
Preferably, described step 5 comprises following sub-step:
Step 51, classification belonging to described input parameter are in conjunction with data consuming time, and what calculate corresponding to each tasks carrying order in training module is on average consuming time;
Step 52, choose on average the shortest tasks carrying consuming time order and to finish the work stream;
Step 53, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
Preferably, described dependence comprise adjustable order relation, can one or more in concurrency relation and alternative relation.
The present invention also proposes a kind of task scheduling optimization device corresponding with two of above-mentioned purpose, and it comprises with lower module:
Receiver module, for receiving an input parameter;
Classifying module, for sorting out described input parameter according to the rule preset, to obtain the classification belonging to described input parameter;
Judge module, the number of times that the classification for judging belonging to described input parameter records in training module, if described number of times is less than predetermined threshold value, then performs randomized blocks, otherwise, perform and optimize module; Described training module records the incidence relation of the classification belonging to input parameter, the tasks carrying order of stream of finishing the work and the data consuming time of stream of finishing the work;
Randomized blocks, for the dependence between each task that basis is preset, random tasks carrying order is adopted to finish the work stream, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module;
Optimize module, to finish the work stream for choosing on average the shortest tasks carrying consuming time order in training module, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
Preferably, described optimization module comprises following submodule:
Computing module, for the classification belonging to described input parameter in conjunction with data consuming time, what calculate corresponding to each tasks carrying order in training module is on average consuming time;
Choosing module, to finish the work stream for choosing on average the shortest tasks carrying consuming time order;
Logging modle, for the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and is recorded to described incidence relation in described training module in newly-increased mode.
The present invention has following beneficial effect:
Developer is without the need to the dependence between the serial of each state and complexity of being concerned about tasks carrying, the task such as parallel, and self-teaching is carried out to task treatment effeciency, thus optimization execution route, use simple, can significantly improve concurrent processing speed, can improve the sharpness of logical design, developer can be absorbed in each task of design, and without the need to the priority of being concerned about each tasks carrying and state relation, reduce design difficulty and the complexity of workflow.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of the task scheduling optimization method of the embodiment of the present invention one;
Fig. 2 is the process flow diagram of the task scheduling optimization method of the embodiment of the present invention two.
Embodiment
Below, by reference to the accompanying drawings and embodiment, the present invention is described further.
Embodiment one
As shown in Figure 1, a kind of task scheduling optimization method, it comprises the following steps:
Step S1, receive an input parameter.Described input parameter can be scope class parameter, and as Price Range, 10-100,200-500 etc., also can be enumerate class parameter, can also be the parameter affecting task quantity.Such as, for ordering system, user wants to sort to the product in certain Price Range, then Price Range 200-500 can be inputted, if conventionally, ordering system will call data in other system (as data system etc.) and finish the work stream according to pre-set tasks carrying order, finally shows ranking results.
Step S2, judge the number of times that described input parameter records in training module, if described number of times is less than predetermined threshold value, represents the frequency of training of described input parameter in training module and also do not reach requirement, then perform step S3, otherwise, perform step S4.Described training module records input parameter, the incidence relation of the tasks carrying order of stream of finishing the work and the data consuming time of stream of finishing the work.
Dependence between each task that step S3, basis are preset, random tasks carrying order is adopted to finish the work stream, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
Wherein, described dependence comprise adjustable order relation, can in concurrency relation and alternative relation.
Namely adjustable order relation shows that the order between task is adjustable: can adjust task order and executed in parallel for some task without the situation that direct sequencing and precondition meet simultaneously and carry out overall usefulness evaluation and test.
Can concurrency relation namely show between task can executed in parallel: meet precondition and task without direct relation can carry out parallelization process for multiple.
Namely alternative relation shows can replace between task: carry out different trials to the impact of the last efficiency of comparison under different parameters for alternative task.
According to above-mentioned dependence, just tasks carrying order can be carried out at random, thus obtain the treatment effeciency of each paths.
Step S4, in training module, choose on average the shortest tasks carrying consuming time order to finish the work stream, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
Concrete, described step S4 specifically comprises following sub-step:
Step S41, according to described input parameter in conjunction with data consuming time, what calculate corresponding to each tasks carrying order in training module is on average consuming time.Such as, tasks carrying order A (task 1-task 2-task 3) has 3 data records consuming time, then get on average consuming time as this tasks carrying order A of this mean value of 3 times, tasks carrying order B (task 2-task 3-task 1) has 4 data records consuming time, gets on average consuming time as this tasks carrying order B of this mean value of 4 times.
Step S42, choose on average the shortest tasks carrying consuming time order and to finish the work stream.Such as, between comparison task execution sequence A, B, which is on average consuming time the shortest, chooses the on average the shortest execution route as this workflow consuming time.
Step S43, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
The present embodiment also proposes a kind of task scheduling optimization device corresponding with said method, and it comprises with lower module:
Receiver module, for receiving an input parameter;
Judge module, for judging the number of times that described input parameter records in training module, if described number of times is less than predetermined threshold value, then performs randomized blocks, otherwise, perform and optimize module; Described training module records input parameter, the incidence relation of the tasks carrying order of stream of finishing the work and the data consuming time of stream of finishing the work;
Randomized blocks, for the dependence between each task that basis is preset, random tasks carrying order is adopted to finish the work stream, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module;
Optimize module, to finish the work stream for choosing on average the shortest tasks carrying consuming time order in training module, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
Concrete, described optimization module comprises following submodule:
Computing module, for according to described input parameter in conjunction with data consuming time, what calculate corresponding to each tasks carrying order in training module is on average consuming time;
Choosing module, to finish the work stream for choosing on average the shortest tasks carrying consuming time order;
Logging modle, for the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and is recorded to described incidence relation in described training module in newly-increased mode.
Known by the description of the present embodiment, the process that traditional task scheduling mode does not have self-optimization to learn, need the flow process fixing workflow, not considering can the inconsistent problem of energy efficiency according to different task input parameters in reality, resource can not be made full use of, some tasks all may be caused unnecessarily to wait for, when task is many, the process of the scheduling of maintenance task is too complicated.
The present embodiment based on the shortcoming for the emphasis of efficiency and the process of conventional operation stream mode task in practical application, the present embodiment done following some to optimize:
A. the efficiency data collection study of tasks carrying.
B. according to the tasks carrying efficiency learning data collected and routing data, carry out model and perform calculating.
C. the separating purification design of task, mainly allows each task only need to be concerned about which task execution should notify after being over, and for process when coming from the completion notice of different previous task.This design does not need the trend paying close attention to very much flow process, only needs to be concerned about the dependence of each task itself and the flow direction of follow-up task.
D. lower to efficiency according to execution efficiency learning data task can do fine-grained deconsolidation process.
By the self-teaching for execution efficiency be supplied on the one hand developer intuitively tasks carrying situation be used for doing the analysis of task bottleneck, the Work flow model being mainly used in task on the other hand calculates the mode of optimal task schedule.
Embodiment two
As shown in Figure 2, a kind of task scheduling optimization method, it comprises the following steps:
Step S1, receive an input parameter.
The rule that step S2, basis are preset is sorted out described input parameter, to obtain the classification belonging to described input parameter.Because the possibility of the kind of input parameter and value is too many, can first classify to input parameter.The parameter etc. that described input parameter can be scope class parameter, enumerate class parameter, affect task quantity.Scope class parameter: the size of the selection Main Basis range difference of range parameter is determined, difference is classified as a class in a certain scope, such as Price Range, difference is the playback category-A of 0-100, is so that 30-100 (difference 70) and 500-560 (difference 60) can be classified as category-A for input parameter.Enumerate class parameter: enumerating class parameter can determine whether wanting random packet as classification, if number can not choose each enumerated value at most as a class according to the number enumerated.Affect the parameter of task quantity: each parameter is as a class, and such as parameter A represents scene, may value be 1,2,3 for scene, so different tasks may be selected to perform for 1,2,3.
Step S3, judge the number of times that the classification belonging to described input parameter records in training module, if described number of times is less than predetermined threshold value, then perform step S4, otherwise, perform step S5; Described training module records the incidence relation of the classification belonging to input parameter, the tasks carrying order of stream of finishing the work and the data consuming time of stream of finishing the work.
Dependence between each task that step S4, basis are preset, random tasks carrying order is adopted to finish the work stream, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.Described dependence comprise adjustable order relation, can one or more in concurrency relation and alternative relation.
Step S5, in training module, choose on average the shortest tasks carrying consuming time order to finish the work stream, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
Concrete, described step S5 comprises following sub-step:
Step S51, classification belonging to described input parameter are in conjunction with data consuming time, and what calculate corresponding to each tasks carrying order in training module is on average consuming time;
Step S52, choose on average the shortest tasks carrying consuming time order and to finish the work stream;
Step S53, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
Embodiment also proposes a kind of task scheduling optimization device corresponding with said method, and it comprises with lower module:
Receiver module, for receiving an input parameter;
Classifying module, for sorting out described input parameter according to the rule preset, to obtain the classification belonging to described input parameter;
Judge module, the number of times that the classification for judging belonging to described input parameter records in training module, if described number of times is less than predetermined threshold value, then performs randomized blocks, otherwise, perform and optimize module; Described training module records the incidence relation of the classification belonging to input parameter, the tasks carrying order of stream of finishing the work and the data consuming time of stream of finishing the work;
Randomized blocks, for the dependence between each task that basis is preset, random tasks carrying order is adopted to finish the work stream, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module;
Optimize module, to finish the work stream for choosing on average the shortest tasks carrying consuming time order in training module, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
Concrete, described optimization module comprises following submodule:
Computing module, for the classification belonging to described input parameter in conjunction with data consuming time, what calculate corresponding to each tasks carrying order in training module is on average consuming time;
Choosing module, to finish the work stream for choosing on average the shortest tasks carrying consuming time order;
Logging modle, for the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and is recorded to described incidence relation in described training module in newly-increased mode.
The present invention is the efficiency optimization process of the treatment scheme to asynchronous task, have employed training pattern and carries out optimizing scheduling to task process.
The present invention adopts real data as the continuous self-optimization model of statistical basis, constantly makes model closer to optimized mode.
The present invention is optimized scheduling by adopting the classification of input parameter for the order of executing the task, and can reach the most efficient tasks carrying efficiency.
For a person skilled in the art, according to technical scheme described above and design, other various corresponding change and distortion can be made, and all these change and distortion all should belong within the protection domain of the claims in the present invention.
Claims (10)
1. task scheduling optimization method, is characterized in that, comprises the following steps:
Step 1, receive an input parameter;
Step 2, judge the number of times that described input parameter records in training module, if described number of times is less than predetermined threshold value, then perform step 3, otherwise, perform step 4; Described training module records input parameter, the incidence relation of the tasks carrying order of stream of finishing the work and the data consuming time of stream of finishing the work;
Dependence between each task that step 3, basis are preset, random tasks carrying order is adopted to finish the work stream, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module;
Step 4, in training module, choose on average the shortest tasks carrying consuming time order to finish the work stream, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
2. task scheduling optimization method as claimed in claim 1, it is characterized in that, described step 4 comprises following sub-step:
Step 41, according to described input parameter in conjunction with data consuming time, what calculate corresponding to each tasks carrying order in training module is on average consuming time;
Step 42, choose on average the shortest tasks carrying consuming time order and to finish the work stream;
Step 43, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
3. task scheduling optimization method as claimed in claim 1, is characterized in that, described dependence comprise adjustable order relation, can one or more in concurrency relation and alternative relation.
4. task scheduling optimization device, is characterized in that, comprises with lower module:
Receiver module, for receiving an input parameter;
Judge module, for judging the number of times that described input parameter records in training module, if described number of times is less than predetermined threshold value, then performs randomized blocks, otherwise, perform and optimize module; Described training module records input parameter, the incidence relation of the tasks carrying order of stream of finishing the work and the data consuming time of stream of finishing the work;
Randomized blocks, for the dependence between each task that basis is preset, random tasks carrying order is adopted to finish the work stream, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module;
Optimize module, to finish the work stream for choosing on average the shortest tasks carrying consuming time order in training module, the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
5. task scheduling optimization device as claimed in claim 4, it is characterized in that, described optimization module comprises following submodule:
Computing module, for according to described input parameter in conjunction with data consuming time, what calculate corresponding to each tasks carrying order in training module is on average consuming time;
Choosing module, to finish the work stream for choosing on average the shortest tasks carrying consuming time order;
Logging modle, for the tasks carrying order of this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and is recorded to described incidence relation in described training module in newly-increased mode.
6. task scheduling optimization method, is characterized in that, comprises the following steps:
Step 1, receive an input parameter;
The rule that step 2, basis are preset is sorted out described input parameter, to obtain the classification belonging to described input parameter;
Step 3, judge the number of times that the classification belonging to described input parameter records in training module, if described number of times is less than predetermined threshold value, then perform step 4, otherwise, perform step 5; Described training module records the incidence relation of the classification belonging to input parameter, the tasks carrying order of stream of finishing the work and the data consuming time of stream of finishing the work;
Dependence between each task that step 4, basis are preset, random tasks carrying order is adopted to finish the work stream, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module;
Step 5, in training module, choose on average the shortest tasks carrying consuming time order to finish the work stream, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
7. task scheduling optimization method as claimed in claim 6, it is characterized in that, described step 5 comprises following sub-step:
Step 51, classification belonging to described input parameter are in conjunction with data consuming time, and what calculate corresponding to each tasks carrying order in training module is on average consuming time;
Step 52, choose on average the shortest tasks carrying consuming time order and to finish the work stream;
Step 53, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
8. task scheduling optimization method as claimed in claim 6, is characterized in that, described dependence comprise adjustable order relation, can one or more in concurrency relation and alternative relation.
9. task scheduling optimization device, is characterized in that, comprises with lower module:
Receiver module, for receiving an input parameter;
Classifying module, for sorting out described input parameter according to the rule preset, to obtain the classification belonging to described input parameter;
Judge module, the number of times that the classification for judging belonging to described input parameter records in training module, if described number of times is less than predetermined threshold value, then performs randomized blocks, otherwise, perform and optimize module; Described training module records the incidence relation of the classification belonging to input parameter, the tasks carrying order of stream of finishing the work and the data consuming time of stream of finishing the work;
Randomized blocks, for the dependence between each task that basis is preset, random tasks carrying order is adopted to finish the work stream, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module;
Optimize module, to finish the work stream for choosing on average the shortest tasks carrying consuming time order in training module, the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and in newly-increased mode, described incidence relation is recorded in described training module.
10. task scheduling optimization device as claimed in claim 9, it is characterized in that, described optimization module comprises following submodule:
Computing module, for the classification belonging to described input parameter in conjunction with data consuming time, what calculate corresponding to each tasks carrying order in training module is on average consuming time;
Choosing module, to finish the work stream for choosing on average the shortest tasks carrying consuming time order;
Logging modle, for the tasks carrying order of the classification belonging to this input parameter, stream of finishing the work and the data consuming time of stream of finishing the work are formed incidence relation, and is recorded to described incidence relation in described training module in newly-increased mode.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410820409.8A CN104503833A (en) | 2014-12-22 | 2014-12-22 | Task scheduling optimization method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410820409.8A CN104503833A (en) | 2014-12-22 | 2014-12-22 | Task scheduling optimization method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104503833A true CN104503833A (en) | 2015-04-08 |
Family
ID=52945233
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410820409.8A Pending CN104503833A (en) | 2014-12-22 | 2014-12-22 | Task scheduling optimization method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104503833A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106371918A (en) * | 2016-08-23 | 2017-02-01 | 北京云纵信息技术有限公司 | Task cluster scheduling management method and apparatus |
| CN107291533A (en) * | 2016-03-31 | 2017-10-24 | 阿里巴巴集团控股有限公司 | Determine method, the device of upstream node bottleneck degree and system bottleneck degree |
| CN110098918A (en) * | 2019-03-28 | 2019-08-06 | 中至数据集团股份有限公司 | Decrypt dispatching method, device, readable storage medium storing program for executing and computer equipment |
| CN111597028A (en) * | 2020-05-19 | 2020-08-28 | 北京百度网讯科技有限公司 | Method and device for task scheduling |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102222072A (en) * | 2010-04-19 | 2011-10-19 | 腾讯科技（深圳）有限公司 | Method and device for information classification |
| CN103116582A (en) * | 2011-11-16 | 2013-05-22 | 阿里巴巴集团控股有限公司 | Information retrieval method and relevant system and device |
| CN103324534A (en) * | 2012-03-22 | 2013-09-25 | 阿里巴巴集团控股有限公司 | Operation scheduling method and operation scheduler |
-
2014
- 2014-12-22 CN CN201410820409.8A patent/CN104503833A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102222072A (en) * | 2010-04-19 | 2011-10-19 | 腾讯科技（深圳）有限公司 | Method and device for information classification |
| CN103116582A (en) * | 2011-11-16 | 2013-05-22 | 阿里巴巴集团控股有限公司 | Information retrieval method and relevant system and device |
| CN103324534A (en) * | 2012-03-22 | 2013-09-25 | 阿里巴巴集团控股有限公司 | Operation scheduling method and operation scheduler |
Non-Patent Citations (1)
| Title |
|---|
| 许兰等: "一种基于历史信息的自适应动态网格作业调度方法", 《计算机应用与软件》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107291533A (en) * | 2016-03-31 | 2017-10-24 | 阿里巴巴集团控股有限公司 | Determine method, the device of upstream node bottleneck degree and system bottleneck degree |
| CN107291533B (en) * | 2016-03-31 | 2020-10-30 | 阿里巴巴集团控股有限公司 | Method and device for determining upstream node bottleneck degree and system bottleneck degree |
| CN106371918A (en) * | 2016-08-23 | 2017-02-01 | 北京云纵信息技术有限公司 | Task cluster scheduling management method and apparatus |
| CN106371918B (en) * | 2016-08-23 | 2019-10-18 | 北京云纵信息技术有限公司 | Task cluster schedule management method and device |
| CN110098918A (en) * | 2019-03-28 | 2019-08-06 | 中至数据集团股份有限公司 | Decrypt dispatching method, device, readable storage medium storing program for executing and computer equipment |
| CN111597028A (en) * | 2020-05-19 | 2020-08-28 | 北京百度网讯科技有限公司 | Method and device for task scheduling |
| CN111597028B (en) * | 2020-05-19 | 2023-08-25 | 北京百度网讯科技有限公司 | Method and device for task scheduling |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111309479B (en) | Method, device, equipment and medium for realizing task parallel processing | |
| CN109492774B (en) | Deep learning-based cloud resource scheduling method | |
| CN104298550B (en) | A kind of dynamic dispatching method towards Hadoop | |
| Bensmaine et al. | A new heuristic for integrated process planning and scheduling in reconfigurable manufacturing systems | |
| TWI547817B (en) | Method, system and apparatus of planning resources for cluster computing architecture | |
| CN105260237B (en) | The task scheduling system and its dispatching method of heterogeneous polynuclear platform | |
| WO2015066979A1 (en) | Machine learning method for mapreduce task resource configuration parameters | |
| CN104503833A (en) | Task scheduling optimization method and device | |
| CN110389816A (en) | Method, apparatus and computer program product for scheduling of resource | |
| CN108628672A (en) | Method for scheduling task, system, terminal device and storage medium | |
| CN108897798A (en) | Method and device for classifying customer service work orders and electronic equipment | |
| CN113467944B (en) | Resource deployment device and method for complex software system | |
| CN111949795A (en) | Work order automatic classification method and device | |
| CN106776025A (en) | A kind of computer cluster job scheduling method and its device | |
| CN104318459A (en) | Automatic market analysis method based on multi-mode learning | |
| CN111198754A (en) | Task scheduling method and device | |
| CN116820714A (en) | Scheduling method, device, equipment and storage medium of computing equipment | |
| CN108268312A (en) | Method for scheduling task and scheduler | |
| CN101495978B (en) | Reduction of message flow between bus-connected consumers and producers | |
| CN110232486B (en) | Multi-workshop comprehensive scheduling method based on K shortest path | |
| Kavitha et al. | Ai Integration in Data Driven Decision Making for Resource Management in Internet of Things (Iot): A Survey | |
| CN110019207A (en) | Data processing method and device and script display methods and device | |
| CN106295892A (en) | The computational methods of a kind of enterprise commission and device | |
| Fernandes et al. | Worker assignment in dual resource constrained systems subject to machine failures: a simulation study | |
| CN106776974A (en) | A kind of big data monitoring method based on configuration, device and platform |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20180416 Address after: 510000 Guangzhou City, Guangzhou, Guangdong, Fangcun Avenue, one of the 314 self compiled Applicant after: Guangzhou Pinwei Software Co., Ltd. Address before: 510370, No. 314, Fangcun Road, Liwan District, Guangdong, Guangzhou two Applicant before: GUANGZHOU VIP NETWORK TECHNOLOGY CO., LTD. |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150408 |