CN108628667A - A kind of heuristic dynamic task scheduling system and its implementation based on multiple attribute decision making (MADM) - Google Patents
A kind of heuristic dynamic task scheduling system and its implementation based on multiple attribute decision making (MADM) Download PDFInfo
- Publication number
- CN108628667A CN108628667A CN201710157889.8A CN201710157889A CN108628667A CN 108628667 A CN108628667 A CN 108628667A CN 201710157889 A CN201710157889 A CN 201710157889A CN 108628667 A CN108628667 A CN 108628667A
- Authority
- CN
- China
- Prior art keywords
- task
- module
- madm
- decision making
- attribute decision
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/48—Program initiating; Program switching, e.g. by interrupt
- G06F9/4806—Task transfer initiation or dispatching
- G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
- G06F9/4881—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
- G06F9/4887—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues involving deadlines, e.g. rate based, periodic
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
- G06F9/505—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals considering the load
Abstract
The present invention provides a kind of heuristic dynamic task scheduling system and its implementation based on multiple attribute decision making (MADM),Including global task queue module,Scheduler module and local task queue module,The global task queue module includes multiple scheduler task set,Global task queue module downlink is connect with Scheduler module,Scheduler module includes global task choosing module and process cores selecting module,Scheduler module downlink is equipped with local task queue module,Local task queue module includes multiple and is parallel relationship between each other,The present invention comprehensively utilizes heuristic search thought and multiple attribute decision making (MADM) thought,In global task choosing,Based on heuristic search,Ensure to find the task of a minimum target cost in linear time complexity,When being scheduled for task choosing process cores,Based on multiple attribute decision making (MADM),Consider the information about dynamic parameters of currently processed core,It is scheduled for the highest process cores of task choosing comprehensive evaluation value.
Description
Technical field
The present invention relates to task scheduling technique field more particularly to a kind of heuristic dynamic tasks based on multiple attribute decision making (MADM)
Scheduling system and its implementation.
Background technology
With the fast development of hardware technology, multi-core processor is applied to daily life, scientific algorithm etc. more and more
In field.The performance of multi-core processor has had both the advantage of low-power consumption considerably beyond traditional single core processor.It plays
Go out advantage of the multi-core processor on hardware, needs corresponding software technology support.Wherein, task scheduling is influenced at multinuclear
As soon as reason device plays the key point of its performance, the performance that give full play of each kernel of multi-core processor must have efficiently
Multinuclear task scheduling algorithm.
Multinuclear task scheduling has proved to be the problem of a NP hardly possible at present, mostly uses heuristic or heredity greatly now
Method is solved.In the existing method, such as document《A kind of dynamic dispatching algorithm of new Real-Time Multiprocessor Systems》In
Method have higher task scheduling success rate, but select process cores execute task when, have ignored the utilization rate of process cores.
When task is intensive, the free time of part core can be caused long, the utilization rate between core is uneven.For another example document《Multiprocessor
System dynamic dispatching load balancing saving algrithm》In method, the method adds pair in the target function value of calculating task
The weight of task processing time considers, by the calculating for changing object function so that the short task of run time can preferentially be located
Reason avoids the task of long operational time from occupying process cores for a long time, causes short task scheduling unsuccessful.But working as has multiple process cores
When may be selected, using the method randomly selected, there is this selection strategy certain blindness may be led when number of tasks is excessive
The load of process cores is caused to aggravate instead.
Invention content
The present invention provides a kind of heuristic dynamic task scheduling system and its implementation based on multiple attribute decision making (MADM), it is comprehensive
Conjunction based on heuristic search, is ensured using heuristic search thought and multiple attribute decision making (MADM) thought in global task choosing
Task of a minimum target cost is found in linear time complexity, when being scheduled for task choosing process cores, with more
Based on attribute decision, the information about dynamic parameters of currently processed core is considered, be the highest processing of task choosing comprehensive evaluation value
Core is scheduled.
In order to solve the above technical problems, the embodiment of the present application provides a kind of heuristic dynamic based on multiple attribute decision making (MADM) times
Business scheduling system, including global task queue module, Scheduler module and local task queue module, the global task team
Row module includes multiple scheduler task set, and global task queue module downlink is connect with Scheduler module, the scheduler
Module includes global task choosing module and process cores selecting module, and the Scheduler module downlink is equipped with local task queue
Module, the local task queue module include multiple and are parallel relationship between each other.
As the preferred embodiment of this programme, the implementation method of the task scheduling system includes the following steps：
Step 1：If the set of tasks currently reached in scheduling queue is, by task-set T
It is arranged according to off period non-decreasing sequence, assigns the off period the small higher dispatching priority of task.
Step 2：K task in window Wnd, which carries out feasibility inspection, to be checked to feasibility, with the current local scheduling of determination
Whether meet " strong feasible ".Feasible condition is by force：To any one in K task, a process cores can be found so that
Task is completed within the off period.The value formula of K is as follows.
Wherein symbolRepresent the minimum value for taking the two.
Step 3：If meeting " strong feasible ", step 4 is arrived, step 6 is otherwise arrived.
Step 4：Calculate the value of the object function H (i) of task in feasibility window.The calculation formula of H (i) is as follows.
Wherein,Represent taskDeadline,For taskIt can start the time executed earliest.
Step 5：The task of selection H (i) value minimum is scheduled, and carries out multiple attribute decision making (MADM), and optimal processing core is selected to execute
Task arrives step 8.
Step 6：It traces back to last layer to be scheduled, backtracking number adds one.
Step 7：The small task of H (i) values time is selected in this layer choosing to be scheduled, and is carried out multiple attribute decision making (MADM), is selected optimal place
It manages core and executes task.
Step 8：Feasibility inspection window is moved backward into a task.
Step 9：Step 2- steps 8 are repeated, until there is any one following situation：Task in T is all adjusted
Spend the possibility for either reaching maximum traceback number BackMax or not recalling again.
One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages：
Heuristic search thought and multiple attribute decision making (MADM) thought are comprehensively utilized, in global task choosing, based on heuristic search,
The ensureing to find a minimum target cost in linear time complexity of the task, is scheduled for task choosing process cores
When, based on multiple attribute decision making (MADM), consider the information about dynamic parameters of currently processed core, is task choosing comprehensive evaluation value highest
Process cores be scheduled.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention, for those of ordinary skill in the art, without having to pay creative labor, also
It can be obtain other attached drawings according to these attached drawings.
Fig. 1 is the system model figure of this method work；
Fig. 2 is the model framework figure of this method；
In Fig. 1-Fig. 2,1, global task queue module, 2, Scheduler module, 3, local task queue module, 4, global task choosing
Select module, 5, process cores selecting module.
Specific implementation mode
The present invention provides a kind of heuristic dynamic task scheduling system and its implementation based on multiple attribute decision making (MADM), it is comprehensive
Conjunction based on heuristic search, is ensured using heuristic search thought and multiple attribute decision making (MADM) thought in global task choosing
Task of a minimum target cost is found in linear time complexity, when being scheduled for task choosing process cores, with more
Based on attribute decision, the information about dynamic parameters of currently processed core is considered, be the highest processing of task choosing comprehensive evaluation value
Core is scheduled.
In order to better understand the above technical scheme, in conjunction with appended figures and specific embodiments to upper
Technical solution is stated to be described in detail.
As Figure 1-Figure 2, a kind of heuristic dynamic task scheduling system based on multiple attribute decision making (MADM), including global task
Queue module 1, Scheduler module 2 and local task queue module 3, the global task queue module 1 include multiple scheduling
Set of tasks, global 1 downlink of task queue module are connect with Scheduler module 2, and the Scheduler module 2 includes global task
Selecting module 4 and process cores selecting module 5,2 downlink of Scheduler module is equipped with local task queue module 3, described
Local task queue module 3 includes multiple and is parallel relationship between each other.
Wherein, in practical applications, the implementation method of the task scheduling system includes the following steps：
Step 1：If the set of tasks currently reached in scheduling queue is, by task-set T
It is arranged according to off period non-decreasing sequence, assigns the off period the small higher dispatching priority of task.
Step 2：K task in window Wnd, which carries out feasibility inspection, to be checked to feasibility, with the current local scheduling of determination
Whether meet " strong feasible ".Feasible condition is by force：To any one in K task, a process cores can be found so that
Task is completed within the off period.The value formula of K is as follows.
Wherein symbolRepresent the minimum value for taking the two.
Step 3：If meeting " strong feasible ", step 4 is arrived, step 6 is otherwise arrived.
Step 4：Calculate the value of the object function H (i) of task in feasibility window.The calculation formula of H (i) is as follows.
Wherein,Represent taskDeadline,For taskIt can start the time executed earliest.
Step 5：The task of selection H (i) value minimum is scheduled, and carries out multiple attribute decision making (MADM), and optimal processing core is selected to execute
Task arrives step 8.
Step 6：It traces back to last layer to be scheduled, backtracking number adds one.
Step 7：The small task of H (i) values time is selected in this layer choosing to be scheduled, and is carried out multiple attribute decision making (MADM), is selected optimal place
It manages core and executes task.
Step 8：Feasibility inspection window is moved backward into a task.
Step 9：Step 2- steps 8 are repeated, until there is any one following situation：Task in T is all adjusted
Spend the possibility for either reaching maximum traceback number BackMax or not recalling again.
The above described is only a preferred embodiment of the present invention, be not intended to limit the present invention in any form, though
So the present invention has been disclosed as a preferred embodiment, and however, it is not intended to limit the invention, any technology people for being familiar with this profession
Member, without departing from the scope of the present invention, when the technology contents using the disclosure above make a little change or modification
For the equivalent embodiment of equivalent variations, as long as being the content without departing from technical solution of the present invention, according to the technical essence of the invention
To any simple modification, equivalent change and modification made by above example, in the range of still falling within technical solution of the present invention.
Claims (2)
1. a kind of heuristic dynamic task scheduling system based on multiple attribute decision making (MADM), including global task queue module（1）, scheduling
Device module（2）With local task queue module（3）, which is characterized in that the global task queue module（1）Including multiple tune
Spend set of tasks, global task queue module（1）Downlink and Scheduler module（2）Connection, the Scheduler module（2）Including
Global task choosing module（4）With process cores selecting module（5）, the Scheduler module（2）Downlink is equipped with local task team
Row module（3）, the local task queue module（3）Including it is multiple and between each other be parallel relationship.
2. a kind of heuristic dynamic task scheduling system based on multiple attribute decision making (MADM) according to claim 1, feature exist
In the implementation method of the task scheduling system includes the following steps：
Step 1：If the set of tasks currently reached in scheduling queue is, task-set T is pressed
It is arranged according to off period non-decreasing sequence, assigns the off period the small higher dispatching priority of task；
Step 2：K task progress feasibility inspection in window Wnd is checked to feasibility, with the current local scheduling of determination whether
Meeting " strong feasible " strong feasible condition is：To any one in K task, a process cores can be found so that task
It is completed within the off period, the value formula of K is as follows；
Wherein symbolRepresent the minimum value for taking the two；
Step 3：If meeting " strong feasible ", step 4 is arrived, step 6 is otherwise arrived；
Step 4：The value of the object function H (i) of task in feasibility window is calculated, the calculation formula of H (i) is as follows：
Wherein,Represent taskDeadline,For taskIt can start the time executed earliest；
Step 5：The task of selection H (i) value minimum is scheduled, and carries out multiple attribute decision making (MADM), and optimal processing core is selected to execute task,
To step 8；
Step 6：It traces back to last layer to be scheduled, backtracking number adds one；
Step 7：The small task of H (i) values time is selected in this layer choosing to be scheduled, and is carried out multiple attribute decision making (MADM), is selected optimal processing core
Execution task；
Step 8：Feasibility inspection window is moved backward into a task；
Step 9：Step 2- steps 8 are repeated, until there is any one following situation：Task in T all dispatched or
The possibility that person reaches maximum traceback number BackMax or do not recall again.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710157889.8A CN108628667A (en) | 2017-03-16 | 2017-03-16 | A kind of heuristic dynamic task scheduling system and its implementation based on multiple attribute decision making (MADM) |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710157889.8A CN108628667A (en) | 2017-03-16 | 2017-03-16 | A kind of heuristic dynamic task scheduling system and its implementation based on multiple attribute decision making (MADM) |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108628667A true CN108628667A (en) | 2018-10-09 |
Family
ID=63686508
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710157889.8A Pending CN108628667A (en) | 2017-03-16 | 2017-03-16 | A kind of heuristic dynamic task scheduling system and its implementation based on multiple attribute decision making (MADM) |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108628667A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110825510A (en) * | 2019-11-05 | 2020-02-21 | 中国人民解放军国防科技大学 | Task-driven multi-satellite cooperative task allocation method and system |
| CN110928648A (en) * | 2019-12-10 | 2020-03-27 | 浙江工商大学 | Heuristic and intelligent computing-fused cloud workflow segmentation online scheduling optimization method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104484233A (en) * | 2014-10-31 | 2015-04-01 | 北京邮电大学 | Method of allocating resources |
-
2017
- 2017-03-16 CN CN201710157889.8A patent/CN108628667A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104484233A (en) * | 2014-10-31 | 2015-04-01 | 北京邮电大学 | Method of allocating resources |
Non-Patent Citations (1)
| Title |
|---|
| 乔颖等: ""一种新的实时多处理器系统的动态调度算法"", 《软件学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110825510A (en) * | 2019-11-05 | 2020-02-21 | 中国人民解放军国防科技大学 | Task-driven multi-satellite cooperative task allocation method and system |
| CN110928648A (en) * | 2019-12-10 | 2020-03-27 | 浙江工商大学 | Heuristic and intelligent computing-fused cloud workflow segmentation online scheduling optimization method |
| CN110928648B (en) * | 2019-12-10 | 2022-05-20 | 浙江工商大学 | Heuristic and intelligent computing-fused cloud workflow segmentation online scheduling optimization method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | Dynamic cloud task scheduling based on a two-stage strategy | |
| TWI503742B (en) | Multiprocessors systems and processes scheduling methods thereof | |
| WO2015106533A1 (en) | Coprocessor-based job scheduling processing method and device | |
| CN112000388B (en) | Concurrent task scheduling method and device based on multi-edge cluster cooperation | |
| CN103092683A (en) | Scheduling used for analyzing data and based on elicitation method | |
| CN104572305A (en) | Load-balanced cluster rendering task dispatching method | |
| US9104491B2 (en) | Batch scheduler management of speculative and non-speculative tasks based on conditions of tasks and compute resources | |
| CN105955809B (en) | Thread scheduling method and system | |
| CN112363821A (en) | Computing resource scheduling method and device and computer equipment | |
| CN111104211A (en) | Task dependency based computation offload method, system, device and medium | |
| US20180329750A1 (en) | Resource management method and system, and computer storage medium | |
| CN104268018A (en) | Job scheduling method in Hadoop cluster and job scheduler | |
| US20170083382A1 (en) | Power-aware work stealing | |
| US20170097854A1 (en) | Task placement for related tasks in a cluster based multi-core system | |
| CN103365711A (en) | Task scheduling mechanism and method applied to business platform of internet of Things | |
| Tang et al. | Nanily: A qos-aware scheduling for dnn inference workload in clouds | |
| CN107533479A (en) | Power knows scheduling and power manager | |
| CN108628667A (en) | A kind of heuristic dynamic task scheduling system and its implementation based on multiple attribute decision making (MADM) | |
| CN108984286A (en) | A kind of resource regulating method and system of cloud computing platform | |
| Tsai et al. | Scheduling multiple scientific and engineering workflows through task clustering and best-fit allocation | |
| CN114371933A (en) | Method and system for dynamically scheduling multi-core fusion computing processor | |
| US11954518B2 (en) | User-defined metered priority queues | |
| Singh et al. | Market-inspired dynamic resource allocation in many-core high performance computing systems | |
| CN111367655A (en) | Method, system and storage medium for GPU resource scheduling in cloud computing environment | |
| CN105955816A (en) | Event scheduling method and device |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181009 |