CN101582831A - Scheduling method of grid resources of multi-Qos - Google Patents
Scheduling method of grid resources of multi-Qos Download PDFInfo
- Publication number
- CN101582831A CN101582831A CNA2008100375647A CN200810037564A CN101582831A CN 101582831 A CN101582831 A CN 101582831A CN A2008100375647 A CNA2008100375647 A CN A2008100375647A CN 200810037564 A CN200810037564 A CN 200810037564A CN 101582831 A CN101582831 A CN 101582831A
- Authority
- CN
- China
- Prior art keywords
- resource
- time
- scheduling
- expense
- value
- 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
Aiming at the problem of scheduling resources in the grid computing, the invention provides a scheduling method of grid resources of multi-Qos, which selects the most economic resource to schedule when in each scheduling. Compared with the prior art, the invention can complete the task at the least scheduling driving function value while meeting the user QoS, and realizes the satisfaction of a user to the max.
Description
Technical field
The present invention relates to the scheduling of resource technology in the grid computing, particularly relate to the grid resource scheduling method of a kind of many QoS.
Background technology
Grid computing is accompanied by Internet technology and develops rapidly, the special novel computation schema that calculates at complicated science, whole Network integration is become a huge supercomputer, realize the overall sharing of computational resource, storage resources, data resource, information resources, knowledge resource, Expert Resources.The grid computing resource becomes solution large scale industry science trends calculated in modern society.Yet the dispersiveness of grid computing resource, the cost that decision is calculated based on the distribution of cross-node communication is higher than the Distributed Calculation in the local area network (LAN) far away, and frequent communication causes sizable time-delay expense.Improving concurrency is the important means that reduces communication overhead.Because the execution speed of the Internet resources that disperse on the geographical position is different with expense, also to consider the demand of user during scheduling to aspects such as time of scheduler task and expenses.
Since the dispersiveness of gridding resource, the difference of resource kind, resource is had by different individuals and organizations, and the difference of the operation expense of resource, difference of task character or the like factor, scheduling of resource are very complicated processes.The scheduling strategy that most of existing grid resources and dispatching patcher are taked is only with the throughput and the utilance that improve system and to finish the work in earliest time be target, resource access cost and user are not taken into account the requirement of time, expense, do not considered user's demand well.User's demand differs greatly under the background of different application.Because the order of importance and emergency difference of task, some customer requirements is no more than in expense under the prerequisite of Budget, and task is more early finished good more.Some requires under not super Deadline prerequisite running time, and it is few more good more to spend.Both take into account the customer requirements that also has.
Summary of the invention
Technical problem to be solved by this invention is exactly the grid resource scheduling method that a kind of many QoS are provided for the defective that overcomes above-mentioned prior art existence.
Purpose of the present invention can be achieved through the following technical solutions: the grid resource scheduling method of a kind of many QoS, it is characterized in that, and may further comprise the steps:
(1) the different demand parameter of the different task of record grid user submission;
When (2) being each task scheduling gridding resource,, calculate the user time-expense scaled value of each resource, determine the resource of user time-expense scaled value minimum according to this task corresponding demand parameter;
(3) above-mentioned least resource and user's demand parameter compare, and judge whether this resource satisfies user's demand, if, execution in step (4), if not, then scheduling failure;
(4) give the user with this resource allocation.
Described demand parameter comprises time restriction value, overhead constraints value, time weighting value, expense weighted value.
The user time of described each resource of calculating-expense scaled value passes through function:
F(V
i，C
i，L
j)＝(L
j/V
i)×(α+β×C
i)
G(V
i，C
i，T
i，L
j)＝F(V
i，C
i，L
j)+α×T
i+β×C
total
Finish;
Wherein:
1＜＝i＜n；
R
iBe system resource, the line number of its per unit time operation and the expense of per second run time version are used V respectively
iAnd C
iExpression; A
jBe task, L
jBe A
jThe task line number, task A
jBe dispatched to resource R
iCarry out, then time overhead=L
j/ V
i, run time version cost=(L
j/ V
i) * C
iα is the time weighting value, and β is the expense weighted value, alpha+beta=1; F (V
i, C
i, L
j) be task A
jBe dispatched to resource R
iThe cost function of last execution;
T
iBe resource R
iThe total time of executing the task; C
TotalBe the overhead that task is carried out, G (V
i, C
i, T
i, L
j) be the user time-expense scaled value of each resource.
The user time of described each resource-expense scaled value is dispatched F (V when scheduling is initial
i, C
i, L
j) the minimum gridding resource R of value
k, be L the running time of revising this resource after this finishing scheduling
1/ V
k
The user time of described each resource-expense scaled value is dispatched G (V when carrying out the j time scheduling
i, C
i, T
i, L
j) the minimum gridding resource R of value
p, be T the running time of revising this resource after this finishing scheduling
p+ L
j/ V
p
The user time of described each resource-expense scaled value is dispatched G (V when carrying out the n time scheduling
i, C
i, T
i, L
n) the minimum gridding resource of value.
Compared with prior art, the present invention can finish the work with the scheduling resource of minimum when satisfying user QoS, realizes user's satisfaction substantially.
Description of drawings
Fig. 1 is a flow chart of the present invention;
Fig. 2 is the running time and the expense schematic diagram of embodiments of the invention.
Embodiment
The invention will be further described below in conjunction with accompanying drawing.
Shown in Fig. 1～2, the grid resource scheduling method of a kind of many QoS may further comprise the steps:
(1) the different demand parameter of the different task of record grid user submission;
When (2) being each task scheduling gridding resource,, calculate the user time-expense scaled value of each resource, determine the resource of user time-expense scaled value minimum according to this task corresponding demand parameter;
(3) above-mentioned least resource and user's demand parameter compare, and judge whether this resource satisfies user's demand, if, execution in step (4), if not, then scheduling failure;
(4) give the user with this resource allocation.
Described demand parameter comprises time restriction value, overhead constraints value, time weighting value, expense weighted value.
The user time of described each resource of calculating-expense scaled value passes through function:
F(V
i，C
i，L
j)＝(L
j/V
i)×(α+β×C
i)
G(V
i，C
i，T
i，L
j)＝F(V
i，C
i，L
j)+α×T
i+β×C
total
Finish;
Wherein:
1＜＝i＜n；
R
iBe system resource, the line number of its per unit time operation and the expense of per second run time version are used V respectively
iAnd C
iExpression; A
jBe task, L
jBe A
jThe task line number, task A
jBe dispatched to resource R
iCarry out, then time overhead=L
j/ V
i, run time version cost=(L
j/ V
i) * C
iα is the time weighting value, and β is the expense weighted value, alpha+beta=1; F (V
i, C
i, L
j) be task A
jBe dispatched to resource R
iThe cost function of last execution;
T
iBe resource R
iThe total time of executing the task; C
TotalBe the overhead that task is carried out, G (V
i, C
i, T
i, L
j) be the user time-expense scaled value of each resource;
The user time of described each resource-expense scaled value is dispatched F (V when scheduling is initial
i, C
i, L
j) the minimum gridding resource R of value
k, be L the running time of revising this resource after this finishing scheduling
1/ V
k
The user time of described each resource-expense scaled value is dispatched G (V when carrying out the j time scheduling
i, C
i, T
i, L
j) the minimum gridding resource R of value
p, be T the running time of revising this resource after this finishing scheduling
p+ L
j/ V
p
The user time of described each resource-expense scaled value is dispatched G (V when carrying out the n time scheduling
i, C
i, T
i, L
n) the minimum gridding resource of value.
The present invention is based on greedy algorithm, make it through a preliminary treatment after, be greedy criterion with the schedule driven function, realize that the user dispatches the most economical scheduling strategy under these mission requirements.
Greedy algorithm (greedy algorithm) promptly adopts the method for progressively constructing optimal solution.In each stage, all make one and look optimum decision-making (under certain standard).In a single day decision-making is made, just can not change again.The foundation of making greedy decision-making is called greedy criterion (greedy criterion).
The scheduling cost function
Be provided with when the user submits task to and finish this task receptible time restriction of institute and overhead constraints, represent with parameter Deadline and Budget respectively.System resource R
iThe per unit time line number (is unit with MI) and the expense of per second run time version of operation used V parameter respectively
iAnd C
iExpression.Suppose task A
jLine number be L
j, task A
jBe dispatched to resource R
iCarry out, then time overhead=L
j/ V
i, run time version cost=(L
j/ V
i) * C
iTo different application programs, the user is different to the requirement of time and expense:
(1) some customer requirements is no more than in expense under the prerequisite of Budget, and task is more early finished good more.
(2) some requires under not super Deadline prerequisite running time, and it is few more good more to spend.
(3) both take into account the customer requirements that also has.
So for the better demand that must satisfy the user, two parameter alpha must be set, β composes to the weight of finishing required by task time and cost as the user, and satisfy alpha+beta=1.For example, to above-mentioned (1): α=1, β=0; To above-mentioned (2): α=0, β=1; To above-mentioned (3): 0＜α, β＜1 and satisfy alpha+beta=1.
Overhead=(L that above-mentioned cum rights is heavy
j/ V
i) * (alpha+beta * C
i).Task A
jBe dispatched to resource R
iOverhead F (the V that carries out
i, C
i, L
j)=(L
j/ V
i) * (alpha+beta * C
i).We claim F (V
i, C
i, L
j) be task A
jBe dispatched to resource R
iThe cost function of last execution.
The schedule driven function
Suppose that a certain moment grid user submitted n task to system, existing task obtained task list A according to the length sort descending:
A：A
1A
2......A
n
L：L
1L
2......L
n
M list of available resources arranged in the system:
R：R
1R
2......R
m
V：V
1V
2......V
m
C：C
1C
2......C
m
Resource R
iThe total time of executing the task is used T
iExpression, the overhead C that task is carried out
TotalExpression.Make G (V
i, C
i, T
i, L
j)=F (V
i, C
i, L
j)+α * T
i+ β * C
Total, 1＜=i＜n wherein.Then as task A
jBe dispatched on the resource carry out after, G (V
i, C
i, T
i, L
j)=F (V
i, C
i, L
j)+α * T
i+ β * C
Total, F (V wherein
i, C
i, L
j) be the j time scheduling overhead, α * T
iAmount to expense, then G (V for the total run time of j-1 scheduling back resource i
i, C
i, T
i, L
j) be the expense of amounting to of the total run time of resource i after preceding j time the scheduling.
Grid resource scheduling
During each selection scheduling, search G (V
i, C
i, T
i, L
j) the minimum resource of value is dispatched, and the running time T of modification respective resources
i
During the submission task:
T when initial
i=0,1＜=i＜=n wherein, G (V
i, C
i, T
i, L
1)=F (V
i, C
i, L
1)+α * T
i+ β * C
Total=F (V
i, C
i, L
1), function G (V
i, C
i, T
i, L
1) the value minimum, i.e. F (V
i, C
i, L
1) the value minimum.A
1F (V is given in scheduling
i, C
i, L
1) the minimum resource execution of value.Relatively whether Tiao Du time and cost have exceeded Deadline and Budget that the user can bear.If do not exceed, then distribute; Otherwise scheduling failure.If successfully scheduling supposes that this resource is R
k, then after this finishing scheduling, T
k=L
1/ V
k
Carrying out the j time when scheduling 1＜j＜n wherein, G (V
i, C
i, T
i,, L
j)=F (V
i, C
i, L
j)+α * T
i+ β * C
Total, A
jG (V is given in scheduling
i, C
i, T
i, L
j) the minimum resource execution of value.Relatively whether Tiao Du time and cost have exceeded Deadline and Budget that the user can bear.If do not exceed, then distribute; Otherwise scheduling failure.If successfully scheduling supposes that this resource is R
p, then after this finishing scheduling, T
p=T
p+ L
j/ V
p
When carrying out the n time scheduling, G (V
i, C
i, T
i, L
n)=F (V
i, C
i, L
n)+α * T
i+ β * C
Total, A
nG (V is given in scheduling
i, C
i, T
i, L
n) the minimum resource execution of value.Relatively whether Tiao Du time and cost have exceeded Deadline and Budget that the user can bear.If do not exceed, then distribute; Otherwise scheduling failure.
Superiority proves
Adopt this method to dispatch the schedule driven functional value minimum that obtains than any other method, promptly the user time-expense scaled value minimum, satisfy the requirement of user best to time and expense.
The proof hypothesis adopts the D that is scheduling to of this method, adopts the B that is scheduling to of additive method, and n task hundred classified as from big to small by length: L
1, L
2... L
nM resource arranged by the speed of service: R
1, R
2... R
mT
DAnd C
D TotalFor operation total time and the overhead under the scheduling D, be 0 when initial; T
BAnd C
B TotalFor operation total time and the overhead under the scheduling B, be 0 when initial.
Now adopt second mathematical induction to prove to separate D to separate than more excellent one of B:
When carrying out dispatching the first time, according to this dispatching algorithm, the resource of a G functional value minimum is chosen in the each scheduling of D, so G (V
D i, C
D i, T
D i, L
1)＜=G (V
B j, C
B jT
B j, L
1), i.e. α * T
D i+ β * C
D Total＜=α * T
B i+ β * C
B Total, carry out scheduling back D for the first time and be separating than more excellent one of B;
Suppose that after finishing the k-1 time scheduling, D separates than more excellent one of B, promptly satisfies MAX (α * T
D i+ β * C
D Total)＜=MAX (α * T
B i+ β * C
B Total), 1＜=i＜=k-1 wherein, 1＜=j＜=k-1;
(3) then carry out the k time scheduling, if after having dispatched for the k time
MAX (α * T
D'
i+ β * C
D'
Total)＞MAX (α * T
B'
i+ β * C
B'
Total), 1＜=i＜=k wherein, 1＜=j＜=k;
Now with the method proof is impossible anyway: if after the k time scheduling, D is A
kResource R is given in scheduling
D pCarry out, B is A
kResource R is given in scheduling
B qCarry out, by
As can be known:
MAX(α×T
D’
i+β×C
D’
total)＝G(V
D p，C
D p，T
D p，L
k)
＞MAX(α×T
B’
j+β×C
B’
total)
＞＝G(V
B q，C
B q，T
B q，L
k)，
Be G (V
D p, C
D p, T
D p, L
k)＞G (V
B q, C
B q, T
B q, L
k)
If D selects other resource R
D iDispatch, then G (V
D i, C
D i, T
D i, L
k)＞=G (V
D p, C
D p, T
D p, L
k)＞G (V
B q, C
B q, T
B q, L
1), because D selects R
D pScheduling is the optimal solution under the D scheme, and other separate all big than its expense, i.e. G (V
B q, C
B q, T
B q, L
k)＜MIN (G (V
D i, C
D i, T
D i, L
k)).
And by MAX (α * T
D i+ β * C
D Total)＜=MAX (α * T
B i+ β * C
B Total), 1＜=i＜=k-1 wherein, 1＜=j＜=k-1, relative B of the time of implementation of different resource is average in the scheduling scheme of D as can be known, and the concurrency of D is relatively good.And being scheduled of task is scheduled successively according to the order that code length successively decreases, so if exist q to make G (V
B q, C
B q, T
B q, L
k)＜MIN (G (V
D i, C
D i, T
D i, L
k)) must exist j to make MAX (α * T
B'
i+ β * C
B'
Total)-G (V
D j, C
D j, T
D j, L
K-1)＞=F (V
j, C
j, L
k) set up.A
kResource R is given in scheduling
D jCarry out, have after scheduling is finished:
MAX(α×T
D’
i+β×C
D’
total)
＜＝G(V
D j，C
D j，T
D j，L
k-1)+F(V
j，C
j，L
k)
＜＝MAX(α×T
B’
i+β×C
B’
total)，
This and (2) contradiction.
To sum up, got by mathematical induction, the dispatching method D that the greedy algorithm that utilization is promoted obtains is better than being different from arbitrarily the feasible schedule scheme of A, has realized the optimization of the scheduling of resource under the agreed terms.
Embodiment
Suppose that a grid user has 8 tasks to carry out, length is respectively 8000,7000,6000,5000,4000,3000,2000,1000, and unit is MI, and Deadline and Budget that the user is provided with are 1000, and system has 3 available resources R at this moment
1, R
2, R
3, execution speed is respectively 300MIPS, 500MIPS, and 800MIPS, expense is respectively 3.0,5.0,8.0 (cents/sec of unit).
As time weight timeWeight=1, during expense weight costWeight=0, scheduling process is as shown in the table:
| Task ID | Task length | Begin to carry out constantly | Execute constantly | Expense (cents) | Available resources |
| 0 | 8000 | 19.32 | 77.32 | 80.0 | Resourse_0 |
| 1 | 7000 | 24.76 | 81.51 | 43.75 | Resourse_1 |
| 2 | 6000 | 30.2 | 85.7 | 22.5 | Resourse_2 |
| 3 | 5000 | 77.32 | 131.57 | 50.0 | Resourse_0 |
| 4 | 4000 | 81.51 | 134.51 | 25.0 | Resourse_1 |
| 5 | 3000 | 131.57 | 183.32 | 30.0 | Resourse_0 |
| 6 | 2000 | 85.7 | 136.2 | 7.5 | Resourse_2 |
| 7 | 1000 | 134.51 | 183.76 | 6.25 | Resourse_1 |
8 tasks are assigned to resource R respectively
0, R
1, R
2, R
0, R
1, R
0, R
2, R
1Last execution, total run time are 183.76sec, and overhead is 256cents, and the schedule driven functional value is 183.76.
TimeWeight is made as 0.1,0.2,0.3,0.4,0.5,0.6 successively, 0.7 0.8,0.9,1 o'clock, running time and expense were as shown in Figure 2, as we know from the figure, timeWeight becomes greatly gradually, and total run time has the trend that diminishes, and costWeight diminishes gradually, and overhead has the trend that becomes big.This algorithm can be made different scheduling strategies to time and the different of expense requirement according to the user, and this scheduling strategy can be met customer need preferably.
Claims (6)
1. the grid resource scheduling method of QoS more than a kind is characterized in that, may further comprise the steps:
(1) the different demand parameter of the different task of record grid user submission;
When (2) being each task scheduling gridding resource,, calculate the user time-expense scaled value of each resource, determine the resource of user time-expense scaled value minimum according to this task corresponding demand parameter;
(3) above-mentioned least resource and user's demand parameter compare, and judge whether this resource satisfies user's demand, if, execution in step (4), if not, then scheduling failure;
(4) give the user with this resource allocation.
2. the grid resource scheduling method of a kind of many QoS according to claim 1 is characterized in that, described demand parameter comprises time restriction value, overhead constraints value, time weighting value, expense weighted value.
3. the grid resource scheduling method of a kind of many QoS according to claim 2 is characterized in that, the user time of described each resource of calculating-expense scaled value passes through function:
F(V
i，C
i，L
j)＝(L
j/V
i)×(α+β×C
i)
G(V
i，C
i，T
i，L
j)＝F(V
i，C
i，L
j)+α×T
i+β×C
total
Finish;
Wherein:
1＜＝i＜n；
R
iBe system resource, the line number of its per unit time operation and the expense of per second run time version are used V respectively
iAnd C
iExpression; A
jBe task, L
jBe A
jThe task line number, task A
jBe dispatched to resource R
iCarry out, then time overhead=L
j/ V
i, run time version cost=(L
j/ V
i) * C
iα is the time weighting value, and β is the expense weighted value, alpha+beta=1; F (V
i, C
i, L
j) be task A
jBe dispatched to resource R
iThe cost function of last execution;
T
iBe resource R
iThe total time of executing the task; C
TotalBe the overhead that task is carried out, G (V
i, C
i, T
i, L
j) be the user time-expense scaled value of each resource.
4. the grid resource scheduling method of a kind of many QoS according to claim 3 is characterized in that, the user time of described each resource-expense scaled value is dispatched F (V when scheduling is initial
i, C
i, L
j) the minimum gridding resource R of value
k, be L the running time of revising this resource after this finishing scheduling
l/ V
k
5. the grid resource scheduling method of a kind of many QoS according to claim 4 is characterized in that, the user time of described each resource-expense scaled value is dispatched G (V when carrying out the j time scheduling
i, C
i, T
i, L
j) the minimum gridding resource R of value
p, be T the running time of revising this resource after this finishing scheduling
p+ L
j/ V
p
6. the grid resource scheduling method of a kind of many QoS according to claim 5 is characterized in that, the user time of described each resource-expense scaled value is dispatched G (V when carrying out the n time scheduling
i, C
i, T
i, L
n) the minimum gridding resource of value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008100375647A CN101582831B (en) | 2008-05-16 | 2008-05-16 | Scheduling method of grid resources of multi-Qos |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2008100375647A CN101582831B (en) | 2008-05-16 | 2008-05-16 | Scheduling method of grid resources of multi-Qos |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101582831A true CN101582831A (en) | 2009-11-18 |
| CN101582831B CN101582831B (en) | 2012-06-13 |
Family
ID=41364799
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2008100375647A Expired - Fee Related CN101582831B (en) | 2008-05-16 | 2008-05-16 | Scheduling method of grid resources of multi-Qos |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101582831B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102184121A (en) * | 2011-05-13 | 2011-09-14 | 南京财经大学 | Grid service quality scheduling method based on Markov chain |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107451740A (en) * | 2017-08-02 | 2017-12-08 | 长威信息科技发展股份有限公司 | A kind of method and system of the scheduling of resource changed based on grid cell |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100435159C (en) * | 2006-08-24 | 2008-11-19 | 上海交通大学 | Method for implementing mesh based optical mesh emulation platform |
| CN100495984C (en) * | 2007-06-07 | 2009-06-03 | 武汉理工大学 | Grid task dispatching method based on time and expense restriction |
| CN100511157C (en) * | 2007-09-20 | 2009-07-08 | 武汉理工大学 | Network resource dispatching method based on market economy model |
-
2008
- 2008-05-16 CN CN2008100375647A patent/CN101582831B/en not_active Expired - Fee Related
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102184121A (en) * | 2011-05-13 | 2011-09-14 | 南京财经大学 | Grid service quality scheduling method based on Markov chain |
| CN102184121B (en) * | 2011-05-13 | 2012-11-14 | 南京财经大学 | Grid service quality scheduling method based on Markov chain |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101582831B (en) | 2012-06-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10474504B2 (en) | Distributed node intra-group task scheduling method and system | |
| CN106651089A (en) | Modeling and optimization method for production scheduling problem distribution set robust model | |
| CN103473122B (en) | Workflow system resource scheduling method in cloud computing environment | |
| CN105260818A (en) | Online optimized scheduling method for workflow groups with deadline constraint in mixed cloud environment | |
| CN110308967A (en) | A kind of workflow cost based on mixed cloud-delay optimization method for allocating tasks | |
| CN108154317B (en) | Workflow group scheduling method based on example self-adaptive distribution integration in multi-cloud environment | |
| Rewers et al. | A study of priority rules for a levelled production plan | |
| CN103401939A (en) | Load balancing method adopting mixing scheduling strategy | |
| CN110414863A (en) | A kind of intelligence manufacture workshop resource regulating method | |
| CN101237469A (en) | Method for optimizing multi-QoS grid workflow based on ant group algorithm | |
| Park et al. | Due-date scheduling on parallel machines with job splitting and sequence-dependent major/minor setup times | |
| CN109298920A (en) | Based on the quasi- mixing key task dispatching method for dividing thought | |
| US8281313B1 (en) | Scheduling computer processing jobs that have stages and precedence constraints among the stages | |
| CN106934537A (en) | The sub- time limit based on the scheduling of reverse operation stream obtains optimization method | |
| Li et al. | A QoS-based scheduling algorithm for instance-intensive workflows in cloud environment | |
| Singh et al. | A comparative study of various scheduling algorithms in cloud computing | |
| CN101582831B (en) | Scheduling method of grid resources of multi-Qos | |
| CN110264097A (en) | More scientific workflows based on cloud environment concurrently execute dispatching method | |
| Zhang et al. | A discrete job-shop scheduling algorithm based on improved genetic algorithm | |
| Dahal et al. | Scheduling in multiprocessor system using genetic algorithms | |
| Tarasova et al. | Accounting for large jobs for a single-processor online model | |
| CN109783189A (en) | A kind of quiescent operation stream scheduling method and device | |
| CN103679404A (en) | Multi-project resource balancing and optimizing method | |
| Salmani et al. | A fuzzy-based multi-criteria scheduler for uniform multiprocessor real-time systems | |
| CN112700169A (en) | Business process task allocation method and system based on prediction and personnel feedback |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120613 Termination date: 20140516 |