CN105912409A - Task scheduling method and device - Google Patents

Abstract

The invention discloses a task scheduling method and device, relates to the technical field of system resource processing and solves technical problems that scheduling algorithm consideration factors are relatively single in the prior art, and QoS (Quality of Service) demands of a user are hard to be met in a resource scheduling process. The task scheduling method comprises the steps: receiving a to-be-completed job submitted by a client, and dividing the job into multiple tasks, wherein the job comprises QoS-required time and cost parameters of the user; aiming at each task to determine a scheduling solution for completing the task, wherein the scheduling solution determining process specifically comprises: finding out at least one scheduling path meeting the task according to an ant algorithm, and determining an optimal path from the at least one scheduling path according to a QoS constraint function based on the time and cost parameters.

Description

A kind of method for scheduling task and device

Technical field

The present invention relates to system resource processing technology field, particularly relate to a kind of method for scheduling task and device.

Background technology

MapReduce is a kind of programming model, for the concurrent operation of large-scale dataset (more than 1TB). Concept " Map (mapping) " and " Reduce (reduction) ", be their main thought, and it is the most convenient The program of oneself will not operated in distributed system in the case of distributed parallel programming by programming personnel On.Current software realizes being to specify Map (mapping) function, is used for one group of key-value pair to map Become one group of new key-value pair, it is intended that concurrent Reduce (reduction) function, be used for ensureing the key of all mappings Each of value centering shares identical key group.MapReduce is by dividing the large-scale operation of data set The each node issued on network realizes reliability；Each node can periodically return the work that it is completed With up-to-date state.After system receives an operation (Job), automatically wait to locate by an operation (Job) The big data of reason are divided into a lot of data blocks, and each data block calculates task (Task) corresponding to one, And Automatic dispatching calculates node and processes corresponding data block.Operation and task scheduling function are mainly responsible for distribution Calculate node (Map node or Reduce node) with scheduling, be responsible for monitoring the execution shape of these nodes simultaneously State, and the Synchronization Control that responsible Map node performs.

In MapReduce, Jobtracker (job trace) node is responsible for constantly monitoring Jobclient (work Industry client), wait Job to be committed and be divided into multiple Map Tasks and Reduce Tasks.So Perform, by scheduling Tasktrackers (task tracker), the Job that user submits to afterwards according to scheduling strategy. The Job that reasonably Jobclient should be able to be submitted to by scheduling strategy maps, comprehensively with suitably resource in system Consider that many factors meets user's actual QoS (Quality of Service, service quality) demand and maintains The resource utilization that system is higher.But, existing dispatching algorithm Consideration is the most single, is difficult in money While the scheduling in source, meet the QoS demand of user.

Summary of the invention

The embodiment of the present invention provides a kind of method for scheduling task and device, is used for solving scheduling in prior art and calculates Method Consideration is the most single, is difficult to while the scheduling of resource, meets the technology of the QoS demand of user Problem.

A kind of method for scheduling task, including:

Receive the operation that the needs of client submission complete, and described operation is divided into multiple task, its In, described operation comprises the time cost parameter of user's QoS demand；

For each task, having determined the scheduling scheme of described task, wherein, this determines scheduling scheme Process specifically includes:

At least one scheduling path meeting described task is searched according to ant algorithm；

Dispatch path true from described at least one according to QoS constraint function based on described time cost parameter Make optimal path.

Optionally, the method also includes: often finds a scheduling path and all adds up fortune in described scheduling path Row deadline of wanting of described required by task and complete expense.

Wherein, described QoS constraint function based on described time cost parameter, particularly as follows:

Con=a × con_Time+b × con_Charge

Wherein, a is time parameter, and b is cost parameters, and con_Time is time-constrain function, con_Charge For expense restriction function, a+b=1.

Optionally, during described time-constrain function is described scheduling path based on described statistics, the deadline is Little, the deadline is maximum and runs described scheduler task terminates the required time and obtain.

Optionally, described expense restriction function completes expense in being described scheduling path based on described statistics Little, the expense that completes is maximum and runs described scheduler task terminates required expense and obtain.

In the method that the embodiment of the present invention provides, meet described task by using to search according to ant algorithm At least one scheduling path；And according to QoS constraint function based on described time cost parameter from described at least Article one, scheduling path is determined the technological means of optimal path, dispatching algorithm in prior art can be solved and examine Worry factor is the most single, is difficult to while the scheduling of resource, meets the technical problem of the QoS demand of user, Achieve the time of not only Job being run by user, cost requirement as Consideration, and can also meet Its actual QoS requirement, and maintain each node load relative equilibrium, improve the technology of resource utilization ratio Effect.

Based on same inventive concept, the embodiment of the present invention continues to provide a kind of task scheduling apparatus, including:

Receiver module, the operation that the needs submitted to for receiving client complete, and described operation is divided For multiple tasks；Wherein, described operation comprises the time cost parameter of user's QoS demand；

Search module, for for each task, search according to ant algorithm and meet described scheduler task extremely A few scheduling path；

Determine module, for basis QoS constraint function based on described time cost parameter from described at least one Bar scheduling determines optimal path in path.

Optionally, this device also includes:

Statistical module, for all adding up in described scheduling path whenever lookup module searches to scheduling path Run deadline that described required by task wants and complete expense.

Optionally, described QoS constraint function based on described time cost parameter, particularly as follows:

Con=a × con_Time+b × con_Charge

Wherein, a is time parameter, and b is cost parameters, and con_Time is time-constrain function, con_Charge For expense restriction function, a+b=1.

Optionally, during described time-constrain function is described scheduling path based on described statistics, the deadline is Little, the deadline is maximum and runs described scheduler task terminates the required time and obtain.

Optionally, described expense restriction function completes expense in being described scheduling path based on described statistics Little, the expense that completes is maximum and runs described scheduler task terminates required expense and obtain.

During what the embodiment of the present invention provided transfers accounts, have and meet described task at least according to ant algorithm lookup Article one, scheduling path；And according to QoS constraint function based on described time cost parameter from described at least one Scheduling determines the function of optimal path in path, can solve in prior art dispatching algorithm Consideration relatively For single, it is difficult to while the scheduling of resource, meets the technical problem of the QoS demand of user, it is achieved that Not only Job is run by user the time, cost requirement as Consideration, and it is actual to meet it QoS demand, and maintain each node load relative equilibrium, improve the technique effect of resource utilization ratio.

Other features and advantages of the present invention will illustrate in the following description, and, partly from explanation Book becomes apparent, or understands by implementing the present invention.The purpose of the present invention and other advantages can Realize by structure specifically noted in the description write, claims and accompanying drawing and obtain ?.

Below by drawings and Examples, technical scheme is described in further detail.

Accompanying drawing explanation

Accompanying drawing is for providing a further understanding of the present invention, and constitutes a part for description, with this Bright embodiment is used for explaining the present invention together, is not intended that limitation of the present invention.In the accompanying drawings:

The flow chart of a kind of method for scheduling task that Fig. 1 provides for the embodiment of the present invention one；

The flow chart of the another kind of method for scheduling task that Fig. 2 provides for the embodiment of the present invention two；

The structural representation of a kind of task scheduling apparatus that Fig. 3 provides for the embodiment of the present invention three.

Detailed description of the invention

Below in conjunction with accompanying drawing, the preferred embodiments of the present invention are illustrated, it will be appreciated that described herein Preferred embodiment is merely to illustrate and explains the present invention, is not intended to limit the present invention.

Embodiment one

See Fig. 1, a kind of method for scheduling task that the embodiment of the present invention provides, the method includes:

101, receive the operation that the needs of client submission complete, and described operation is divided into multiple Business, wherein, comprises the time cost parameter of user's QoS demand in described operation；

The operation and the process that division operation is multiple task that receive client submission all can be according to prior aries Perform, time different, operation includes the time cost parameter of user's QoS demand, this time cost Parameter is the requirement for the time for fulfiling assignment and expense, it is also possible to as each task Requirement.Time cost parameter be i.e. time factor and expense factor, express time and price lay particular stress on ratio, as Time 80%, price 20%, illustrate that the QoS demand of user is desirable to task and performs as early as possible, charge height does not closes System.

For each task, having determined the scheduling scheme of described task, wherein, this determines scheduling scheme Process specifically includes (it should be understood that the scheduling scheme provided in the embodiment of the present invention is all to appoint with one Business for unit description how for completing the resource regulating method selected by this task):

102, for each task, search at least one scheduling road meeting described task according to ant algorithm Footpath.

Concrete, this implementation of 102 can refer to the 201-206 in following embodiment two.

Optionally, after 102, the method may also include that often finding a scheduling path all adds up described tune Degree path is run deadline that described required by task wants and completes expense.

103, according to QoS constraint function based on described time cost parameter from described at least one scheduling road Footpath is determined optimal path.

Wherein, described QoS constraint function based on described time cost parameter, specially formula (1):

Con=a × con_Time+b × con_Charge (1)

Wherein, a is time parameter (time factor), and b is cost parameters (expense factor), and con_Time is Time-constrain function, con_Charge is expense restriction function, a+b=1.

Optionally, during described time-constrain function is described scheduling path based on described statistics, the deadline is Little, the deadline is maximum and runs described scheduler task terminates the required time and obtain, the concrete time is about The expression formula of bundle function can refer to following formula (2).

Optionally, described expense restriction function completes expense in being described scheduling path based on described statistics Little, the expense that completes is maximum and runs described scheduler task terminates required expense and obtain, concrete expense is about The expression formula of bundle function can refer to following formula (3).

In the method that the embodiment of the present invention provides, meet described task by using to search according to ant algorithm At least one scheduling path；And according to QoS constraint function based on described time cost parameter from described at least Article one, scheduling path is determined the technological means of optimal path, dispatching algorithm in prior art can be solved and examine Worry factor is the most single, is difficult to while the scheduling of resource, meets the technical problem of the QoS demand of user, Achieve the time of not only Job being run by user, cost requirement as Consideration, and can also meet Its actual QoS requirement, and maintain each node load relative equilibrium, improve the technology of resource utilization ratio Effect.

Following for readily appreciating, the embodiment of the present invention specifically describes based on described time cost parameter The derivation of QoS constraint function (1):

1, table 1 below is the definition of relevant parameter of Task (task) mathematical description

Table 1

Parameter	Explanation
		A	The Task sum divided
Ti	I-th Task
		V	The sum of system interior joint
NODEj	Jth NODE
		PTF(Ti,NODEj)	A*V matrix task Ti Estimated Time Of Operation on NODEj
RCU(NODEj)	Task is the expense of run unit time on NODEj

The mathematical description of Task can be obtained by upper table 1:

(T_i,NODE_j,PTF(T_i,NODE_j),RCU(NODE_j)) wherein (i ∈ [1, A], j ∈ [1, V])

2, table 2 below is that Task runs time and the relevant parameter explanation of expense

Table 2

Parameter	Explanation
		m	The quantity of the Task distributed on NODEj
Time(Ti,NODEj)	The operation time required for the upper Ti of NODEj
		T_Lengthi	The length of Ti
IFSi	Task deducts and needs the length of remaining information after executable portion
		N_BWj	The bandwidth of NODEj
PNj	The anticipated disposal ability of NODEj
		pesNumj/perMipsj	The speed of the quantity of processor/each processor in NODEj

Each NODE completes the calculating of the Estimated Time Of Operation V_Time needed for distributing to its all Tasks As formula is:

$N\_Time\left({\mathrm{NODE}}_j\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}Time(T_i,{\mathrm{NODE}}_j)$

Wherein:

$Time(T_i,{\mathrm{NODE}}_j)=\frac{T\_{\mathrm{Length}}_i}{{\mathrm{PN}}_j}+\frac{{\mathrm{IFS}}_i}{N\_{\mathrm{BW}}_j}$

PN_j=pesNum_j×perMips_j

Owing to the Tasks in each node is parallel running, so the time needed for Task end of run

(finishTime) computing formula is that following formula (4) is:

FinishTime=MAX (N_Time (NODE_j)) (4)

All needed for Task end of run, cost formula is following formula (5):

$finishCh\arg e=\underset{j=1}{\overset{V}{\mathrm{\Σ}}}N\_Time\left({\mathrm{NODE}}_j\right)\×RCU\left({\mathrm{NODE}}_j\right)---\left(5\right)$

3, table 3 is designated as the explanation of QoS constraint function relevant parameter under being:

Table 3

Parameter	Definition
		con_Time/con_Charge	Time/expense restriction function
finishTimeMIN/finishTimeMAX	The Task operation time on the NODE that configuration is best/worst
		finishChargeMIN/finishChargeMAX	Task runs required minimum/ultimate cost on different NODE

$con\_Time=\frac{finishTime-{\mathrm{finishTime}}_{MIN}}{{\mathrm{finishTime}}_{MAX}-{\mathrm{finishTime}}_{MIN}}---\left(2\right)$

$con\_Ch\arg e=\frac{finishCh\arg e-{\mathrm{finishCharge}}_{MIN}}{{\mathrm{finishCharge}}_{MAX}-{\mathrm{finishCharge}}_{MIN}}---\left(3\right)$

Wherein:

$finish\_{\mathrm{Time}}_{MIN}=\frac{\underset{i=1}{\overset{A}{\mathrm{\Σ}}}T\_{\mathrm{Length}}_i}{V\×MAX\left({\mathrm{PN}}_j\right)}+\frac{\underset{i=1}{\overset{A}{\mathrm{\Σ}}}{\mathrm{IFS}}_i}{V\×N\_{\mathrm{BW}}_j}$

$finish\_{\mathrm{Time}}_{MAX}=\frac{\underset{i=1}{\overset{A}{\mathrm{\Σ}}}T\_{\mathrm{Length}}_i}{V\×MIN\left({\mathrm{PN}}_j\right)}+\frac{\underset{i=1}{\overset{A}{\mathrm{\Σ}}}{\mathrm{IFS}}_i}{V\×N\_{\mathrm{BW}}_j}$

finishCh arg e_MIN=finishTime_MIN×MIN(RCU(NODE_j))

finishCh arg e_MAX=finishTime_MAX×MAX(RCU(NODE_j))

4, QoS constraint function formula (1) is finally given

Con=a × con_Time+b × con_Charge (1)

Wherein: a, b are the time cost factor respectively, value is all between 0～1, and a+b=1.They Value is to be formulated by the actual QoS requirement of user.The factor corresponding to parameter that sensitivity is high is set as bigger Value.Such as, if a takes 1, b takes 0, then QoS-LBACO can be run the principle of minimal time by Task Carry out task scheduling.

Embodiment two

Based on above-mentioned QoS constraint function, the embodiment of the present invention specifically provides method for scheduling task, the party Method is as a example by a task (task), the tune by completing the scheduling of resource that this task is carried out of description The determination process of degree scheme.Specifically, as in figure 2 it is shown, the method includes:

201, after loop initialization number of times count=0, initialize the pheromone of resource node in all-network Value；

Wherein, in initialization MapReduce, on each Node, the value of pheromone is realized by following formula:

τ_Nj=pesNum_j×perMips_j+N_BW_j

202, random is put into each Formica fusca (realizing the device of ant algorithm) on each resource node；

203, as a example by Formica fusca X, Formica fusca X is that task rotates next resource joint according to transition function Point；

Wherein, Formica fusca, when selecting Node for Task, first has to other Node in calculating system and processes letter The ability of breath element concentration and load balancing degrees, then according to the probability that Node is selected, with " roulette " Method is that task selects next VM.Relevant parameter is described as follows shown in table 4

Table 4

Parameter	Definition
		ηNj	Ti selects the expected value of NODEj
VPj	The attribute VPj=τ of NODEVj(0)
		LBNj	NODEj load balancing degrees in systems
BAT	The average performance times of each NODE in optimal solution before next time circulating

T Formica fusca is that the probability function of Task Ti selection NODEj is:

$P(t,T_i,{\mathrm{NODE}}_j)=\frac{{{\mathrm{\τ}}_{\mathrm{Nj}}}^{\mathrm{\α}}\left(t\right)\×{{\mathrm{\η}}_{\mathrm{Nj}}}^{\mathrm{\β}}}{\underset{i=1}{\overset{V}{\mathrm{\Σ}}}{{\mathrm{\τ}}_i}^{\mathrm{\α}}\left(t\right)\×{{\mathrm{\η}}_i}^{\mathrm{\β}}}$

This formula is only set up when i, j ∈ 1...V, and in the case of other, probability is zero.

Wherein:

η_Nj=NP_j×LB_Nj

$P(t,T_i,{\mathrm{NODE}}_j)=\frac{{{\mathrm{\τ}}_{Nj}}^{\mathrm{\α}}\left(t\right)\×{{\mathrm{\η}}_{Nj}}^{\mathrm{\β}}}{\underset{i=1}{\overset{V}{\mathrm{\Σ}}}{{\mathrm{\τ}}_i}^{\mathrm{\α}}\left(t\right)\×{{\mathrm{\η}}_i}^{\mathrm{\β}}}$

α, β are respectively intended to control the disturbance degree of τ and η.

204, it may be judged whether find a kind of scheduling path？If so, 205 are performed；Otherwise, 203 are performed；

205, calculate and in this scheduling path, run deadline that this required by task wants and complete expense, and Carry out local message element renewal.

Wherein, deadline finishtime added up in each scheduling path calculated and complete expense Finishcharge, be easy on each path of comparison incurs the time and expense it is known that size order therein, It is easy in following 206 QoS constraint function uses.

Wherein, Pheromone update formula is:

τ_Nj(t+1)=(1-ρ) × τ_Nj(t)+Δτ_Nj

Wherein, ρ ∈ (0,1] the representative information element disappearance factor.Δτ_NjDetermine by con for pheromone increment, letter More capable local updating and the overall situation two kinds of situations of renewal of including of breath element:

After a Formica fusca finds a paths, it is a task when finding a node, on path All node can carry out local message element renewal.

Now,

Δτ_Nj=H₁/con(S_ki) (6)

Ski represents the allocative decision that i-th Formica fusca is found in kth time circulation, and H1 is constant.

In once circulation, after whole Formica fuscas find a paths the most respectively, find the optimal solution that this circulates, Then local message element renewal is carried out for the Nodes on path.

Δτ_Nj=H₂/MIN(con(S_ki)) (7)

Wherein MIN (con (S_ki)) represent the optimal solution that in kth time circulation, all Formica fuscas are found by Jobclient.H₂ For constant.

206, it may be judged whether every Formica fusca all finds a kind of allocative decision？The most then perform 207；Otherwise, Perform 203；

207, find out the optimal path of this circulation.

It is determined with specific reference to formula (1).

208, the resource node on path, optimal path place is carried out full detail element renewal；

209, cycle-index count adds 1, and judges that whether the value of count is more than the cycle-index preset M-count.The most then perform 210；Otherwise, 202 are performed.

210, export optimal path.

The method for scheduling task that the embodiment of the present invention provides, is a kind of load balancing ant colony based on QoS constraint Optimized algorithm (QoS-LBACO) dispatching method, not only runs the requirement of time, cost by user to Job As Consideration, and its actual QoS requirement can also be met, and maintain each node load relative equilibrium, Improve resource utilization ratio.

The overall thought of the QoS-LBACO that the embodiment of the present invention proposes is: by by " ACO (ant group algorithm) Solve TSP (travelling salesman) problem " method be applied in the QoS-LBACO dispatching method of the present invention real Existing.

Table 5 below is that " ACO (ant group algorithm) solves TSP (travelling salesman) problem " dispatches plan with QoS-LBACO A simple corresponding relation slightly

Table 5

Setting up QoS-LBACO mathematical model according to upper table in conjunction with QoS constraint function, Formica fusca can be according to transfer Probability function roulette method is that next Map and Reduce Task selects VM, and circulation always is it is known that every Individual Task finds suitable node.Meanwhile, after Formica fusca once finds one " path ", it is first calculated Runtime and expense.Then in " path " found by each Formica fusca, Runtime and expense are entered Row compares, and the pheromone value updated in network topology in respective nodes, finally further according to the actual QoS of user Demand selects an optimal scheduling scheme, such as network minimal or shortest time or expense relatively low with constantly Between the QoS demand such as the shortest.

Embodiment three

Realizing for the ease of the method in above-described embodiment one and two, the embodiment of the present invention continues to provide a task Dispatching device, as it is shown on figure 3, this device includes:

Receiver module 31, the operation that the needs submitted to for receiving client complete, and described operation is drawn It is divided into multiple task；Wherein, described operation comprises the time cost parameter of user's QoS demand；

Search module 32, for for each task, search according to ant algorithm and meet described scheduler task At least one scheduling path；

Determine module 33, for according to QoS constraint function based on described time cost parameter from described at least Article one, optimal path is determined in scheduling path.

Optionally, this device also includes:

Statistical module, for all adding up in described scheduling path whenever lookup module searches to scheduling path Run deadline that described required by task wants and complete expense.

Optionally, described QoS constraint function based on described time cost parameter, particularly as follows:

Con=a × con_Time+b × con_Charge

Wherein, a is time parameter, and b is cost parameters, and con_Time is time-constrain function, con_Charge For expense restriction function, a+b=1.

Optionally, during described time-constrain function is described scheduling path based on described statistics, the deadline is Little, the deadline is maximum and runs described scheduler task terminates the required time and obtain.

Optionally, described expense restriction function completes expense in being described scheduling path based on described statistics Little, the expense that completes is maximum and runs described scheduler task terminates required expense and obtain.

During what the embodiment of the present invention provided transfers accounts, have and meet described task at least according to ant algorithm lookup Article one, scheduling path；And according to QoS constraint function based on described time cost parameter from described at least one Scheduling determines the function of optimal path in path, can solve in prior art dispatching algorithm Consideration relatively For single, it is difficult to while the scheduling of resource, meets the technical problem of the QoS demand of user, it is achieved that Not only Job is run by user the time, cost requirement as Consideration, and it is actual to meet it QoS demand, and maintain each node load relative equilibrium, improve the technique effect of resource utilization ratio.

Other features and advantages of the present invention will illustrate in the following description, and, partly from explanation Book becomes apparent, or understands by implementing the present invention.The purpose of the present invention and other advantages can Realize by structure specifically noted in the description write, claims and accompanying drawing and obtain ?.

Those skilled in the art are it should be appreciated that embodiments of the invention can be provided as method, system or meter Calculation machine program product.Therefore, the present invention can use complete hardware embodiment, complete software implementation or knot The form of the embodiment in terms of conjunction software and hardware.And, the present invention can use and wherein wrap one or more Computer-usable storage medium containing computer usable program code (include but not limited to disk memory and Optical memory etc.) form of the upper computer program implemented.

The present invention is with reference to method, equipment (system) and computer program product according to embodiments of the present invention The flow chart of product and/or block diagram describe.It should be understood that can by computer program instructions flowchart and / or block diagram in each flow process and/or flow process in square frame and flow chart and/or block diagram and/ Or the combination of square frame.These computer program instructions can be provided to general purpose computer, special-purpose computer, embedding The processor of formula datatron or other programmable data processing device is to produce a machine so that by calculating The instruction that the processor of machine or other programmable data processing device performs produces for realizing at flow chart one The device of the function specified in individual flow process or multiple flow process and/or one square frame of block diagram or multiple square frame.

These computer program instructions may be alternatively stored in and computer or the process of other programmable datas can be guided to set In the standby computer-readable memory worked in a specific way so that be stored in this computer-readable memory Instruction produce and include the manufacture of command device, this command device realizes in one flow process or multiple of flow chart The function specified in flow process and/or one square frame of block diagram or multiple square frame.

These computer program instructions also can be loaded in computer or other programmable data processing device, makes Sequence of operations step must be performed to produce computer implemented place on computer or other programmable devices Reason, thus the instruction performed on computer or other programmable devices provides for realizing flow chart one The step of the function specified in flow process or multiple flow process and/or one square frame of block diagram or multiple square frame.

Obviously, those skilled in the art can carry out various change and modification without deviating from this to the present invention Bright spirit and scope.So, if the present invention these amendment and modification belong to the claims in the present invention and Within the scope of its equivalent technologies, then the present invention is also intended to comprise these change and modification.

Claims (10)

1. a method for scheduling task, it is characterised in that including:

Receive the operation that the needs of client submission complete, and described operation is divided into multiple task, its In, described operation comprises the time cost parameter of user's QoS demand；

For each task, having determined the scheduling scheme of described task, wherein, this determines scheduling scheme Process specifically includes:

At least one scheduling path meeting described task is searched according to ant algorithm；

Dispatch path true from described at least one according to QoS constraint function based on described time cost parameter Make optimal path.

Method the most according to claim 1, it is characterised in that the method also includes: often find Article one, scheduling path is all added up and is run deadline that described required by task wants in described scheduling path and complete Expense.

Method the most according to claim 2, it is characterised in that described join based on described time cost The QoS constraint function of number, particularly as follows:

Con=a × con_Time+b × con_Charge

Wherein, a is time parameter, and b is cost parameters, and con_Time is time-constrain function, con_Charge For expense restriction function, a+b=1.

Method the most according to claim 3, it is characterised in that described time-constrain function be based on In the described scheduling path of described statistics, the deadline is minimum, the deadline is maximum and runs described scheduler task Time needed for end and obtain.

Method the most according to claim 3, it is characterised in that described expense restriction function be based on Complete expense minimum in the described scheduling path of described statistics, the expense that completes is maximum and runs described scheduler task Expense needed for end and obtain.

6. a task scheduling apparatus, it is characterised in that including:

Receiver module, the operation that the needs submitted to for receiving client complete, and described operation is divided For multiple tasks；Wherein, described operation comprises the time cost parameter of user's QoS demand；

Search module, for for each task, search according to ant algorithm and meet described scheduler task extremely A few scheduling path；

Determine module, for basis QoS constraint function based on described time cost parameter from described at least one Bar scheduling determines optimal path in path.

Device the most according to claim 6, it is characterised in that this device also includes:

Statistical module, for all adding up in described scheduling path whenever lookup module searches to scheduling path Run deadline that described required by task wants and complete expense.

Device the most according to claim 7, it is characterised in that described join based on described time cost The QoS constraint function of number, particularly as follows:

Con=a × con_Time+b × con_Charge

Wherein, a is time parameter, and b is cost parameters, and con_Time is time-constrain function, con_Charge For expense restriction function, a+b=1.

Device the most according to claim 8, it is characterised in that described time-constrain function be based on In the described scheduling path of described statistics, the deadline is minimum, the deadline is maximum and runs described scheduler task Time needed for end and obtain.

Device the most according to claim 8, it is characterised in that described expense restriction function be based on Complete expense minimum in the described scheduling path of described statistics, the expense that completes is maximum and runs described scheduler task Expense needed for end and obtain.