CN104536828A - Cloud computing task scheduling method and system based on quantum-behaved particle swarm algorithm - Google Patents

Abstract

The invention provides a cloud computing task scheduling method and system based on a quantum-behaved particle swarm algorithm. The method comprises the steps of defining and storing an objective function of task scheduling, randomly initializing the speed and position of a particle swarm, calculating adaptive values of all particles, determining the optimal positions of all the particles and the optimal position of the particle swarm according to the adaptive values of all the particles, updating the speeds and positions of the particles, judging whether the number of iterations reaches the preset number of iterations or not, and if yes, outputting a cloud computing task scheduling result. According to the method, task scheduling is carried out by defining the objective function, the quantum-behaved particle swarm algorithm is used, various resources of cloud computing is fully used, the quantum-behaved particle swarm algorithm is used for looking for spare nodes, and each task is made to find an optimal path according to the optimal solution of the quantum-behaved particle swarm to finish the task scheduling, and therefore the utilization rate of resources and the task scheduling efficiency are effectively improved.

Description

Based on method for scheduling task and the system of the cloud computing of quanta particle swarm optimization

Technical field

Field of computer technology of the present invention, particularly relates to a kind of method for scheduling task of the cloud computing based on quanta particle swarm optimization.

Background technology

Thin cloud is exactly a branch of cloud computing, refers to cloud computing some application in individual subrange.With regard to education sector, what thin cloud technology was applied at present is not very wide.Under thin cloud Technical Architecture, colleges and universities do not need the infrastructure of building oneself, according to infrastructure such as the flexible rental service devices of oneself school information construction demand, thus can reduce investment construction, save informatization cost.Under many tenants pattern, single hardware infrastructure is shared by multiple tenant simultaneously and is used, when tenant's load increases, performance can be caused to disturb (performance interference) to the contention of shared resource between different tenant, therefore, need to carry out performance isolation (performance isolation) towards tenant.For the monitoring of CPU class resource of this time, conventional art is still difficult to the monitoring resource demand met under many tenants pattern of high degree of share.Direct monitoring method, owing to there is system compatible sex chromosome mosaicism, is difficult to be applied to the cloud computing environment that there is a large amount of heterogeneous system.Based on the indirect analysis method of program translation, system overhead is too high, is difficult to be applied to actual production environment.The method of Corpus--based Method analysis does not rely on any operating system, there is good system compatibility, and owing to using monitoring tolerance conventional in production environment as the input of statistical study, therefore can not produce system overhead, but still there is the problems such as poor in timeliness, low to load change adaptability, estimation result accuracy is low.

Mezmaz M etc. proposes a kind of parallel double target mixed method, scheduling problem is modeled as minimum completion time and energy consumption biobjective scheduling problem, and utilize pareto genetic algorithm to ask to change multi-objective optimization question, the Bi-objective genetic algorithm of proposition has taken into account the minimum of the minimum of energy consumption and application deadline.Cui etc. improve genetic algorithm, propose a kind of resource regulating method, task are used directed acyclic graph theory, adopt short task priority strategy to shorten task execution time.SUN D W etc. propose the preference multidimensional QO cloud scheduling of resource optimized algorithm based on immune clone, improve the convergence capabilities of optimum solution, and can keep load balancing.Lin W etc. propose a kind of Dynamic Resource Allocation for Multimedia algorithm based on threshold values, reconfigure virtual resource allocation algorithm, heighten resource utilization according to the change tread of load, reduce the cost of use of user.Just current, most of energy consumption model only considers the utilization factor of cpu, does not consider the energy consumption problem of the various physical resources under the various states such as the master dormant under thin cloud technology, free time and work comprehensively.

Summary of the invention

Based on this, be necessary to provide a kind of cloud computing method for scheduling task based on quanta particle swarm optimization making full use of cloud computing physical resource.Based on a cloud computing method for scheduling task for quanta particle swarm optimization, comprise the following steps:

Define and the objective function of store tasks scheduling, described objective function is:

wherein, x _ijrepresent that i-th subtask is run on a jth virtual resource node, n is the quantity of task; M is the quantity of dummy node; Tx _ijrepresent the time that i-th subtask is run at a jth virtual resource node;

The speed of random initializtion population and position;

Calculate the adaptive value of all particles, and determine optimal location and colony's optimal location of all particles according to the adaptive value of all particles;

The more speed of new particle and position;

Judge whether iterations arrives default iterations;

If so, cloud computing task scheduling result is then exported.

Wherein in a kind of embodiment, before the speed of described random primary group and the step of position, also comprise step:

Pre-set the correlation parameter of quanta particle swarm optimization.

Wherein in a kind of embodiment, the adaptive value of all particles of described calculating, and determine that the optimal location of all particles and the formula of colony's optimal location are according to the adaptive value of all particles:

$\mathrm{sig}\left(v_{\mathrm{ij}}^{k+1}\right)=\frac{1}{1+\exp (-v_{\mathrm{ij}}^{k+1})}$ With $\mathrm{sig}\left(x_{\mathrm{ij}}^{k+1}\right)=\frac{1}{1+\exp (-x_{\mathrm{ij}}^{k+1})};$ Wherein, described K is iterations, the position vector of particle, value is 1 or 0, for speed.

Wherein in a kind of embodiment, the formula of the speed and position that upgrade described particle is:

V _i(t+1)=ω (t) × V _i(t)+c ₁× r ₁× (lbest _i(t)-P _i(t))+c ₂× r ₂× (gbest (t)-P _i(t)) and P _i(t+1)=P _i(t)+V _i(t+1); Wherein, particle rapidity is the number of iterations that Vi, t represent current, and Pi is i-th particle; C1 and c2 represents speedup factor, r1 and r2 represents the random number be evenly distributed in interval (0,1); ω _maxfor the maximum magnitude of ω, ω _minfor the minimum value scope of ω, t is current iteration algebraically, and PM is population greatest iteration algebraically.

Wherein in a kind of embodiment, if judge, iterations does not arrive default iterations, then return and continue to perform the adaptive value of described calculating all particles, determining that the optimal location of all particles and the step of colony's optimal location are until judge that iterations arrives default iterations.

Based on a cloud computing task scheduling system for quanta particle swarm optimization, comprising:

Objective function memory module, for defining and the objective function of store tasks scheduling, described objective function is: wherein, x _ijrepresent that i-th subtask is run on a jth virtual resource node; N is the quantity of task; M is the quantity of dummy node; Tx _ijrepresent the time that i-th subtask is run at a jth virtual resource node;

Initialization module, for speed and the position of random initializtion population;

Best position calculation module: for calculating the adaptive value of all particles, determines optimal location and colony's optimal location of all particles;

Update module, for speed and the position of more new particle;

Judge module, for judging whether iterations arrives default iterations;

Task scheduling modules, during for judging that when judge module iterations arrives the iterations preset, exports cloud computing task scheduling result.

Wherein in a kind of embodiment, also comprise: presetting module, for pre-setting the correlation parameter of quanta particle swarm optimization.

Wherein in a kind of embodiment, described best position calculation module calculates the adaptive value of all particles, and determines that the optimal location of all particles and the formula of colony's optimal location are according to the adaptive value of all particles:

$\mathrm{sig}\left(v_{\mathrm{ij}}^{k+1}\right)=\frac{1}{1+\exp (-v_{\mathrm{ij}}^{k+1})}$ With $\mathrm{sig}\left(x_{\mathrm{ij}}^{k+1}\right)=\frac{1}{1+\exp (-x_{\mathrm{ij}}^{k+1})};$ Wherein, described K is iterations, the position vector of particle, value is 1 or 0, for speed.

Wherein in a kind of embodiment, described update module more

The formula of the speed and position that upgrade described particle is:

V _i(t+1)=ω (t) × V _i(t)+c ₁× r ₁× (lbest _i(t)-P _i(t))+c ₂× r ₂× (gbest (t)-P _i(t)) and P _i(t+1)=P _i(t)+V _i(t+1); Wherein, particle rapidity is the number of iterations that Vi, t represent current, and Pi is i-th particle; C1 and c2 represents speedup factor, r1 and r2 represents the random number be evenly distributed in interval (0,1); ω _maxfor the maximum magnitude of ω, ω _minfor the minimum value scope of ω, t is current iteration algebraically, and PM is population greatest iteration algebraically.

Wherein in a kind of embodiment, described best position calculation module, also for judging that iterations does not arrive default iterations when judge module, calculating the adaptive value of all particles, determining optimal location and colony's optimal location of all particles.

The above-mentioned cloud computing method for scheduling task based on quanta particle swarm optimization and system, the method is by objective definition function, and use particle cluster algorithm to carry out task scheduling, make full use of the multiple resources of cloud computing, particle algorithm is used to find idle node, allow each task can find optimal path according to the optimum solution of quantum particle swarm, scheduling of finishing the work, thus effectively raise the utilization factor of resource and the efficiency of task scheduling.

Accompanying drawing explanation

Fig. 1 is a kind of process flow diagram of the cloud computing method for scheduling task based on quanta particle swarm optimization of embodiment;

Fig. 2 is the process flow diagram of the cloud computing method for scheduling task based on quanta particle swarm optimization of another kind of embodiment;

Fig. 3 is a kind of cloud computing task scheduling modules figure based on quanta particle swarm optimization of embodiment.

Embodiment

In order to make technical scheme of the present invention and advantage clearly understand, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.

As shown in Figure 1, a kind of cloud computing method for scheduling task based on quanta particle swarm optimization, comprises the following steps:

S100: define and the objective function of store tasks scheduling, objective function is:

$f1=\min \left(F\right)=\min \left(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{\mathrm{Tx}}_{\mathrm{ij}}\right);$

Wherein, x _ijrepresent that i-th subtask is run on a jth virtual resource node, n is the quantity of task; M is the quantity of dummy node.

Assuming that within a certain period of time, there is n separate and need the subtask that arranges, representing T={t with task linear list T ₁, t ₂, t ₃... t _n, thin cloud service platform is m virtual resource node (m<=n) altogether, represents, VR={vr with linear VR ₁, vr ₂, vr ₃... vr _n, and specify that a subtask can only be run on a dummy node, use x _ijrepresent that i-th subtask is run on a jth virtual resource node.If i-th subtask is run on a jth virtual resource node, then obtain x _ij=1, otherwise the time Tx being 0, i-th subtask is run at a jth virtual resource node _ijrepresent.The time then completing all subtasks is set up objective function $f1=\min \left(F\right)=\min \left(\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}{\mathrm{Tx}}_{\mathrm{ij}}\right),$ The final purpose of algorithm makes minimize.

S300: the speed of random initializtion population and position.

The initial individuals of each particle, history optimal location and individual optimal-adaptive value evaluated, preserved according to the initial adaptive value of objective function to each particle, initial global history optimal location and optimal-adaptive value preserved simultaneously.

In the method for scheduling task of present embodiment, each value of particle position vector is 0 or 1, and particle rapidity no longer represents the speed of particle flight in continuous space, and its meaning is to calculate the probability that particle position vector value is 0 or 1.

In population, the position of each particle represents a feasible task scheduling approach, and according to relations of distribution matrix X, the position of i-th particle can be expressed as:

$X_i=\left(\begin{array}{cccc}{x}_{11}& x_{12}& ...& x_{1n}\\ x_{21}& x_{22}& ...& x_{2n}\\ .& .& & .\\ .& .& & .\\ .& .& & .\\ x_{m1}& x_{m2}& ...& x_{\mathrm{mn}}\end{array}\right);$

The speed definition of particle i is:

$X_i=\left(\begin{array}{cccc}{x}_{11}& x_{12}& ...& x_{1n}\\ x_{21}& x_{22}& ...& x_{2n}\\ .& .& & .\\ .& .& & .\\ .& .& & .\\ x_{m1}& x_{m2}& ...& x_{\mathrm{mn}}\end{array}\right).$

S400: the adaptive value calculating all particles, and optimal location and the colony's optimal location of determining all particles according to the adaptive value of all particles.

S500: the more speed of new particle and position.

S600: judge whether iterations arrives default iterations.

S700: if so, then export cloud computing task scheduling result.

The above-mentioned cloud computing method for scheduling task based on quanta particle swarm optimization, the method is by objective definition function, and use particle cluster algorithm to carry out task scheduling, make full use of the multiple resources of cloud computing, particle algorithm is used to find idle node, allow each task can find optimal path according to the optimum solution of quantum particle swarm, scheduling of finishing the work, thus effectively raise the utilization factor of resource and the efficiency of task scheduling.

In another embodiment, as shown in Figure 2, before step S300, also step S200 is comprised: the correlation parameter pre-setting quanta particle swarm optimization.Parameter comprises the parameter indexs such as speedup factor, compressibility factor, iteration weights, maximum iteration time and carries out initialization etc.

In another embodiment, S400

Calculate the adaptive value of all particles, and determine that the optimal location of all particles and the formula of colony's optimal location are according to the adaptive value of all particles:

$\mathrm{sig}\left(v_{\mathrm{ij}}^{k+1}\right)=\frac{1}{1+\exp (-v_{\mathrm{ij}}^{k+1})}$ With $\mathrm{sig}\left(x_{\mathrm{ij}}^{k+1}\right)=\frac{1}{1+\exp (-x_{\mathrm{ij}}^{k+1})}$ Wherein, described K is iterations the position vector of particle, value is 1 or 0, for speed.

In another embodiment, in S500

The speed of new described particle and the formula of position are:

V _i(t+1)=ω (t) × V _i(t)+c ₁× r ₁× (lbest _i(t)-P _i(t))+c ₂× r ₂× (gbest (t)-P _i(t)) and P _i(t+1)=P _i(t)+V _i(t+1);

Wherein, particle rapidity is the number of iterations that Vi, t represent current, and Pi is i-th particle; C1 and c2 represents speedup factor, r1 and r2 represents the random number be evenly distributed in interval (0,1); ω _maxfor the maximum magnitude of ω, ω _minfor the minimum value scope of ω and ω _max=1.2, ω _min=0.4, t is current iteration algebraically, and PM is population greatest iteration algebraically.

In another embodiment, as shown in Figure 2, if S600 judges that iterations does not arrive default iterations, then return and continue to perform the adaptive value that step S400 calculates all particles, determining that the optimal location of all particles and the step of colony's optimal location are until judge that iterations arrives the iterations preset.

The present invention also comprises a kind of cloud computing task scheduling system based on quanta particle swarm optimization, as shown in Figure 3, comprising:

Objective function memory module 100: for defining and the objective function of store tasks scheduling, described objective function is:

$f1=\min \left(F\right)=\min \left(\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}{\mathrm{Tx}}_{\mathrm{ij}}\right);$

Wherein, x _ijbe that i-th subtask is run on a jth virtual resource node.

Assuming that within a certain period of time, have N number of separate and need arrange subtask, represent T={t with task linear list T ₁, t ₂, t ₃... t _n, thin cloud service platform is M virtual resource node (M<=N) altogether, represents, VR={vr with linear VR ₁, vr ₂, vr ₃... vr _n, and specify that a subtask can only be run on a dummy node, use x _ijrepresent that i-th subtask is run on a jth virtual resource node.If i-th subtask is run on a jth virtual resource node, then obtain x _ij=1, otherwise the time Tx being 0, i-th subtask is run at a jth virtual resource node _ijrepresent.The time then completing all subtasks is set up objective function $f1=\min \left(F\right)=\min \left(\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}{\mathrm{Tx}}_{\mathrm{ij}}\right),$ The final purpose of algorithm makes minimize.

Initialization module 300: for speed and the position of random initializtion population.

The initial individuals of each particle, history optimal location and individual optimal-adaptive value evaluated, preserved according to the initial adaptive value of objective function to each particle, initial global history optimal location and optimal-adaptive value preserved simultaneously.

In the task scheduling system of present embodiment, each value of particle position vector is 0 or 1, and particle rapidity no longer represents the speed of particle flight in continuous space, and its meaning is to calculate the probability that particle position vector value is 0 or 1.

In population, the position of each particle represents a feasible task scheduling approach, and according to relations of distribution matrix X, the position of i-th particle can be expressed as:

$X_i=\left(\begin{array}{cccc}{x}_{11}& x_{12}& ...& x_{1n}\\ x_{21}& x_{22}& ...& x_{2n}\\ .& .& & .\\ .& .& & .\\ .& .& & .\\ x_{m1}& x_{m2}& ...& x_{\mathrm{mn}}\end{array}\right)$

The speed definition of particle i is:

$X_i=\left(\begin{array}{cccc}{x}_{11}& x_{12}& ...& x_{1n}\\ x_{21}& x_{22}& ...& x_{2n}\\ .& .& & .\\ .& .& & .\\ .& .& & .\\ x_{m1}& x_{m2}& ...& x_{\mathrm{mn}}\end{array}\right).$

Best position calculation module 400: for calculating the adaptive value of all particles, and optimal location and the colony's optimal location of determining all particles according to the adaptive value of all particles

Update module 500: for speed and the position of more new particle.

Judge module 600: for judging whether iterations arrives default iterations

Task scheduling modules 700: if so, then export cloud computing task scheduling result.

The above-mentioned cloud computing task scheduling system based on quanta particle swarm optimization, this system is by objective definition function, and use particle cluster algorithm to carry out task scheduling, make full use of the multiple resources of cloud computing, particle algorithm is used to find idle node, allow each task can find optimal path according to the optimum solution of quantum particle swarm, scheduling of finishing the work, thus effectively raise the utilization factor of resource and the efficiency of task scheduling.

In another embodiment, this system also comprises: presetting module 200, for pre-setting the correlation parameter of quanta particle swarm optimization.Parameter comprises the parameter indexs such as speedup factor, compressibility factor, iteration weights, maximum iteration time and carries out initialization etc.

In another embodiment, best position calculation module 400

Calculate the adaptive value of all particles, and determine that the optimal location of all particles and the formula of colony's optimal location are according to the adaptive value of all particles:

$\mathrm{sig}\left(v_{\mathrm{ij}}^{k+1}\right)=\frac{1}{1+\exp (-v_{\mathrm{ij}}^{k+1})}$ With $\mathrm{sig}\left(x_{\mathrm{ij}}^{k+1}\right)=\frac{1}{1+\exp (-x_{\mathrm{ij}}^{k+1})};$ Wherein, described K is iterations, the position vector of particle, value is 1 or 0, for speed.

In another embodiment, update module 500 more the speed of new particle and the formula of position be: V _i(t+1)=ω (t) × V _i(t)+c ₁× r ₁× (lbest _i(t)-P _i(t))+c ₂× r ₂× (gbest (t)-P _i(t)) and P _i(t+1)=P _i(t)+V _i(t+1);

Wherein, particle rapidity is the number of iterations that Vi, t represent current, and Pi is i-th particle; C1 and c2 represents speedup factor, r1 and r2 represents the random number be evenly distributed in interval (0,1); ω _maxfor the maximum magnitude of ω, ω _minfor the minimum value scope of ω and ω _max=1.2, ω _min=0.4, t is current iteration algebraically, and PM is population greatest iteration algebraically.

In another embodiment, excellent position computation module 400, also for judging that iterations does not arrive default iterations when judge module 600, calculating the adaptive value of all particles, determining optimal location and colony's optimal location of all particles.

The above embodiment only have expressed several embodiment of the present invention, and it describes comparatively concrete and detailed, but therefore can not be interpreted as the restriction to the scope of the claims of the present invention.It should be pointed out that for the person of ordinary skill of the art, without departing from the inventive concept of the premise, can also make some distortion and improvement, these all belong to protection scope of the present invention.Therefore, the protection domain of patent of the present invention should be as the criterion with claims.

Claims (10)

1., based on a cloud computing method for scheduling task for quanta particle swarm optimization, it is characterized in that, comprise the following steps:

Define and the objective function of store tasks scheduling, described objective function is:

$f1=\min \left(F\right)=\min \left(\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{m}{\mathrm{\&Sigma;}}}{\mathrm{Tx}}_{\mathrm{ij}}\right);$

Wherein, x _ijrepresent that i-th subtask is run on a jth virtual resource node, n is the quantity of task; M is the quantity of dummy node; Tx _ijrepresent the time that i-th subtask is run at a jth virtual resource node;

The speed of random initializtion population and position;

Calculate the adaptive value of all particles, and determine optimal location and colony's optimal location of all particles according to the adaptive value of all particles;

The more speed of new particle and position;

Judge whether iterations arrives default iterations;

If so, cloud computing task scheduling result is then exported.

2. the cloud computing method for scheduling task based on quanta particle swarm optimization according to claim 1, is characterized in that, before the speed of described random primary group and the step of position, also comprises step:

Pre-set the correlation parameter of quanta particle swarm optimization.

3. the cloud computing method for scheduling task based on quanta particle swarm optimization according to claim 1, it is characterized in that, the adaptive value of all particles of described calculating, and determine that the optimal location of all particles and the formula of colony's optimal location are according to the adaptive value of all particles:

$\mathrm{sig}\left(v_{\mathrm{ij}}^{k+1}\right)=\frac{1}{1+\exp (-v_{\mathrm{ij}}^{k+1})}$ With $\mathrm{sig}\left(x_{\mathrm{ij}}^{k+1}\right)=\frac{1}{1+\exp (-x_{\mathrm{ij}}^{k+1})};$ Wherein, described K is iterations, the position vector of particle, value is 1 or 0, for speed.

4. the cloud computing method for scheduling task based on quanta particle swarm optimization according to claim 1, is characterized in that, the formula of the speed and position that upgrade described particle is:

V _i(t+1)=ω (t) × V _i(t)+c ₁× r ₁× (lbest _i(t)-P _i(t))+c ₂× r ₂× (gbest (t)-P _i(t)) and P _i(t+1)=P _i(t)+V _i(t+1); Wherein, particle rapidity is the number of iterations that Vi, t represent current, and Pi is i-th particle; C1 and c2 represents speedup factor, r1 and r2 represents the random number be evenly distributed in interval (0,1); ω _maxfor the maximum magnitude of ω, ω _minfor the minimum value scope of ω, t is current iteration algebraically, and PM is population greatest iteration algebraically.

5. the cloud computing method for scheduling task based on quanta particle swarm optimization according to claim 4, it is characterized in that, if judge, iterations does not arrive default iterations, then return and continue to perform the adaptive value of described calculating all particles, determining that the optimal location of all particles and the step of colony's optimal location are until judge that iterations arrives default iterations.

6., based on a cloud computing task scheduling system for quanta particle swarm optimization, it is characterized in that, comprising:

Objective function memory module, for defining and the objective function of store tasks scheduling, described objective function is: wherein, x _ijrepresent that i-th subtask is run on a jth virtual resource node; N is the quantity of task; M is the quantity of dummy node; Tx _ijrepresent the time that i-th subtask is run at a jth virtual resource node;

Initialization module, for speed and the position of random initializtion population;

Best position calculation module: for calculating the adaptive value of all particles, determines optimal location and colony's optimal location of all particles;

Update module, for speed and the position of more new particle;

Judge module, for judging whether iterations arrives default iterations;

Task scheduling modules, during for judging that when judge module iterations arrives the iterations preset, exports cloud computing task scheduling result.

7. the cloud computing task scheduling system based on quanta particle swarm optimization according to claim 6, is characterized in that, also comprise: presetting module, for pre-setting the correlation parameter of quanta particle swarm optimization.

8. the cloud computing task scheduling system based on quanta particle swarm optimization according to claim 6, it is characterized in that, described best position calculation module calculates the adaptive value of all particles, and determines that the optimal location of all particles and the formula of colony's optimal location are according to the adaptive value of all particles:

$\mathrm{sig}\left(v_{\mathrm{ij}}^{k+1}\right)=\frac{1}{1+\exp (-v_{\mathrm{ij}}^{k+1})}$ With $\mathrm{sig}\left(x_{\mathrm{ij}}^{k+1}\right)=\frac{1}{1+\exp (-x_{\mathrm{ij}}^{k+1})};$ Wherein, described K is iterations, the position vector of particle, value is 1 or 0, for speed.

9. the cloud computing task scheduling system based on quanta particle swarm optimization according to claim 6, it is characterized in that, described update module more

The formula of the speed and position that upgrade described particle is:

V _i(t+1)=ω (t) × V _i(t)+c ₁× r ₁× (lbest _i(t)-P _i(t))+c ₂× r ₂× (gbest (t)-P _i(t)) and P _i(t+1)=P _i(t)+V _i(t+1); Wherein, particle rapidity is the number of iterations that Vi, t represent current, and Pi is i-th particle; C1 and c2 represents speedup factor, r1 and r2 represents the random number be evenly distributed in interval (0,1); ω _maxfor the maximum magnitude of ω, ω _minfor the minimum value scope of ω, t is current iteration algebraically, and PM is population greatest iteration algebraically.

10. the cloud computing task scheduling system based on quanta particle swarm optimization according to claim 9, described best position calculation module, also for judging that iterations does not arrive default iterations when judge module, calculate the adaptive value of all particles, determine optimal location and colony's optimal location of all particles.