CN105843689A - Virtual machine deployment method and system - Google Patents
Virtual machine deployment method and system Download PDFInfo
- Publication number
- CN105843689A CN105843689A CN201610228790.8A CN201610228790A CN105843689A CN 105843689 A CN105843689 A CN 105843689A CN 201610228790 A CN201610228790 A CN 201610228790A CN 105843689 A CN105843689 A CN 105843689A
- Authority
- CN
- China
- Prior art keywords
- sigma
- population
- physical machine
- individual
- virtual machine
- 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
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5061—Partitioning or combining of resources
- G06F9/5077—Logical partitioning of resources; Management or configuration of virtualized resources
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2209/00—Indexing scheme relating to G06F9/00
- G06F2209/50—Indexing scheme relating to G06F9/50
- G06F2209/503—Resource availability
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a virtual machine deployment method and system. The method comprises the steps that an initial solution of a default model is acquired and is used for expressing the mapping relations of multiple physical machines and multiple virtual machines; a default algorithm is used for operation on the initial solution, and the optimal solution of the default model under the default algorithm is obtained; virtual machine deployment is performed according to the obtained optimal solution, wherein parameters of the default model comprise capacities of CPUs, memories and hard disks of the virtual machines. According to the virtual machine deployment method, the initial solution is obtained through the default model, an initial-solution-based scheme about virtual machine deployment on the physical machines is optimized with the default algorithm, load balancing coordination of the physical machines is realized, and the utilization rate of the physical machines is increased. According to the default model selected for the virtual machine deployment method and system, factors including the capacities of CPUs, the memories and the hard disks of the virtual machines are taken into consideration comprehensively, load balancing coordination of the physical machines can be better realized, and the utilization rate of the physical machines is increased.
Description
Technical field
The present invention relates to virtual machine Optimization deployment technical field, particularly relate to the method and system of a kind of deploying virtual machine.
Background technology
Along with the development of internet, Intel Virtualization Technology can on separate unit physical server virtual dissolve multiple separate
Virtual machine (VM, VirtualMachine), current Intel Virtualization Technology has been widely used in data center at different levels, has especially taken
Business device Intel Virtualization Technology is accepted and successful implementation by users especially.
According to the feature of user's request, the description to virtual machine performance generally considers the factors such as CPU, internal memory and hard disk.Example
As, as it is shown in figure 1, during deploying virtual machine, virtual machine VM4, VM5 have been deployed on physical machine HOST1 and HOST2,
During to continue to be deployed in these two physical machine by other three virtual machines VM1, VM2, VM3, it is necessary for considering physics simultaneously
Machine and the self performance of virtual machine, not only make to keep between different physical services resource load balancing, also needs to consider simultaneously
Load balancing between Same Physical Service Source different resource dimension.The feasible scheme of one of which is to be deployed in by VM3
HOST1, VM1, VM2 are deployed in HOST2.N platform virtual machine is mapped in M platform physical services resource, total kind of portion possible for MN
Management side case.Meanwhile, this problem is again multi-objective optimization question, need consider the Homes Using TV of physical machine, physical machine utilization rate and
The most how the non-load balanced case of physical machine, optimize virtual machine deployment scheme in physical machine, improves the utilization of physical machine
Rate becomes the problem needing solution badly.
Summary of the invention
For defect of the prior art, the invention provides the method and system of a kind of deploying virtual machine, by void
Plan machine deployment scheme in physical machine is optimized, and improves the utilization rate of physical machine.
First aspect, a kind of method that the invention provides deploying virtual machine, including: obtain the initial solution of preset model,
Described initial solution is for representing the mapping relations of multiple physical machine and multiple virtual machines;
Use preset algorithm that described initial solution is carried out computing, obtain described preset model under described preset algorithm
Excellent solution;
The deployment of virtual machine is carried out according to the optimal solution obtained;
Wherein, the parameter in described preset model includes the CPU of virtual machine, internal memory and hard disk size.
Second aspect, the invention provides the system of a kind of deploying virtual machine, including:
Initial solution acquisition module, for obtaining the initial solution of preset model, described initial solution is used for representing multiple physical machine
Mapping relations with multiple virtual machines;
Optimal solution acquisition module, is used for using preset algorithm that described initial solution is carried out computing, obtains described preset model
Optimal solution under described preset algorithm;
Deployment module, for carrying out the deployment of virtual machine according to the optimal solution obtained;
Wherein, the parameter in described preset model includes the CPU of virtual machine, internal memory and hard disk size.
As shown from the above technical solution, in the method and system of the deploying virtual machine that the present invention provides, selected is pre-
If model has considered the factors such as the CPU of virtual machine, internal memory and hard disk size, it is possible to preferably realize to physical machine load all
The coordination of weighing apparatus, improves the utilization rate of physical machine.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
In having technology to describe, the required accompanying drawing used is briefly described, it should be apparent that, the accompanying drawing in describing below is only this
Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, it is also possible to
Other accompanying drawing is obtained according to these figures.
Fig. 1 is that in prior art, a kind of virtual machine disposes process schematic in physical machine;
The schematic flow sheet of the method for a kind of deploying virtual machine that Fig. 2 provides for one embodiment of the invention;
The schematic diagram of the chromosome coding that Fig. 3 provides for one embodiment of the invention;
The process schematic that population at individual is performed intersection operation that Fig. 4 provides for one embodiment of the invention;
The process schematic that population at individual is performed mutation operation that Fig. 5 provides for one embodiment of the invention;
Fig. 6 obtains the utilization rate of physical machine and the reality obtaining number of times for the employing BLGA algorithm that one embodiment of the invention provides
Test result schematic diagram;
Fig. 7 obtains physical machine load balancing variance for the employing BLGA algorithm that one embodiment of the invention provides and obtains number of times
Experimental result schematic diagram;
The employing BLGA algorithm that Fig. 8 provides for one embodiment of the invention obtains physical machine with other algorithms of the prior art
The contrast schematic diagram of utilization rate;
The employing BLGA algorithm that Fig. 9 provides for one embodiment of the invention obtains physical machine with other algorithms of the prior art
The contrast schematic diagram of load balancing variance;
The structural representation of the system of a kind of deploying virtual machine that Figure 10 provides for one embodiment of the invention.
Detailed description of the invention
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Describe, it is clear that described embodiment is only a part of embodiment of the present invention rather than whole embodiments wholely.Based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under not making creative work premise
Embodiment, broadly falls into the scope of protection of the invention.
Deploying virtual machine problem in the present embodiment can be further depicted as, by Virtual Service money separate for N platform
Source VM={vm1,vm2,…,vmNIt is deployed in M platform physical machine Datecenter={host1,host2,…,hostMComposition data
Center, wants to improve data center resource utilization rate in deployment cycle T, ensures systematic entirety energy simultaneously.
Optimize before deploying virtual machine illustrating how, first the computational methods to some parameters in physical machine carry out as
Give a definition and illustrate.
Definition 1, the load of physical machine CPU.In physical machine Hj, virtual machine quantity is m, then the load of physical machine CPU is this thing
On reason machine Hj, all virtual machines use the ratio of CPU amount and the CPU total amount of this physical machine Hj.
Wherein, HC represents the CPU amount of physical machine Hj, and VCi represents the CPU usage amount of virtual machine.
Definition 2, the load of physical machine internal memory.In physical machine Hj, virtual machine quantity is m, then the load of physical machine internal memory is for being somebody's turn to do
In physical machine Hj, all virtual machines use the ratio of amount of ram and the memory amount of this physical machine Hj.
Wherein, HM represents the amount of ram of physical machine Hj, and VMi represents the internal memory usage amount of virtual machine.
Definition 3, the load of physical machine hard disk.In physical machine Hj, virtual machine quantity is m, then the load of physical machine hard disk is for being somebody's turn to do
In physical machine Hj, all virtual machines use the ratio of hard disk amount and the hard disk total amount of this physical machine Hj.
Wherein, HD represents the hard disk amount of physical machine Hj, and VDi represents the hard disk usage amount of virtual machine.
Definition 4, the load of physical machine Hj, the i.e. utilization rate of single one physical machine.The load of physical machine Hj be then corresponding CPU,
The load weighting sum of internal memory and hard disk.
HL=ω1HLC+ω2HLD+ω3HLM (4.4)
Wherein, ω1, ω2, ω3Representing different resource load weight, different weights represents that user is when using virtual machine
Preference profile to different performance.
Definition 5, the resource utilization of physical machine utilization rate, i.e. data center.There is M platform physical machine HOST and N platform is empty
Plan machine VM, then physical machine utilization rate is the usage amount total amount of all of virtual machine CPU, internal memory and hard disk and all openings
Physical machine CPU, internal memory and hard disk total amount ratio weighting sum, its value actual equal to all physical machine load average
Value.
Wherein, HTjRepresent whether physical machine Hj is in opening, ω1, ω2, ω3Represent different resource load weight,
Different weights represent user's request emphasis need not, meanwhile, physical machine utilization rate comprises opens physical machine minimum number
Target.
Definition 6, the load variance of physical machine.It is expressed as the physical machine load dispersion degree relative to average load, specifically
It is expressed as:
Wherein, AHL represents the average load of physical machine.
If S is the deployment scheme of virtual machine, then deploying virtual machine problem can be converted into a multi-objective optimization question.I.e.
The utilization rate MAX HAU (S) maximizing physical machine minimizes load variance MIN σ (S) of physical machine simultaneously.Meanwhile, virtual machine and
Physical machine also needs to retrain below satisfied in performance.
xij∈ 0,1}, i=1,2,3 ..., N;J=1,2,3 ..., M (4.8)
Wherein, N represents the virtual machine quantity that needs are disposed;M represents the physical machine quantity of data center;VCi、VMi、VDiPoint
Biao Shi the CPU of virtual machine i, internal memory, hard disk size;HCj、HMj、HDjRepresent that virtual machine j has CPU, internal memory, hard disk respectively big
Little.In formula (4.8), xijFor binary number, xijIt is 1 when being, represents that virtual machine i is deployed in physical machine j, otherwise, xijFor
0;Formula (4.9)~(4.11) represent that a certain class resource sum of the virtual machine demand being deployed in physical machine is necessarily less than respectively
In this physical machine total amount;Formula (4.12) represents that unique constraints, i.e. any one virtual machine must be deployed in a physical machine.
For the feature of the problem in above-mentioned background technology, virtual machine deployment scheme in physical machine can be considered one three
Tie up variable-sized bin packing.But N platform virtual machine is mapped in M platform physical services resource, total kind of deployment side possible for MN
Case.Meanwhile, this problem is again multi-objective optimization question, needs to consider the Homes Using TV of physical machine, physical machine utilization rate and physics
The non-load balanced case of machine.To this, following embodiment proposes a kind of Revised genetic algorithum and solves the problems referred to above.
The schematic flow sheet of the method for a kind of deploying virtual machine that Fig. 2 provides for one embodiment of the invention, it is adaptable to void
Plan machine deployment scheme in physical machine is optimized, as in figure 2 it is shown, the method comprises the following steps:
201, obtaining the initial solution of preset model, described initial solution is for representing reflecting of multiple physical machine and multiple virtual machines
Penetrate relation;Wherein, the parameter in described preset model includes the CPU of virtual machine, internal memory and hard disk size.
Described preset model includes object function and constraints, and wherein, object function is:
The object function of preset model is:
Constraints is specifically as follows:
xij∈ 0,1}, i=1,2,3 ..., N;J=1,2,3 ..., M;
Wherein, σ represents the load variance of physical machine, and HAU represents the utilization rate of physical machine, HTjRepresent whether physical machine is located
In opening, HLjRepresent the load of physical machine, ω1, ω2, ω3Representing load weight, AHL represents the average negative of physical machine
Carry;N represents virtual machine quantity;M represents physical machine quantity;VCi、VMi、VDiRepresent that the CPU of virtual machine i, internal memory, hard disk are big respectively
Little;HCj、HMj、HDjRepresent that physical machine j has CPU, internal memory, hard disk size respectively;xijFor binary number, xijIt is 1 when being, table
Show that virtual machine i is deployed in physical machine j, otherwise, xijIt is 0.
202, use preset algorithm that described initial solution is carried out computing, obtain described preset model under described preset algorithm
Optimal solution.
In embodiments of the present invention, HAU is benefit function, and σ is cost-effectivenes function, in step 202., uses pre-imputation
It is the constraints obtaining and meeting this above-mentioned preset model and HAU maximum, σ minimum that method carries out the purpose of computing to initial solution
Solve.
Specifically, in the specific implementation, can be using described initial solution as population, described population includes multiple population
Body, each population at individual represents the mapping relations of virtual machine and physical machine;And by genetic algorithm, initial solution is carried out budget, this
Time, above-mentioned step 202 can specifically include:
By genetic algorithm, the population at individual in described population carried out evolution iteration, obtain the population at individual after evolving with
And include the population of the population at individual after described evolution.
Said method by preset model obtain initial solution, by preset algorithm to virtual machine in initial solution in physical machine
Deployment scheme be optimized, it is achieved that the coordination to physical machine load balancing, improve the utilization rate of physical machine.
As a kind of optional embodiment of above-mentioned genetic algorithm, in iteration mistake each time in above-mentioned genetic algorithm
Cheng Zhong, selects to need the individual step carrying out genetic manipulation to may include that entered to described initial solution by preset algorithm
Before row operation, described method also includes:
In iterative process each time, according to the utilization rate of the physical machine of population at individual each in population and physical machine
Load variance obtains the fitness of population at individual;According to the fitness of described population at individual, use roulette method described population
In choose treat heredity population at individual.
Specifically, the process of the fitness obtaining population at individual here can be particularly as follows: obtain every according to equation below
The utilization rate of the physical machine of one population at individual and the fitness of the load variance acquisition population at individual of physical machine:
Wherein, Fitness represents the fitness of population at individual, λ1, λ2Represent weighted factor, and λ1> 0, λ2> 0.
According to the fitness of described population at individual, roulette method is used to choose the population treating evolution iteration in described population
Individual.
203, the deployment of virtual machine is carried out according to the optimal solution obtained.
Understandable, owing to initial solution represents the mapping relations of virtual machine and physical machine, optimal solution the most here is also
It is the mapping relations representing virtual machine with physical machine.Arrange virtual according to virtual machine in physical machine with the mapping relations of physical machine
The process of machine is referred to prior art, is no longer described in detail in the embodiment of the present invention.
Below by specific embodiment, the process obtaining optimal solution in said method is described in detail, including following
Step:
301, initializing population, the model set up by above-mentioned steps 201 obtains the initial solution of population, the kind so obtained
Group is all the population meeting default constraints, according to utilization rate HAU of multiple physical machine in group's individuality each of in population
And the load variances sigma of physical machine obtains the fitness of each described population at individual in initial population, according to each described population
Individual fitness, obtains the fitness Fitness of described initial population, and wherein fitness function is as follows:
Wherein, HAU is referred to formula (4.5), and σ is referred to formula (4.6) and draws, λ1, λ2Represent adding of corresponding factor of evaluation
Weight factor, and λ1> 0, λ2> 0.
Population can be understood as including the corresponding pass that multiple population at individual, each population at individual include virtual machine and physical machine
System.
Concrete, before performing above-mentioned steps 301, need to perform the following step not shown in Fig. 3:
300A, first physical machine and virtual machine are encoded, it may be assumed that
(1) to the physical machine coding opened, different physical machine is distinguished;
(2) which deploying virtual machine clear and definite is in which physical machine.
Meanwhile, Solve problems requires to minimize the quantity opening physical machine, the therefore chromosome coding gene of Solve problems
Be length be uncertain.
Fig. 3 gives a kind of virtual machine deployment way in physical machine.Wherein, virtual machine VM3, VM5, VM6 is deployed in
Physical machine HOST1, virtual machine VM1 is deployed in physical machine HOST2;Virtual machine VM2, VM7 are deployed in physical machine HOST3, virtual machine
VM4 is deployed in physical machine HOST4.Wherein physical machine and virtual machine are the relation of one-to-many.
300B, solve the intraindividual initial solution of each population.
Owing to this problem needs to consider utilization rate and the physical machine load-balancing performance of physical machine, for initial population
Quantity, the initial solution generating algorithm step of deploying virtual machine is as follows:
Step 1 is according to formula (4.4), the load of Computational Physics machine current state;
Step 2 calculates the stock number of each virtual machine and accounts for the proportion p of whole virtual machine;
Step 3 with Probability p by deploying virtual machine to the physical machine having turned on so that the load of physical machine is minimum;If
There is not the physical machine meeting condition, then reopen a physical machine, update the load of physical machine simultaneously.
Wherein in step 301, fitness function fitness function is the result appraisal standard in genetic algorithm, and fitness is more
Greatly, the effect of solution is the best.The evaluation criterion of tradition bin packing mainly uses chest quantity few, ensures chest utilization rate simultaneously
High.But, under IaaS environment, deploying virtual machine problem is different from traditional bin packing, and physical machine to be considered opens quantity
And the utilization rate of physical resource, it is also contemplated that physical machine loads the impact on physical machine performance.The load variance of physical machine
The least stability showing physical machine and robustness are the best.Therefore, amid all these factors, design evaluatio fitness function is such as
Under:
Wherein, λ1, λ2Represent the weighted factor of corresponding factor of evaluation, and λ1> 0, λ2> 0.
302, described initial population fitness less than or equal to preset fitness time, use roulette method described initially
In population, selected population is individual, the digital r between stochastic generation [0,1] simultaneously.
303, judge that described r whether less than or equal to presetting crossover probability, the most then performs step 304 or performs step
307, otherwise perform step 301.
304, at described r less than or equal to when presetting crossover probability, the population at individual selected is performed intersection operation, and is formed
New population at individual, it is thus achieved that include the first optimization population of new population at individual.
Perform to intersect to the population at individual selected described in above-mentioned steps 304 and operate, including the following son not shown in Fig. 3
Step:
3041, in initial population, two population at individuals of wheel disc algorithms selection are used, as shown in Figure 4;
3042, in the first population at individual A, select virtual machine, and select in the second population at individual B and the first population
The identical virtual machine of the mark of virtual machine selected in body is as hybridization virtual machine VM1 and VM4;
3043, use the principle sought common ground while reserving difference, retain the physical machine in the first population at individual A and the second population at individual B altogether
With virtual machine VM3, VM6 and VM7 of mapping, and remaining virtual machine is deleted, obtain the individual C of to be confirmed the third group;
3044, greed principle is used to be inserted into Same Physical machine by the virtual machine of deletion hybridizes virtual machine VM1 and VM4
In HOST2, and removal of impurities in the virtual machine of deletion is handed over virtual machine VM2 and VM5 outside virtual machine be inserted into other physical machine
In HOST1 and HOST3, obtain the third crowd of individual D after optimizing.
305, calculate the described first fitness optimizing population, optimize the fitness of population less than or equal to pre-described first
If during fitness, optimize the deployment in physical machine of the virtual machine of population at individual in population to described first by mutation operation
Scheme is optimized.
306, judge whether the described first fitness optimizing population is less than or equal to preset fitness, the most then perform step
Rapid 307, otherwise, perform step 301.
307, the fitness in described initial population optimizes the suitable of population less than or equal to presetting fitness or described first
Response, less than or equal to when presetting fitness, optimizes selected population in population in described initial population or described first individual, simultaneously
Digital q between stochastic generation [0,1].
308, judge that whether described q is less than or equal to presetting genetic probability;The most then perform step 309, otherwise perform step
301。
309, at described q less than or equal to when presetting genetic probability, the population at individual selected is performed mutation operation.
Described in above-mentioned steps 309, the population at individual selected is performed mutation operation, including the following son not shown in Fig. 3
Step:
3091, selection the 4th population at individual E in population is optimized in described initial population or first, as shown in Figure 5;
3092, on described 4th population at individual E, randomly choose two physical machine HOST2 and HOST3, and selecting often
Individual physical machine selects virtual machine VM1 and VM2;
3093, virtual machine VM1 and VM2 selected in said two physical machine HOST2 and HOST3 is swapped, obtain
The 4th population at individual F after optimization.
310, obtain and perform the fitness of the population at individual after mutation operation, and after comprising described execution mutation operation
The fitness of the second optimization population of population at individual, until the described second fitness optimizing population more than default fitness is
Only.
In above-mentioned steps 310, the most also include presetting iterations, namely perform the number of times of mutation operation, certainly
According to follow-up experimental result it can be seen that different number of run has different results, before performing above-mentioned steps, can
To preset as required.
Concrete, can use roulette method in above-mentioned steps after selected population individuality, population at individual is performed
Intersect and operate, then judge fitness;Or population at individual is performed mutation operation, then judges fitness;Or first carry out intersection
Performing mutation operation after operation, finally judge fitness, aforesaid way is not selected by the present embodiment, it is preferable that use
Performing mutation operation after first carrying out intersection operation, in the population so obtained, the utilization rate of physical machine can be higher.
Combining and above-mentioned steps, the selection to concrete genetic operator is described in detail again.
(1) selection opertor
This algorithm uses classical roulette selection method, mainly according to the ratio of each population at individual Yu whole population
Size determines the select probability of this population at individual.Shown in comprising the following steps that:
Step 1 calculates the fitness of population at individual according to formula (4.13);
The fitness that step 2 is obtained according to step 1, calculates each ideal adaptation ratio;
Assuming there are four individualities in population, individual a, the fitness function value of b, c, d is respectively 1,2,3,4, then population
Fitness function value and be 1+2+3+4=10, then individual a enters follow-on probability, the most selected probability, just
It is the fitness value ratio divided by total fitness sum of oneself, namely 1/10=10%;In like manner, the probability of b, c, d is
20%, 30%, 40%.Select according to the probability so obtained.
Step 3 utilizes roulette method to select individuality.
Using roulette method, population at individual fitness is the biggest, then to be chosen possibility the biggest for this population at individual, it is ensured that plant
The individual preservation that group's fitness is bigger is gone down so that solving result has preferable global convergence, meanwhile, there is also population
Body fitness the most selected little possibility, this is avoided solving result to sink into locally optimal solution very well.
(2) crossover operator
Crossover operator is a core operator in genetic algorithm, and the performance of genetic algorithm is largely by crossover operator
Determining, crossover operator selects to must take into two principles: principle one: after intersection, new explanation must be a feasible solution;Principle two:
Search volume can be increased after intersection.Comprise the following steps that, as shown in Figure 4:
Step 1 population at individual selects, and uses round robin algorithm to randomly choose two population at individuals T1, T2;
Step 2 virtual machine selects, and randomly chooses the virtual machine under physical machine as hybridization portion;
Step 3 deletes virtual machine, uses principle of seeking common ground while reserving differences, i.e. retains physical machine co-map in population at individual T1, T2
Virtual machine, the demapping section that deletion of physically machine is different;
Step 4 inserts virtual machine, uses greed principle that the virtual machine of deletion is reinserted into physical machine.
(3) mutation operator
Mutation operator is an Important Operators in genetic algorithm, and the introducing of mutation operator maintains the diversity of population,
This operator mainly adjusts the gene position of population at individual.Comprise the following steps that, as shown in Figure 5:
Step 1 sets mutation probability Pm, randomly chooses a population at individual according to mutation probability;
Step 2 randomly chooses two physical machine on the population at individual chosen;
The virtual machine that step 3 randomly chooses in two physical machine respectively swaps;
If after step 4 end step 3, this population at individual is feasible solution, i.e. meets and presets constraints (model), then
This mutation process terminates;Otherwise re-start mutation operation.
The solution procedure of above-mentioned genetic algorithm (BLGC algorithm) is summarized as follows:
Step 1 initial population, sets relevant parameter, general including maximum iteration time Gmax, crossover probability Pc and heredity
Rate Pm;
Step 2 calculates the fitness of population at individual according to formula (4.13);
Step 3 judges whether to meet end condition, if meeting condition, then exports optimal solution, and algorithm terminates.Otherwise enter
Step 4;
Step 4 utilizes roulette method, and the population at individual in selected population enters the next generation;
Digital r between step 5 stochastic generation one [0,1], it is judged that whether r < Pc meets.If met, intersect
Operation, forms new population at individual.Otherwise forward step 6 to;
Digital q between step 6 stochastic generation one [0,1], it is judged that whether q < Pm meets, if it is satisfied, then become
ETTHER-OR operation.Forward step 2 to.
In order to illustrate to use the effect of optimization of embodiment of the present invention scheme, illustrate below by experimental data.
In experimentation, save as 3G in main frame, be 3.20GHz for 650G, CPU, and utilize under Windows XP
My Eelipse8.5 and jdk1.6.0_10 is carried out, and uses java programming realization.
First provide relevant experiment parameter, the data center of 12 physical machine compositions of Setup Experiments, there are 4 types
Physical machine, each 3 of each type physical machine.Four kinds of physical machine type configuration are as shown in table 1:
Table 1
In an experiment, 25 virtual machines are set altogether, owing to virtual machine size is determined by user completely, so dissimilar use
Family is different to virtual machine demand.Some users belong to computation-intensive user, relatively big to the demand of CPU and internal memory, have
User belong to storage intensive user, bigger to hard disk demand.Therefore, in experiment the design parameter of virtual machine at certain model
Enclose interior stochastic generation, to prevent the contingency of experimental result.
Virtual machine parameter area is set in an experiment: CPU (50HZ~500HZ);Internal memory (384M~1024M);Hard disk
(50G~200G).Real needs based on these 25 virtual machines of stochastic generation are as shown in table 2:
Table 2
In this experiment, setting this algorithm greatest iteration test Gmax=200, population scale is 20, probability of crossover Pc=
0.8, mutation probability Pm=0.15, λ1=100, λ2=1.Wherein, ω1, ω2, ω3It is 1/3, i.e. thinks CPU, internal memory and hard
Dish is of equal importance.Taking many experiments to take BLGA algorithm, the utilization rate of physical machine and load balancing variance experimental result are such as
Shown in Fig. 6.
Result is run repeatedly it can be seen that the fluctuation range of physical machine utilization rate is at [0.796-0.874] from BLGA algorithm,
Although the utilization rate of physical machine exists situation about fluctuating up and down, substantially remaining in more than 0.8, physical machine utilization rate situation is preferable,
This is to improve the service utilization of resources as target mainly due to BLGA algorithm, is intersected by genetic algorithm, mutation operator selects
Selecting the more preferable chromosome of fitness, so physical machine can reach more preferable resource utilization, this test result indicate that BLGA pair
It is a kind of key tactics in solving deploying virtual machine problem.Corresponding load balancing variance has similar result, such as Fig. 7 institute
Show.
In order to verify that BLGA algorithm can preferably solve Service Source portion based on load balancing relative to other algorithms
Administration's problem, it is thus achieved that more excellent satisfactory solution, illustrates that this algorithm has validity, uses the matching algorithm that most preferably successively decreases (BFD) herein, adds
Power Smallest connection algorithm (WLC) and method based on distance coupling compare analysis, the utilization rate of physical machine and physical machine
Load variance experimental result as shown in Figure 8 and Figure 9:
It can be seen that BLGA strategy physical machine utilization rate apparently higher than WLC strategy, BFD strategy and to
Flux matched method, it is considered to the WLC strategy of load balancing is higher than BFD strategy and Vectors matching strategy.Mainly due to BLGA strategy it is
A kind of intelligent optimization algorithm, has bigger search volume, can carry out virtual machine coupling from the angle of the overall situation;Meanwhile,
BLGA strategy considers the problem of load balancing that physical machine is disposed, and is coordinated by physical machine load, improves physical machine on the whole
Utilization rate, and on the basis of different physical machine loads are coordinated, the load coordinating Same Physical machine different dimensions resource is equal
Weighing apparatus, improves the utilization rate of physical machine to greatest extent, it is to avoid the phenomenon of resource fragmentation.
It can be seen that the load variance of the physical machine of BLGA strategy is significantly lower than BFD strategy and Vectors matching
Method, is slightly better than WLC strategy.It is primarily due to BLGA strategy and improves physical machine utilization rate, focus on collaborative different physical resource
Load capacity, it is considered to the non-load balanced case of physical machine, therefore, the non-load balanced case of BLGA strategy is substantially better than BFD strategy
And Vectors matching method.Although WLC strategy also focuses on the non-load balanced case of physical machine, but search volume is calculated much smaller than BLGA
Method, so BLGA strategy is better than WLC strategy.
The embodiment of the present invention additionally provides the structural representation of the system of a kind of deploying virtual machine, it is adaptable to exist virtual machine
Deployment scheme in physical machine is optimized, and as shown in Figure 10, this system includes:
Initial solution acquisition module 10, for obtaining the initial solution of preset model, described initial solution is used for representing multiple physics
Machine and the mapping relations of multiple virtual machines;
Optimal solution acquisition module 11, is used for using preset algorithm that described initial solution is carried out computing, obtains described default mould
Type optimal solution under described preset algorithm;
Deployment module 12, for carrying out the deployment of virtual machine according to the optimal solution obtained;
Wherein, the parameter in described preset model includes the CPU of virtual machine, internal memory and hard disk size.
One of the present embodiment preferred embodiment in, the object function of described preset model is:
Wherein,
The constraints of described preset model is:
xij∈ 0,1}, i=1,2,3 ..., N;J=1,2,3 ..., M;
Wherein, σ represents the load variance of physical machine, and HAU represents the utilization rate of physical machine, HTjRepresent whether physical machine is located
In opening, HLjRepresent the load of physical machine, ω1, ω2, ω3Representing load weight, AHL represents the average negative of physical machine
Carry;N represents virtual machine quantity;M represents physical machine quantity;VCi、VMi、VDiRepresent that the CPU of virtual machine i, internal memory, hard disk are big respectively
Little;HCj、HMj、HDjRepresent that physical machine j has CPU, internal memory, hard disk size respectively;xijFor binary number, xijIt is 1 when being, table
Show that virtual machine i is deployed in physical machine j, otherwise, xijIt is 0.
Optimal solution acquisition module 11, specifically for using preset algorithm that described initial solution is carried out computing, is met institute
State constraints and make the solution that HAU is maximum and σ is minimum.
One of the present embodiment preferred embodiment in, using described initial solution as population, described population includes many
Individual population at individual, each population at individual represents the mapping relations of virtual machine and physical machine;
Accordingly, described optimal solution acquisition module, specifically for:
By genetic algorithm, the population at individual in described population carried out evolution iteration, obtain the population at individual after evolving with
And include the population of the population at individual after described evolution.
One of the present embodiment preferred embodiment in, described optimal solution acquisition module 11 includes:
Fitness acquisition module, is used in iterative process each time, according to the physical machine of population at individual each in population
Utilization rate and physical machine load variance obtain population at individual fitness;
Treat Advanced group species individuality acquisition module, for the fitness according to described population at individual, use roulette method in institute
State and population is chosen the population at individual treating heredity.
One of the present embodiment preferred embodiment in, described fitness acquisition module, specifically for:
Wherein,
Wherein, Fitness represents the fitness of population at individual, and σ represents the load variance of physical machine, and HAU represents physical machine
Utilization rate, λ1, λ2Represent weighted factor, and λ1> 0, λ2> 0, HTjRepresent whether physical machine is in opening, HLjRepresent
The load of physical machine, ω1, ω2, ω3Representing load weight, AHL represents the average load of physical machine, and N represents virtual machine quantity;M
Represent physical machine quantity;VCi、VMi、VDiRepresent the CPU of virtual machine i, internal memory, hard disk size respectively;HCj、HMj、HDjTable respectively
Show that physical machine j has CPU, internal memory, hard disk size.
One of the present embodiment preferred embodiment in, described optimal solution acquisition module, be used for:
Roulette method is used to select population at individual to be evolved in described population, simultaneously between stochastic generation [0,1]
Numeral r, it is judged that whether described r is less than or equal to presetting crossover probability;
At described r less than or equal to when presetting crossover probability, the population at individual treating evolution performs intersection operation, and is formed new
Population at individual;
Digital q between stochastic generation [0,1], it is judged that whether described q is less than or equal to presetting mutation probability;
At described q less than or equal to when presetting mutation probability, in the population including described new population at individual, select to wait to become
Different population at individual;
Described population at individual to be made a variation is performed mutation operation, it is judged that include the population at individual after performing mutation operation
Whether population meets is preset constraints;
When this population meets and presets constraints, stop performing mutation operation.
It should be noted that system in the present embodiment and said method are relations one to one, in said method
Implementation detail is equally applicable to this system, and the present embodiment is not embodied as details to system and is described in detail.
In the specification of the present invention, illustrate a large amount of detail.It is to be appreciated, however, that embodiments of the invention are permissible
Put into practice in the case of there is no these details.In some instances, it is not shown specifically known method, structure and skill
Art, in order to do not obscure the understanding of this description.
Similarly, it will be appreciated that disclose to simplify the present invention and help understand in each inventive aspect one or many
Individual, above in the description of the exemplary embodiment of the present invention, each feature of the present invention is grouped together into single sometimes
In embodiment, figure or descriptions thereof.But, the method for the disclosure should not explained and i.e. be wanted in reflecting an intention that
Seek the application claims feature more more than the feature being expressly recited in each claim of protection.More precisely, such as
As claims below is reflected, inventive aspect is all features less than single embodiment disclosed above.
Therefore, it then follows claims of detailed description of the invention are thus expressly incorporated in this detailed description of the invention, the most each right is wanted
Ask itself all as the independent embodiment of the present invention.
It will be understood by those skilled in the art that and the module in the equipment in embodiment adaptively can be changed
And they are provided in one or more equipment that this embodiment is different.Can the module in embodiment or unit or
Assembly is combined into a module or unit or assembly, and can put them into multiple submodule or subelement or subgroup in addition
Part.Except at least some in such feature and/or process or unit is mutually exclusive part, any combination can be used
To all features disclosed in this specification (including adjoint claim, summary and accompanying drawing) and the disclosedest any side
Method or all processes of equipment or unit are combined.Unless expressly stated otherwise, this specification (includes that adjoint right is wanted
Ask, make a summary and accompanying drawing) disclosed in each feature can be replaced by providing identical, equivalent or the alternative features of similar purpose.
Although additionally, it will be appreciated by those of skill in the art that embodiments more described herein include other embodiments
Some feature included by rather than further feature, but the combination of the feature of different embodiment means to be in the present invention's
Within the scope of and form different embodiments.Such as, in the following claims, embodiment required for protection appoint
One of meaning can mode use in any combination.
The all parts embodiment of the present invention can realize with hardware, or to run on one or more processor
Software module realize, or with combinations thereof realize.It will be understood by those of skill in the art that and can use in practice
Microprocessor or digital signal processor (DSP) realize in the equipment of a kind of browser terminal according to embodiments of the present invention
The some or all functions of some or all parts.The present invention is also implemented as performing side as described herein
Part or all equipment of method or device program (such as, computer program and computer program).Such
The program realizing the present invention can store on a computer-readable medium, or can have the shape of one or more signal
Formula.Such signal can be downloaded from internet website and obtain, or provides on carrier signal, or with any other shape
Formula provides.
The present invention will be described rather than limits the invention to it should be noted above-described embodiment, and ability
Field technique personnel can design alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference symbol that should not will be located between bracket is configured to limitations on claims.Word " comprises " and does not excludes the presence of not
Arrange element in the claims or step.Word "a" or "an" before being positioned at element does not excludes the presence of multiple such
Element.The present invention and can come real by means of including the hardware of some different elements by means of properly programmed computer
Existing.If in the unit claim listing equipment for drying, several in these devices can be by same hardware branch
Specifically embody.Word first, second and third use do not indicate that any order.These word explanations can be run after fame
Claim.
Last it is noted that various embodiments above is only in order to illustrate technical scheme, it is not intended to limit;To the greatest extent
The present invention has been described in detail by pipe with reference to foregoing embodiments, it will be understood by those within the art that: it depends on
So the technical scheme described in foregoing embodiments can be modified, or the most some or all of technical characteristic is entered
Row equivalent;And these amendments or replacement, do not make the essence of appropriate technical solution depart from various embodiments of the present invention technology
The scope of scheme, it all should be contained in the middle of the claim of the present invention and the scope of specification.
Claims (10)
1. the method for a deploying virtual machine, it is characterised in that including:
Obtaining the initial solution of preset model, described initial solution is for representing the mapping relations of multiple physical machine and multiple virtual machines;
Use preset algorithm that described initial solution is carried out computing, obtain described preset model optimum under described preset algorithm
Solve;
The deployment of virtual machine is carried out according to the optimal solution obtained;
Wherein, the parameter in described preset model includes the CPU of virtual machine, internal memory and hard disk size.
Method the most according to claim 1, it is characterised in that the object function of described preset model is:
Wherein,
The constraints of described preset model is:
xij∈ 0,1}, i=1,2,3 ..., N;J=1,2,3 ..., M;
Wherein, σ represents the load variance of physical machine, and HAU represents the utilization rate of physical machine, HTjRepresent whether physical machine is in unlatching
State, HLjRepresent the load of physical machine, ω1, ω2, ω3Representing load weight, AHL represents the average load of physical machine;N represents
Virtual machine quantity;M represents physical machine quantity;VCi、VMi、VDiRepresent the CPU of virtual machine i, internal memory, hard disk size respectively;HCj、
HMj、HDjRepresent that physical machine j has CPU, internal memory, hard disk size respectively;xijFor binary number, xijIt is 1 when being, represents virtual machine
I is deployed in physical machine j, otherwise, and xijIt is 0;
Described use preset algorithm carries out computing to described initial solution, obtains the optimal solution of described preset model, including:
Use preset algorithm that described initial solution is carried out computing, be met described constraints and make HAU maximum and σ minimum
Solution.
Method the most according to claim 1, it is characterised in that using described initial solution as population, described population includes many
Individual population at individual, each population at individual represents the mapping relations of virtual machine and physical machine;
Accordingly, described use preset algorithm carries out computing to described initial solution, obtains described preset model in described pre-imputation
Optimal solution under method, including:
By genetic algorithm, the population at individual in described population is carried out evolution iteration, obtain the population at individual after evolving and bag
Include the population of the population at individual after described evolution.
Method the most according to claim 3, it is characterised in that the population at individual in described population is entered by genetic algorithm
Travelingization iteration, including:
In iterative process each time, according to utilization rate and the load of physical machine of the physical machine of population at individual each in population
Variance obtains the fitness of population at individual;According to the fitness of described population at individual, roulette method is used to select in described population
Take the population at individual treating heredity.
Method the most according to claim 4, it is characterised in that the described physical machine according to population at individual each in population
The load variance of utilization rate and physical machine obtains the fitness of population at individual, including: obtain each population according to equation below
The utilization rate of individual physical machine and the fitness of the load variance acquisition population at individual of physical machine:
Wherein,
Wherein, Fitness represents the fitness of population at individual, and σ represents the load variance of physical machine, and HAU represents the profit of physical machine
With rate, λ1, λ2Represent weighted factor, and λ1> 0, λ2> 0, HTjRepresent whether physical machine is in opening, HLjRepresent physics
The load of machine, ω1, ω2, ω3Representing load weight, AHL represents the average load of physical machine, and N represents virtual machine quantity;M represents
Physical machine quantity;VCi、VMi、VDiRepresent the CPU of virtual machine i, internal memory, hard disk size respectively;HCj、HMj、HDjRepresent thing respectively
Reason machine j has CPU, internal memory, hard disk size.
6. the system of a deploying virtual machine, it is characterised in that including:
Initial solution acquisition module, for obtaining the initial solution of preset model, described initial solution is used for representing that multiple physical machine is with many
The mapping relations of individual virtual machine;
Optimal solution acquisition module, is used for using preset algorithm that described initial solution is carried out computing, obtains described preset model in institute
State the optimal solution under preset algorithm;
Deployment module, for carrying out the deployment of virtual machine according to the optimal solution obtained;
Wherein, the parameter in described preset model includes the CPU of virtual machine, internal memory and hard disk size.
System the most according to claim 6, it is characterised in that the object function of described preset model is:
Wherein,
The constraints of described preset model is:
xij∈ 0,1}, i=1,2,3 ..., N;J=1,2,3 ..., M;
Wherein, σ represents the load variance of physical machine, and HAU represents the utilization rate of physical machine, HTjRepresent whether physical machine is in unlatching
State, HLjRepresent the load of physical machine, ω1, ω2, ω3Representing load weight, AHL represents the average load of physical machine;N represents
Virtual machine quantity;M represents physical machine quantity;VCi、VMi、VDiRepresent the CPU of virtual machine i, internal memory, hard disk size respectively;HCj、
HMj、HDjRepresent that physical machine j has CPU, internal memory, hard disk size respectively;xijFor binary number, xijIt is 1 when being, represents virtual machine
I is deployed in physical machine j, otherwise, and xijIt is 0;
Described optimal solution acquisition module, specifically for using preset algorithm that described initial solution is carried out computing, is met described
Constraints and make HAU maximum and solution that σ is minimum.
System the most according to claim 6, it is characterised in that using described initial solution as population, described population includes many
Individual population at individual, each population at individual represents the mapping relations of virtual machine and physical machine;
Accordingly, described optimal solution acquisition module, specifically for:
By genetic algorithm, the population at individual in described population is carried out evolution iteration, obtain the population at individual after evolving and bag
Include the population of the population at individual after described evolution.
System the most according to claim 8, it is characterised in that described optimal solution acquisition module includes:
Fitness acquisition module, in iterative process each time, according to the profit of the physical machine of population at individual each in population
The fitness of population at individual is obtained by the load variance of rate and physical machine;
Treat Advanced group species individuality acquisition module, for the fitness according to described population at individual, use roulette method in described kind
The population at individual treating heredity is chosen in Qun.
System the most according to claim 9, it is characterised in that described fitness acquisition module, specifically for according to as follows
The fitness of formula calculating population at individual:
Wherein,
Wherein, Fitness represents the fitness of population at individual, and σ represents the load variance of physical machine, and HAU represents the profit of physical machine
With rate, λ1, λ2Represent weighted factor, and λ1> 0, λ2> 0, HTjRepresent whether physical machine is in opening, HLjRepresent physics
The load of machine, ω1, ω2, ω3Representing load weight, AHL represents the average load of physical machine, and N represents virtual machine quantity;M represents
Physical machine quantity;VCi、VMi、VDiRepresent the CPU of virtual machine i, internal memory, hard disk size respectively;HCj、HMj、HDjRepresent thing respectively
Reason machine j has CPU, internal memory, hard disk size.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610228790.8A CN105843689A (en) | 2016-04-12 | 2016-04-12 | Virtual machine deployment method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610228790.8A CN105843689A (en) | 2016-04-12 | 2016-04-12 | Virtual machine deployment method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105843689A true CN105843689A (en) | 2016-08-10 |
Family
ID=56597362
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610228790.8A Pending CN105843689A (en) | 2016-04-12 | 2016-04-12 | Virtual machine deployment method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105843689A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107491341A (en) * | 2017-08-31 | 2017-12-19 | 福州大学 | A kind of virtual machine distribution method based on particle group optimizing |
CN110308993A (en) * | 2019-06-27 | 2019-10-08 | 大连理工大学 | A kind of cloud computing resources distribution method based on improved adaptive GA-IAGA |
CN111324422A (en) * | 2020-02-24 | 2020-06-23 | 武汉轻工大学 | Multi-target virtual machine deployment method, device, equipment and storage medium |
CN111338757A (en) * | 2020-02-24 | 2020-06-26 | 武汉轻工大学 | Energy-optimized virtual machine deployment method and system |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8423998B2 (en) * | 2010-06-04 | 2013-04-16 | International Business Machines Corporation | System and method for virtual machine multiplexing for resource provisioning in compute clouds |
CN104572251A (en) * | 2015-01-30 | 2015-04-29 | 中国联合网络通信集团有限公司 | Virtual machine deploying method and device |
CN104899100A (en) * | 2015-05-28 | 2015-09-09 | 北京邮电大学 | Resource scheduling method for cloud system |
-
2016
- 2016-04-12 CN CN201610228790.8A patent/CN105843689A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8423998B2 (en) * | 2010-06-04 | 2013-04-16 | International Business Machines Corporation | System and method for virtual machine multiplexing for resource provisioning in compute clouds |
CN104572251A (en) * | 2015-01-30 | 2015-04-29 | 中国联合网络通信集团有限公司 | Virtual machine deploying method and device |
CN104899100A (en) * | 2015-05-28 | 2015-09-09 | 北京邮电大学 | Resource scheduling method for cloud system |
Non-Patent Citations (3)
Title |
---|
汪永康: "IaaS环境下虚拟服务资源分配策略研究", 《万方数据》 * |
罗贺等: "一种基于成本的云服务资源分配策略", 《中国管理科学》 * |
罗贺等: "基于负载均衡的云服务资源配置策略研究", 《中国管理科学》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107491341A (en) * | 2017-08-31 | 2017-12-19 | 福州大学 | A kind of virtual machine distribution method based on particle group optimizing |
CN110308993A (en) * | 2019-06-27 | 2019-10-08 | 大连理工大学 | A kind of cloud computing resources distribution method based on improved adaptive GA-IAGA |
CN110308993B (en) * | 2019-06-27 | 2022-12-13 | 大连理工大学 | Cloud computing resource allocation method based on improved genetic algorithm |
CN111324422A (en) * | 2020-02-24 | 2020-06-23 | 武汉轻工大学 | Multi-target virtual machine deployment method, device, equipment and storage medium |
CN111338757A (en) * | 2020-02-24 | 2020-06-26 | 武汉轻工大学 | Energy-optimized virtual machine deployment method and system |
CN111324422B (en) * | 2020-02-24 | 2024-04-16 | 武汉轻工大学 | Multi-target virtual machine deployment method, device, equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Mohamed et al. | Adaptive guided differential evolution algorithm with novel mutation for numerical optimization | |
Parvizi et al. | Utilization-aware energy-efficient virtual machine placement in cloud networks using NSGA-III meta-heuristic approach | |
Ali | Knowledge-based optimization model for control valve locations in water distribution networks | |
Mohamed et al. | Differential evolution with novel mutation and adaptive crossover strategies for solving large scale global optimization problems | |
CN105843689A (en) | Virtual machine deployment method and system | |
CN110673951B (en) | Mimicry scheduling method, system and medium for general operation environment | |
Sinha et al. | An interactive evolutionary multi-objective optimization algorithm with a limited number of decision maker calls | |
CN105704255A (en) | Server load balancing method based on genetic algorithm | |
Soroudi et al. | Efficient immune‐GA method for DNOs in sizing and placement of distributed generation units | |
US20190043025A1 (en) | Genetic algorithms in blockchain space | |
CN109447264B (en) | Virtual machine placement genetic optimization method based on VHAM-R model in cloud computing environment | |
US8639643B2 (en) | Classification of a document according to a weighted search tree created by genetic algorithms | |
Idrissi et al. | A new approach for a better load balancing and a better distribution of resources in cloud computing | |
Basu et al. | Implementing fuzzy TOPSIS in cloud type and service provider selection | |
CN108764741B (en) | For determining the method and device of the layout of the production equipment in factory set region | |
He et al. | Biobjective optimization-based frequency regulation of power grids with high-participated renewable energy and energy storage systems | |
CN104580518A (en) | Load balance control method used for storage system | |
Liu | Optimal computing budget allocation to the differential evolution algorithm for large-scale portfolio optimization | |
Peng et al. | Security-aware computation offloading for mobile edge computing-enabled smart city | |
Sallam et al. | Two-phase differential evolution framework for solving optimization problems | |
Ouyang et al. | Amended harmony search algorithm with perturbation strategy for large-scale system reliability problems | |
CN112882805A (en) | Profit optimization scheduling method based on task resource constraint | |
CN109298914A (en) | A kind of Docker based on three-tier architecture and virtual machine initial placement method | |
CN107621980A (en) | A kind of virtual machine migration method, cluster control system and control device | |
Ma et al. | Aiming at QoS: A Modified DE Algorithm for Task Allocation in Cloud Computing |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into 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: 20160810 |