CN106845643A - A kind of cloud service system dynamic evolution method based on improved adaptive GA-IAGA - Google Patents

A kind of cloud service system dynamic evolution method based on improved adaptive GA-IAGA Download PDF

Info

Publication number
CN106845643A
CN106845643A CN201710070315.7A CN201710070315A CN106845643A CN 106845643 A CN106845643 A CN 106845643A CN 201710070315 A CN201710070315 A CN 201710070315A CN 106845643 A CN106845643 A CN 106845643A
Authority
CN
China
Prior art keywords
cloud service
class
population
function
individuality
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201710070315.7A
Other languages
Chinese (zh)
Other versions
CN106845643B (en
Inventor
徐洪珍
李卫东
宋文琳
朱雪琴
钟国韵
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
East China Institute of Technology
Original Assignee
East China Institute of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by East China Institute of Technology filed Critical East China Institute of Technology
Priority to CN201710070315.7A priority Critical patent/CN106845643B/en
Publication of CN106845643A publication Critical patent/CN106845643A/en
Application granted granted Critical
Publication of CN106845643B publication Critical patent/CN106845643B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biophysics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Theoretical Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Molecular Biology (AREA)
  • Biomedical Technology (AREA)
  • Genetics & Genomics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Physiology (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Telephonic Communication Services (AREA)

Abstract

The present invention discloses a kind of cloud service system dynamic evolution method based on improved adaptive GA-IAGA, and the dynamic evolution method includes：Dynamic evolution coding is carried out to the cloud service in cloud service system；The cloud service system is initialized；Fitness function is built, the fitness value of each candidate's cloud service in i-th class candidate's cloud service is calculated；Compare the fitness value and the size of the first given threshold of each candidate's cloud service in i-th class candidate's cloud service, therefrom select the i-th class candidate cloud service corresponding more than or equal to the fitness value of the first given threshold as target cloud service；Target cloud service according to all kinds of cloud services generates target cloud service system.The cloud service system dynamic evolution method that the present invention is provided ensure that the dynamic evolution method can be in global scope preferentially, simultaneously, can be selected from the angle of Optimizing Search by selection operation, cloud service generation by generation is optimized, can quickly and efficiently complete to cloud service it is global preferentially.

Description

A kind of cloud service system dynamic evolution method based on improved adaptive GA-IAGA
Technical field
The present invention relates to dynamic evolution field, more particularly to a kind of cloud service system dynamic based on improved adaptive GA-IAGA Evolution method.
Background technology
With continuing to develop for software engineering, cloud computing is increasingly becoming the applied environment of main flow.Under cloud computing environment, service Turn into key concept, show all trend for all servicing.Cloud service system will turn into the mainstream applications form of software.So And, the dynamic of cloud computing environment, open and complexity, and user's request frequent change, it is desirable to cloud service system should Continuous dynamic evolution.Neither one cloud service system can meet user's requirement always, and constantly running is not gone down.In order to adapt to Open cloud computing environment, each cloud service of cloud service system needs to be developed accordingly.Due to clock availability The characteristics of, the current study hotspot that cloud service system dynamic evolution has turned into.
The complexity of cloud environment determines that cloud service Population is larger, when cloud service system dynamic evolution is carried out, passes System optimization method is to seek optimal solution from single initial value iteration, is easily strayed into locally optimal solution, it is impossible to quickly and efficiently complete complete Office preferentially develops.
The content of the invention
It is described dynamic it is an object of the invention to provide a kind of cloud service system dynamic evolution method based on improved adaptive GA-IAGA The overall situation that state evolution method can quickly and efficiently complete cloud service system preferentially develops.
To achieve the above object, the invention provides following scheme：
A kind of cloud service system dynamic evolution method based on improved adaptive GA-IAGA, the dynamic evolution method includes：
Dynamic evolution coding is carried out to the cloud service in cloud service system, the cloud service of the definition composition cloud service system is：
Si={ ni,pi,fi,fdi,fei,Ii,Idi,odi,idi,noi,Ci,Ni(0,1) },
Wherein, i represents cloud service sequence number, SiRepresent i-th cloud service, ni、pi、fi、fdi、fei、Ii、Idi、odi、idi、 noi、CiAnd Ni(0,1) is the encoded components of i-th cloud service, niRepresent the corresponding title of i-th cloud service, piRepresent i-th The parameter of cloud service, the parameter is the inner element of i-th cloud service, pi=(ei1,ei2,...,eiM), wherein M is The number of the inner element of i-th cloud service, eikBe in i-th cloud service, k-th inner element relied on except k-th The number of the inner element outside inner element, fiRepresent the function number of i-th cloud service, fdiRepresent having for i-th cloud service The function number of defect, feiRepresent the function number of the dependence environment of i-th cloud service, IiRepresent the number of ports of i-th cloud service； IdiRepresent the defective number of ports of i-th cloud service, odiRepresent the degree of coupling that of i-th cloud service, idiRepresent i-th cloud What is serviced enters the degree of coupling, noiRepresent the Failure count of i-th cloud service, CiRepresent the class belonging to i-th cloud service, Ni(0,1) The corresponding random number of i-th cloud service is represented, the interval of the random number is (0,1)；
The cloud service system is initialized, the initial of the i-th class cloud service in the cloud service system is generated at random Cloud service population；
Fitness function is built, each individual fitness value in the initial cloud service population of the i-th class is calculated；
First screening is carried out to the initial cloud service population of i-th class and determines candidate's cloud service；
Compare the fitness value of all candidate's cloud services of i-th class and the size of the first given threshold, therefrom select The i-th class candidate cloud service corresponding more than or equal to the fitness value of the first given threshold as target cloud service, if the i-th class The target cloud service of cloud service uses probability selection method choice one to be taken as the i-th final class target cloud more than one, then Business；If the fitness value of all candidate's cloud services of the i-th class cloud service is respectively less than the first given threshold, to all candidate's clouds Service carries out crossover operation and mutation operation to update i-th class candidate's cloud service；
Final goal cloud service according to all kinds of cloud services generates target cloud service system.
Optionally, it is described to generate the initial cloud service population of the i-th class in the cloud service system at random, specifically include：
Using formula cijk=| (cijkmax-cijkmin)×Nij(0,1)+cijkmin| calculate each i-th class cloud service population In j-th initial value of each component of individuality, wherein, | | be bracket function, cijkRepresent j-th in the i-th class cloud service population K-th individual component initial value, cijkmaxJ-th k-th component of individuality is in value model in representing the i-th class cloud service population Enclose interior maximum, cijkminRepresent in the i-th class cloud service population the minimum of j-th k-th component of individuality in span Value；The big I of the initial cloud service population of the i-th class of the random generation gives as needed.
Optionally, the structure fitness function, specifically includes：Build fitness function fij(Q)=w1i*f1ij(FC)+ w2i*f2ij(FD)+w3i*f3ij(IC)+w4i*f4ij(CP)+w5i*f5ij(CH)+w6i*f6ij(EC) in the i-th class cloud service population of calculating J-th fitness of individuality, wherein, wriRepresent the weights of each quality index of the i-th class cloud service, r=1 ..., 6, f1ij(FC) It is j-th function accuracy function of individuality, f in the i-th class cloud service population2ij(FD) for j-th in the i-th class cloud service population Individual functional independence function, f3ij(IC) it is j-th interface correctness function of individuality, f in the i-th class cloud service population4ij (CP) it is j-th coupling function of individuality, f in the i-th class cloud service population5ij(CH) for j-th in the i-th class cloud service population Individual cohesion function, f6ij(EC) it is j-th evolution compatibility function of individuality in the i-th class cloud service population.
Optionally, j-th function accuracy function f of individuality in the i-th class cloud service population1ij(FC) it is：
f1ij(FC)=1-fdij/fij,
Wherein, fdijIt is in j-th function of individuality in the i-th class cloud service population, to detect defective function number, fijIt is j-th function sum of individuality in the i-th class cloud service population；
J-th functional independence function f of individuality in the i-th class cloud service population2ij(FD) it is：
f2ij(FD)=1-feij/fij,
Wherein, feijThe function number of environment is relied on to be detected in j-th individuality in the i-th class cloud service population；
J-th interface correctness function f of individuality in the i-th class cloud service population3ij(IC) it is：
f3ij(IC)=1-Idij/Iij,
Wherein, IdijIt is in j-th interface of individuality in the i-th class cloud service population, to detect defective number of ports, IijIt is j-th interface sum of individuality in the i-th class cloud service population；
J-th coupling function f of individuality in the i-th class cloud service population4ij(CP) it is：
f4ij(CP)=odij/(odij+idij),
Wherein, odijFor j-th individuality goes out the degree of coupling, id in the i-th class cloud service populationijIt is the i-th class cloud service population In j-th individuality enter the degree of coupling；
J-th cohesion function f of individuality in the i-th class cloud service population5ij(CH) it is：
Wherein, M is j-th number of individual inner element, e in the i-th class cloud service populationijkIt is the i-th class cloud service population In j-th individuality the other elements number that is relied on of k-th element；
J-th evolution compatibility function f of individuality in the i-th class cloud service population6ij(EC) it is：
f6ij(EC)=1/ (noij+ 1),
Wherein, noijWhen replacing legacy version for the redaction of j-th individuality in the i-th class cloud service population, the number of times for failing.
Optionally, the initial cloud service population to each i-th class cloud service carries out first screening determination candidate's cloud Service includes：
Compare the individual corresponding fitness value of each cloud service and the second setting threshold in each i-th class cloud service initial population The size of value, therefrom selects the initial cloud serve individual corresponding more than the fitness value of the second given threshold as i-th Class candidate's cloud service；If the corresponding i-th class initial cloud serve individual of the fitness value for being not greater than the second given threshold, Initialization is then re-started to the i-th class cloud service, the initial cloud service population of the i-th class cloud service is generated；
Optionally, the mutation operation includes：The number of times of variation is produced by a random number first, is then passed through every time Mutation operator carries out random variation, and the mutation operator includes increasing that function, to the cloud service, is removed from the cloud service Function, the repair function from the cloud service changes the function number of dependence environment in the cloud service, introduces interface to the cloud Service, interface is removed from the cloud service, and interface is repaired from the cloud service, and that changes the cloud service goes out the degree of coupling, The degree of coupling that enters of the cloud service is changed, the parameter of the cloud service is changed, the Failure count of the cloud service is changed.
Optionally, the crossover operation includes：The number of times for intersecting is produced by a random number first, it is then random every time Select the encoded components of the cloud service for crosspoint, exchange the encoded components of the cloud service.
Optionally, the probability selection method includes：Each target cloud service per class cloud service is individual selected general Rate and its fitness value are proportional, if the individual number of the target cloud service of the i-th class cloud service is k, wherein target cloud service Body SijFitness value be fij, then target cloud service individuality SijThe probability for being selected as final goal cloud service is：
Wherein, f represents fitness value, and j represents sequence number, fijRepresent j-th target cloud service individuality of the i-th class cloud service Fitness value.
According to the specific embodiment that the present invention is provided, the invention discloses following technique effect：The present invention is by genetic algorithm In selection operation, crossover operation and mutation operation be introduced into cloud service system dynamic evolution method, due to crossover operation and change The migration rules of ETTHER-OR operation are random, it is ensured that the dynamic evolution method can be in global scope preferentially.Meanwhile, with one As the undirected search that is carried out of stochastic search methods it is different, can be carried out from a large amount of population at individual by selection operation twice Rapid Optimum is selected, and cloud service generation by generation is optimized evolution, and quickly approaches optimal cloud service, so as to ensure cloud service During system dynamic evolution, the overall situation individual to a large amount of cloud services can be quickly and efficiently completed preferentially.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to institute in embodiment The accompanying drawing for needing to use is briefly described, it should be apparent that, drawings in the following description are only some implementations of the invention Example, for those of ordinary skill in the art, on the premise of not paying creative work, can also obtain according to these accompanying drawings Obtain other accompanying drawings.
Fig. 1 is the flow chart of the dynamic evolution method of embodiment of the present invention cloud service system；
Fig. 2 is the cloud service code pattern before embodiment of the present invention crossover operation；
Fig. 3 is the cloud service code pattern after embodiment of the present invention crossover operation；
Fig. 4 is that embodiment of the present invention itinerary formulates cloud service system structure chart.
Specific embodiment
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 Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than whole embodiments.It is based on Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made Embodiment, belongs to the scope of protection of the invention.
It is an object of the invention to provide a kind of dynamic evolution method of the cloud service system based on improved adaptive GA-IAGA, with fast The overall situation that speed efficiently completes cloud service system preferentially develops.
It is below in conjunction with the accompanying drawings and specific real to enable the above objects, features and advantages of the present invention more obvious understandable The present invention is further detailed explanation to apply mode.
Fig. 1 is the flow chart of the dynamic evolution method of embodiment of the present invention cloud service system, as shown in figure 1, cloud service system The dynamic evolution method step of system is as follows：
Step 100：Determine the species number n of cloud service in cloud service system；
Step 101：N class cloud services are encoded, the cloud service of the definition composition cloud service system is：
Si={ ni,pi,fi,fdi,fei,Ii,Idi,odi,idi,noi,Ci,Ni(0,1) },
Wherein, i represents cloud service sequence number, SiRepresent i-th cloud service, ni、pi、fi、fdi、fei、Ii、Idi、odi、idi、 noi、CiAnd Ni(0,1) is the encoded components of i-th cloud service, niRepresent the corresponding title of i-th cloud service, piRepresent i-th The parameter of cloud service, the parameter is the inner element of i-th cloud service, pi=(ei1,ei2,...,eiM), wherein M is The number of the inner element of i-th cloud service, eikBe in i-th cloud service, k-th inner element relied on except k-th The number of the inner element outside inner element, fiRepresent the function number of i-th cloud service, fdiRepresent having for i-th cloud service The function number of defect, feiRepresent the function number of the dependence environment of i-th cloud service, IiRepresent the number of ports of i-th cloud service； IdiRepresent the defective number of ports of i-th cloud service, odiRepresent the degree of coupling that of i-th cloud service, idiRepresent i-th cloud What is serviced enters the degree of coupling, noiRepresent the Failure count of i-th cloud service, CiRepresent the class belonging to i-th cloud service, Ni(0,1) The corresponding random number of i-th cloud service is represented, the interval of the random number is (0,1).
The coding of the cloud service system of table 1
Step 102：The cloud service system is initialized, the i-th class cloud clothes in the cloud service system are generated at random The initial cloud service population of business, i=1 ..., n, it is individual that the cloud service population includes multiple such cloud service；
Step 103：Fitness function is built, the fitness value of each individual cloud service in the i-th class cloud service population, i is calculated =1 ..., n；
Compare the individual corresponding fitness value of each cloud service and the second setting threshold in each i-th class cloud service initial population The size of value, therefrom selects the initial cloud serve individual corresponding more than the fitness value of the second given threshold as i-th Class candidate's cloud service；If the corresponding i-th class initial cloud serve individual of the fitness value for being not greater than the second given threshold, Initialization is then re-started to the i-th class cloud service, the initial cloud service population of the i-th class cloud service is generated, specially step 104 is arrived Step 108.
Step 104：Whether the individual corresponding fitness value of each cloud service sets more than second in judging the i-th class cloud service population Determine threshold value；
Step 105：If the individual corresponding fitness value of cloud service in the i-th class cloud service population is less than or equal to second Given threshold, then abandon the cloud service individual；
Step 106：If the individual corresponding fitness value of the cloud service of the i-th class cloud service population is more than the second preset value, Then retain the cloud service individual；
Step 107：If individual without residue cloud service in the i-th class cloud service population, the i-th class of generation is reinitialized Cloud service population, the remaining cloud service is individual for the cloud service for remaining is individual；
Step 108：It is if there is remaining cloud service individual in the i-th class cloud service population, the remaining cloud service is individual Used as candidate's cloud service, the remaining cloud service is individual for the cloud service for remaining is individual；
Step 109：Judge whether the fitness value of each candidate's cloud service in the i-th class cloud service population sets more than or equal to first Determine threshold value；
Step 110：If the fitness value of all candidate's cloud services is respectively less than the first setting threshold in the i-th class cloud service population Value, then do crossover operation and mutation operation, to update i-th class candidate's cloud service to all candidate's cloud services；
Step 111：If there is the fitness value of candidate's cloud service in the i-th class cloud service population more than or equal to the first setting Threshold value, then using candidate's cloud service as target cloud service, select the mesh in all candidate's cloud services in the i-th class cloud service population Mark cloud service；
Step 112：Judge whether the target cloud service in the i-th class cloud service population is unique；
Step 113：If the target cloud service in the i-th class cloud service population is not unique, using the choosing of probability selection method One is selected as the i-th final class target cloud service；
Step 114：Final goal cloud service according to all kinds of cloud services generates target cloud service system.
Wherein, in step 102, the cloud service system is initialized, in generating each cloud service system at random The initial cloud service population of the i-th class cloud service, specifically includes：
Using formula cijk=| (cijkmax-cijkmin)×Nij(0,1)+cijkmin| calculate each i-th class cloud service population In j-th initial value of each component of individuality, wherein, | | be bracket function, cijkRepresent j-th in the i-th class cloud service population K-th individual component initial value, cijkmaxJ-th k-th component of individuality is in value model in representing the i-th class cloud service population Enclose interior maximum, cijkminRepresent in the i-th class cloud service population the minimum of j-th k-th component of individuality in span Value；The big I of the initial cloud service population of the i-th class of the random generation gives as needed.
Step 103 builds fitness function, specifically includes：
Build fitness function fij(Q)=w1i*f1ij(FC)+w2i*f2ij(FD)+w3i*f3ij(IC)+w4i*f4ij(CP)+ w5i*f5ij(CH)+w6i*f6ij(EC) j-th fitness of individuality in the i-th class cloud service population is calculated, wherein, wriRepresent the i-th class The weights of each quality index of cloud service, r=1 ..., 6, f1ij(FC) it is j-th function of individuality in the i-th class cloud service population Correctness function, f2ij(FD) it is j-th functional independence function of individuality, f in the i-th class cloud service population3ij(IC) it is the i-th class J-th interface correctness function of individuality, f in cloud service population4ij(CP) it is j-th coupling of individuality in the i-th class cloud service population Conjunction property function, f5ij(CH) it is j-th cohesion function of individuality, f in the i-th class cloud service population6ij(EC) it is the i-th class cloud service J-th evolution compatibility function of individuality in population.
J-th function accuracy function f of individuality in the i-th class cloud service population1ij(FC) it is：
f1ij(FC)=1-fdij/fij,
Wherein, fdijIt is in j-th function of individuality in the i-th class cloud service population, to detect defective function number, fijIt is j-th function sum of individuality in the i-th class cloud service population；
J-th functional independence function f of individuality in the i-th class cloud service population2ij(FD) it is：
f2ij(FD)=1-feij/fij,
Wherein, feijThe function number of environment is relied on to be detected in j-th individuality in the i-th class cloud service population；
J-th interface correctness function f of individuality in the i-th class cloud service population3ij(IC) it is：
f3ij(IC)=1-Idij/Iij,
Wherein, IdijIt is in j-th interface of individuality in the i-th class cloud service population, to detect defective number of ports, IijIt is j-th interface sum of individuality in the i-th class cloud service population；
J-th coupling function f of individuality in the i-th class cloud service population4ij(CP) it is：
f4ij(CP)=odij/(odij+idij),
Wherein, odijFor j-th individuality goes out the degree of coupling, id in the i-th class cloud service populationijIt is the i-th class cloud service population In j-th individuality enter the degree of coupling；
J-th cohesion function f of individuality in the i-th class cloud service population5ij(CH) it is：
Wherein, M is j-th number of individual inner element, e in the i-th class cloud service populationijkIt is the i-th class cloud service population In j-th individuality the other elements number that is relied on of k-th element；
J-th evolution compatibility function f of individuality in the i-th class cloud service population6ij(EC) it is：
f6ij(EC)=1/ (noij+ 1),
Wherein, noijWhen replacing legacy version for the redaction of j-th individuality in the i-th class cloud service population, the number of times for failing.
The mutation operation in step 110 includes：The number of times of variation is produced by a random number first, then every time Random variation is carried out by mutation operator, the mutation operator includes increasing function to the cloud service, from the cloud service Remove function, the repair function from the cloud service changes the function number of dependence environment in the cloud service, introduces interface to institute Cloud service is stated, interface is removed from the cloud service, interface is repaired from the cloud service, that changes the cloud service goes out coupling Degree, changes the degree of coupling that enters of the cloud service, changes the parameter of the cloud service, changes the Failure count of the cloud service.
The crossover operation in step 110 includes：The number of times for intersecting is produced by a random number first, then every time The encoded components of the cloud service are randomly choosed for crosspoint, the encoded components of the cloud service are exchanged.
Specifically, selection operation twice is carried out by step 108,111 and 113 pairs of each cloud service individualities.
The purpose of selection operation is to carry out quickly preferentially screening to a large amount of cloud service individualities.By step 108 carry out First time selection operation, is screened for the first time from a large amount of individualities of every class cloud service population, is selected fitness value and is set more than second Determine the individual candidate's cloud service as such cloud service of cloud service of threshold value；Second selection behaviour is carried out by step 111 and 113 Make, selecting fitness value from every class candidate cloud service first should more than or equal to the individual conduct of the cloud service of the first given threshold The target cloud service of class cloud service, if target cloud service is not unique, leads in all target cloud services of such cloud service Cross probability selection method and select final goal cloud service of unique cloud service as such cloud service, if such cloud service The fitness value of all candidate's cloud services is respectively less than the first given threshold, then carry out mutation operation and intersection to all candidate's cloud services Operation is of future generation hereditary to carry out.The final goal cloud service per class cloud service constitutes optimal cloud service as optimal cloud service System, i.e. target cloud service system.In the probability selection method, each target cloud service per class cloud service is individual selected The probability and its fitness value selected are proportional, if the individual number of the target cloud service of the i-th class cloud service is k, wherein target cloud Serve individual SijFitness value be fij, then target cloud service individuality SijThe probability for being selected as final goal cloud service is：
Wherein, f represents fitness value, and j represents sequence number, fijRepresent j-th target cloud service individuality of the i-th class cloud service Fitness value.
Specifically, an existing cloud service during the mutation operator in step 110 refers to cloud service system carries out some Change operation, so that a new cloud service is formed, to reach the purpose for increasing or changing characteristic.The variation of the present embodiment definition Operator is as follows：
Increase One function FkTo cloud service Si；
From cloud service SiMiddle remove function Fi；
Repair cloud service SiIn defective function Fj；
Introduce an interface IkTo cloud service Si；
From cloud service SiMiddle removal interface Ii；
Repair cloud service SiIn defective interface Ij；
Modification cloud service SiGo out the degree of coupling；
Modification cloud service SiEnter the degree of coupling；
Modification cloud service SiParameter；
Modification cloud service SiFailure count.
The crossover operation of step 110 usually realized by exchanging some of cloud service unit, the purpose of crossover operation be for Combine the optimal properties in two cloud services.For example, it is assumed that cloud service individuality G before crossover operation1And G2Coding as shown in Fig. 2 If the defective function number and defective number of ports of selection cloud service are crosspoint respectively, obtained after crossover operation twice Cloud service individual be respectively G1' and G2', by cloud service individuality G after crossover operation twice1' and G2' coding it is as shown in Figure 3.
The genetic manipulation that cloud service system develops includes selection operation, mutation operation and crossover operation.The present embodiment sets Cloud service evolution end condition be：When all kinds of cloud service populations select only one fitness value more than or equal to the first threshold During the final cloud service of value individual (i.e. optimal cloud service is individual), genetic algorithm terminates.
Used as specific embodiment of the invention, Fig. 4 is that embodiment of the present invention itinerary formulates cloud service system structure chart, As shown in figure 4, an itinerary formulates cloud service system and includes five class cloud services, it is respectively travelling planning service S1, make a reservation for Aircraft/train ticket services S2, predetermined food and beverage sevice S3, predetermined hotel service S4With tourist service S5。
The dynamic evolution that itinerary formulates cloud service system is carried out below, is comprised the following steps that：
1) cloud service that itinerary formulates cloud service system is encoded, as shown in table 2 for itinerary formulates cloud The cloud service coding of service system.
The itinerary of table 2 formulates the coding of cloud service system
2) cloud service system is initialized, the initial of all kinds of cloud services in the cloud service system is generated at random Population.
Assuming that 10 are per the size of class cloud service initialization population, it is specific after the random generation of their individual cloud service Supplemental characteristic is as shown in table 3 below：
The quality index of the initial all kinds of cloud service populations of table 3
3) Proper treatment is built, the appropriateness of each cloud service individuality in five class cloud service populations is calculated respectively using Proper treatment Value；Proper treatment is fij(Q)=w1i*f1ij(FC)+w2i*f2ij(FD)+w3i*f3ij(IC)+w4i*f4ij(CP)+w5i*f5ij(CH)+ w6i*f6ij(EC)。
Illustrated by taking planning service of travelling as an example below：
With the individual S in planning service of travelling11As a example by, calculate individuality S116 values of quality index：
f111(FC)=1-fd11/f11=1-2/5=0.6,
f211(FD)=1-fe11/f11=1-1/5=0.8,
f311(IC)=1-Id11/I11=1-1/5=0.8,
f411(CP)=od11/(od11+id11)=2/ (2+1)=2/3,
f611(EC)=1/ (no11+ 1)=1/ (2+1)=1/3,
Individual appropriateness value is calculated according to appropriateness value computing formula：f11(Q)=0.2*0.6+0.15*0.8+0.2*0.8+ 0.15*2/3+0.15*0.5+0.15*1/3=0.625, using same computational methods calculate respectively tourism planning service in The appropriateness value of other 9 individualities.
4) judge to travel and plan whether the appropriate values of 10 cloud service individualities in service population are more than the second given threshold, such as Fruit is less than or equal to the second given threshold, then abandon the cloud service individual；If greater than the second given threshold, then judge that the cloud takes Whether there is remaining cloud service individual in business population, if individual without remaining cloud service, reinitialize generation such cloud clothes Business population；It is if remaining cloud service is individual, then remaining cloud service is individual as candidate's cloud service.
5) judge whether the fitness value of each candidate's cloud service in travelling planning service is more than or equal to the first given threshold, If the fitness value of all candidate's cloud services is respectively less than the first given threshold in travelling planning service population, to all candidates Crossover operation and mutation operation are done in cloud service, to update candidate's cloud service of travelling planning service；If travelling planning service kind There is candidate cloud service of the fitness value more than or equal to the first given threshold in group, then taken this candidate's cloud service as target cloud Business, selects the target cloud service in all candidate's cloud services in travelling planning service population.
6) judge whether the target cloud service in travelling planning service population is unique, if in travelling planning service population Target cloud service is not unique, then using probability selection method choice one as such final target cloud service.
The target cloud service of travelling planning service in cloud service has been obtained by above step.It is identical with above-mentioned steps, can To obtain the target cloud service of other four class cloud services, the final goal cloud service generation target cloud service according to five class cloud services System.
The cloud service system dynamic evolution method based on improved adaptive GA-IAGA that the present invention passes through is by the choosing in genetic algorithm Select operation, crossover operation and mutation operation to be introduced into cloud service system dynamic evolution method, due to crossover operation and mutation operation Migration rules be random, it is ensured that the dynamic evolution method can be in global scope preferentially.Meanwhile, with it is general with The undirected search that machine searching method is carried out is different, can be carried out from a large amount of population at individual by selection operation twice quick excellent Change selection, cloud service generation by generation is optimized evolution, and quickly approach optimal cloud service, so as to ensure that cloud service system is moved When state develops, the overall situation individual to a large amount of cloud services can be quickly and efficiently completed preferentially.
Each embodiment is described by the way of progressive in this specification, and what each embodiment was stressed is and other The difference of embodiment, between each embodiment identical similar portion mutually referring to.
Specific case used herein is set forth to principle of the invention and implementation method, and above example is said It is bright to be only intended to help and understand the method for the present invention and its core concept；Simultaneously for those of ordinary skill in the art, foundation Thought of the invention, will change in specific embodiments and applications.In sum, this specification content is not It is interpreted as limitation of the present invention.

Claims (8)

1. a kind of cloud service system dynamic evolution method based on improved adaptive GA-IAGA, it is characterised in that the dynamic evolution side Method includes：
Dynamic evolution coding is carried out to the cloud service in cloud service system, the cloud service of the definition composition cloud service system is：
Si={ ni,pi,fi,fdi,fei,Ii,Idi,odi,idi,noi,Ci,Ni(0,1) },
Wherein, i represents cloud service sequence number, SiRepresent i-th cloud service, ni、pi、fi、fdi、fei、Ii、Idi、odi、idi、noi、Ci And Ni(0,1) is the encoded components of i-th cloud service, niRepresent the corresponding title of i-th cloud service, piRepresent i-th cloud service Parameter, the parameter is the inner element of i-th cloud service, pi=(ei1,ei2,...,eiM), wherein M is i-th cloud The number of the inner element of service, eikBe in i-th cloud service, k-th inner element relied on except k-th inner element Outside inner element number, fiRepresent the function number of i-th cloud service, fdiRepresent the defective work(of i-th cloud service Can number, feiRepresent the function number of the dependence environment of i-th cloud service, IiRepresent the number of ports of i-th cloud service；IdiRepresent i-th The defective number of ports of individual cloud service, odiRepresent the degree of coupling that of i-th cloud service, idiRepresent i-th cloud service enters coupling It is right, noiRepresent the Failure count of i-th cloud service, CiRepresent the class belonging to i-th cloud service, Ni(0,1) i-th cloud is represented Corresponding random number is serviced, the interval of the random number is (0,1)；
The cloud service system is initialized, the initial cloud clothes of the i-th class cloud service in the cloud service system are generated at random Business population；
Fitness function is built, each individual fitness value in the initial cloud service population of the i-th class is calculated；
First screening is carried out to the initial cloud service population of i-th class and determines candidate's cloud service；
Compare the fitness value of all candidate's cloud services of i-th class and the size of the first given threshold, therefrom select and be more than Or the i-th class candidate cloud service corresponding equal to the fitness value of the first given threshold is used as target cloud service, if the i-th class cloud takes The target cloud service of business more than one, then using probability selection method choice one as the i-th final class target cloud service；Such as Really the fitness value of all candidate's cloud services of the i-th class cloud service is respectively less than the first given threshold, then to all candidate's cloud services Crossover operation and mutation operation is carried out to update i-th class candidate's cloud service；
Final goal cloud service according to all kinds of cloud services generates target cloud service system.
2. the dynamic evolution method of cloud service system according to claim 1, it is characterised in that described in the random generation The initial cloud service population of the i-th class, specifically includes in cloud service system：
Using formula cijk=| (cijkmax-cijkmin)×Nij(0,1)+cijkmin| calculate jth in each i-th class cloud service population The initial value of each individual component, wherein, | | it is bracket function, cikRepresent in the i-th class cloud service population j-th individuality K-th component initial value, cijkmaxJ-th k-th component of individuality is in span in representing the i-th class cloud service population Maximum, cijkminRepresent in the i-th class cloud service population the minimum value of j-th k-th component of individuality in span；Institute The big I for stating the initial cloud service population of the i-th class of random generation gives as needed.
3. the dynamic evolution method of cloud service system according to claim 1, it is characterised in that the structure fitness letter Number, specifically includes：Build fitness function fij(Q)=w1i*f1ij(FC)+w2i*f2ij(FD)+w3i*f3ij(IC)+w4i*f4ij (CP)+w5i*f5ij(CH)+w6i*f6ij(EC) j-th fitness of individuality in the i-th class cloud service population is calculated, wherein, wriRepresent The weights of each quality index of the i-th class cloud service, r=1 ..., 6, f1ij(FC) it is j-th individuality in the i-th class cloud service population Function accuracy function, f2ij(FD) it is j-th functional independence function of individuality, f in the i-th class cloud service population3ij(IC) it is J-th interface correctness function of individuality, f in i-th class cloud service population4ij(CP) for j-th in the i-th class cloud service population The coupling function of body, f5ij(CH) it is j-th cohesion function of individuality, f in the i-th class cloud service population6ij(EC) it is the i-th class J-th evolution compatibility function of individuality in cloud service population.
4. the dynamic evolution method of cloud service system according to claim 3, it is characterised in that
J-th function accuracy function f of individuality in the i-th class cloud service population1ij(FC) it is：
f1ij(FC)=1-fdij/fij,
Wherein, fdijIt is in j-th function of individuality in the i-th class cloud service population, to detect defective function number, fijFor J-th function sum of individuality in i-th class cloud service population；
J-th functional independence function f of individuality in the i-th class cloud service population2ij(FD) it is：
f2ij(FD)=1-feij/fij,
Wherein, feijThe function number of environment is relied on to be detected in j-th individuality in the i-th class cloud service population；
J-th interface correctness function f of individuality in the i-th class cloud service population3ij(IC) it is：
f3ij(IC)=1-Idij/Iij,
Wherein, IdijIt is in j-th interface of individuality in the i-th class cloud service population, to detect defective number of ports, IijFor J-th interface sum of individuality in i-th class cloud service population；
J-th coupling function f of individuality in the i-th class cloud service population4ij(CP) it is：
f4ij(CP)=odij/(odij+idij),
Wherein, odijFor j-th individuality goes out the degree of coupling, id in the i-th class cloud service populationijIt is jth in the i-th class cloud service population Individual enters the degree of coupling；
J-th cohesion function f of individuality in the i-th class cloud service population5ij(CH) it is：
f 5 i j ( C H ) = 1 M = 1 Σ k = 1 M e i j k M × ( M - 1 ) M > 1 ,
Wherein, M is j-th number of individual inner element, e in the i-th class cloud service populationijkIt is in the i-th class cloud service population The other elements number that j k-th individual element is relied on；
J-th evolution compatibility function f of individuality in the i-th class cloud service population6ij(EC) it is：
f6ij(EC)=1/ (noij+ 1),
Wherein, noijWhen replacing legacy version for the redaction of j-th individuality in the i-th class cloud service population, the number of times for failing.
5. the dynamic evolution method of cloud service system according to claim 1, it is characterised in that described to each described i-th The initial cloud service population of class cloud service carries out first screening and determines that candidate's cloud service includes：
Compare the individual corresponding fitness value of each cloud service in each i-th class cloud service initial population and the second given threshold Size, therefrom selects the initial cloud serve individual corresponding more than the fitness value of the second given threshold and is waited as the i-th class Select cloud service；If the corresponding i-th class initial cloud serve individual of the fitness value for being not greater than the second given threshold, right I-th class cloud service re-starts initialization, generates the initial cloud service population of the i-th class cloud service.
6. the dynamic evolution method of cloud service system according to claim 1, it is characterised in that the mutation operation bag Include：The number of times of variation is produced by a random number first, random variation, the variation are then carried out by mutation operator every time Operator includes increasing function to the cloud service, and the remove function from the cloud service, the repair function from the cloud service is repaiied Change the function number of dependence environment in the cloud service, introduce interface to the cloud service, interface is removed from the cloud service, from Interface is repaired in the cloud service, the degree of coupling that of the cloud service is changed, the degree of coupling that enters of the cloud service is changed, institute is changed The parameter of cloud service is stated, the Failure count of the cloud service is changed.
7. the dynamic evolution method of cloud service system according to claim 1, it is characterised in that the crossover operation bag Include：The number of times for intersecting is produced by a random number first, a coding of the cloud service is then randomly choosed every time Component is crosspoint, exchanges the encoded components of the cloud service.
8. the dynamic evolution method of cloud service system according to claim 1, it is characterised in that the probability selection method Including：The individual selected probability of each target cloud service and its fitness value per class cloud service is proportional, if the i-th class cloud takes The individual number of the target cloud service of business is k, wherein target cloud service individuality SijFitness value be fij, then target cloud service Individual SijThe probability for being selected as final goal cloud service is：
f i j f i 1 + f i 2 + ... f i k
Wherein, f represents fitness value, and j represents sequence number, fijRepresent the individual adaptation of j-th target cloud service of the i-th class cloud service Angle value.
CN201710070315.7A 2017-02-09 2017-02-09 A kind of cloud service system dynamic evolution method based on improved adaptive GA-IAGA Expired - Fee Related CN106845643B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710070315.7A CN106845643B (en) 2017-02-09 2017-02-09 A kind of cloud service system dynamic evolution method based on improved adaptive GA-IAGA

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710070315.7A CN106845643B (en) 2017-02-09 2017-02-09 A kind of cloud service system dynamic evolution method based on improved adaptive GA-IAGA

Publications (2)

Publication Number Publication Date
CN106845643A true CN106845643A (en) 2017-06-13
CN106845643B CN106845643B (en) 2019-04-19

Family

ID=59122324

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710070315.7A Expired - Fee Related CN106845643B (en) 2017-02-09 2017-02-09 A kind of cloud service system dynamic evolution method based on improved adaptive GA-IAGA

Country Status (1)

Country Link
CN (1) CN106845643B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107180286A (en) * 2017-07-18 2017-09-19 浙江财经大学 Manufacturing service supply chain optimization method and system based on modified pollen algorithm

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN203261358U (en) * 2013-05-28 2013-10-30 马传志 Genetic algorithm-based cloud computing server cluster
CN104932938A (en) * 2015-06-16 2015-09-23 中电科软件信息服务有限公司 Cloud resource scheduling method based on genetic algorithm
CN105740051A (en) * 2016-01-27 2016-07-06 北京工业大学 Cloud computing resource scheduling realization method based on improved genetic algorithm

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN203261358U (en) * 2013-05-28 2013-10-30 马传志 Genetic algorithm-based cloud computing server cluster
CN104932938A (en) * 2015-06-16 2015-09-23 中电科软件信息服务有限公司 Cloud resource scheduling method based on genetic algorithm
CN105740051A (en) * 2016-01-27 2016-07-06 北京工业大学 Cloud computing resource scheduling realization method based on improved genetic algorithm

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
刘卫宁: "基于改进量子遗传算法的云计算资源调度", 《计算机应用》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107180286A (en) * 2017-07-18 2017-09-19 浙江财经大学 Manufacturing service supply chain optimization method and system based on modified pollen algorithm
CN107180286B (en) * 2017-07-18 2020-12-01 浙江财经大学 Manufacturing service supply chain optimization method and system based on improved pollen algorithm

Also Published As

Publication number Publication date
CN106845643B (en) 2019-04-19

Similar Documents

Publication Publication Date Title
CN104954220B (en) A kind of virtual network service deployment method and apparatus
CN105095075B (en) Partly legalized fuzz testing case generation method based on finite state machine procotol
CN107067163A (en) A kind of breakdown maintenance work dispatching method and device
CN106777127A (en) The automatic generation method and system of the individualized learning process of knowledge based collection of illustrative plates
CN106375120B (en) A kind of IT infrastructure rapid topology method and device
CN108256431A (en) A kind of hand position identification method and device
CN105989737B (en) A kind of parking induction method
Gao et al. Collective learning in China's regional economic development
CN104035889B (en) Multi-state routing deriving method and multi-state routing deriving system
CN104217579B (en) Transportation network key road segment searching method based on section sub-network redundancy
CN106470122A (en) A kind of network failure locating method and device
CN103473265A (en) Method and device for analyzing layout of flow chart
Elazouni et al. Estimating the acceptability of new formwork systems using neural networks
CN105930944A (en) DEC-POMDP-based collaborative optimization decision method and device
CN107453928A (en) A kind of power telecom network pitch point importance evaluation method and device
CN103473606B (en) The public traffic network entirety transfer performance optimization method of the maximum shortest path length based on GIS
CN104679945B (en) System comprehensive estimation method based on colored Petri network
CN105468703B (en) A kind of reason retroactive method
CN106845643A (en) A kind of cloud service system dynamic evolution method based on improved adaptive GA-IAGA
CN105991571B (en) A kind of information processing method and device
JP6045388B2 (en) Operation management device, operation management method, and program
CN107230020A (en) Evaluate and test the method and its correlation technique and system of high ferro dispatcher's job engineering efficiency
CN107437126A (en) A kind of logistics transportation dispatching method and device
CN110365507B (en) Method and device for realizing automatic link weight assignment
CN105809358A (en) Network management performance statement report requirement support modeling method based on data driving

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20190419

Termination date: 20210209

CF01 Termination of patent right due to non-payment of annual fee