CN109167806A - A kind of uncertain QoS perception web service selection method based on prospect theory - Google Patents
A kind of uncertain QoS perception web service selection method based on prospect theory Download PDFInfo
- Publication number
- CN109167806A CN109167806A CN201810784014.5A CN201810784014A CN109167806A CN 109167806 A CN109167806 A CN 109167806A CN 201810784014 A CN201810784014 A CN 201810784014A CN 109167806 A CN109167806 A CN 109167806A
- Authority
- CN
- China
- Prior art keywords
- service
- user
- value
- qos
- prospect
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000010187 selection method Methods 0.000 title claims abstract description 13
- 230000008447 perception Effects 0.000 title description 8
- 239000011159 matrix material Substances 0.000 claims abstract description 44
- 238000000034 method Methods 0.000 claims abstract description 36
- 230000015572 biosynthetic process Effects 0.000 claims abstract description 11
- 238000003786 synthesis reaction Methods 0.000 claims abstract description 11
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000005315 distribution function Methods 0.000 claims description 3
- 238000013210 evaluation model Methods 0.000 claims description 3
- 206010063659 Aversion Diseases 0.000 claims description 2
- 238000012360 testing method Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 241000208340 Araliaceae Species 0.000 description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 235000008434 ginseng Nutrition 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000035699 permeability Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 210000003813 thumb Anatomy 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/70—Admission control; Resource allocation
- H04L47/80—Actions related to the user profile or the type of traffic
- H04L47/805—QOS or priority aware
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/02—Protocols based on web technology, e.g. hypertext transfer protocol [HTTP]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/51—Discovery or management thereof, e.g. service location protocol [SLP] or web services
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of, and the uncertain QoS based on prospect theory perceives web service selection method.The QoS attribute expectation provided first according to user is used as reference point, obtains income and loss of the QoS attribute relative to reference point of each service；Then risk income matrix and risk of loss matrix are established respectively, for each QoS attribute of each service, income and lost value, the income and loss probability weight for obtaining each QoS attribute are calculated separately according to prospect theory, and obtain the synthesis prospect value of each service；Finally service is ranked up according to the size of comprehensive prospect value, obtains optimal service.The present invention comprehensively considers user demand, and user realizes and selects in the Web service of QoS uncertain condition to income, the attitudes toward risk of loss and user's attitudes toward risk large and small to the probability of QoS attribute value.Method embodies the expectation demand of user, carries out Web service selection under QoS uncertain condition for user and provides a kind of new method.
Description
Technical field
The present invention relates to a kind of, and the uncertain QoS based on prospect theory perceives web service selection method, belongs to service choosing
Select field.
Background technique
With the development of information technology, internet has become the important information infrastructure of modern society, SOA
(Service-Oriented Architecture), Services Oriented Achitecture solve in internet environment as a kind of
Shared, service reuse and business integration problem distributed software system framework is serviced, it is extensive by industry and academia
Receive.In particular with the appearance and popularization of Web service technology, Web service has become the mainstream technology of SOA, greatly pushes
SOA applications in various fields.However, with the rapid development of Web service technology, on internet quantity of service and type also with
Increase, occur it is many there is Web service functionally identical or similar, user not only needs to examine in the selection course of service
Consider the functional attributes of Web service, it is also desirable to consider its nonfunctional space.Therefore, based on QoS (Quality of Service,
Service quality) perception (QoS-aware) Web service selection proposition with development be the key that then to solve the problems, such as this.Web
The QoS of service is one group of nonfunctional space set of service, such as response time, price and reliability, the QoS of service are faces
The QoS guarantee of the indispensable element into the architecture of service, service will become judgement ISP's energy
The key factor of no success.Due to the dynamic and opening of running environment, some QoS indexes are by change of network environment etc.
The influence of many extraneous factors, and the complexity and ISP's scheduling of resource of service user's application context environment
The features such as flexibility, the qos value for servicing Web can change, so that Web service QoS has uncertainty.I.e. by not
Determine that factor influences, QoS attribute value table reveals certain fluctuation or unstable situation.Because being indicated in the form of value by determining
QoS attribute can not accurate description Web service by open environment influence degree and accurately reflect the actual performance of service,
It is unfavorable for the selection of best candidate service, therefore the proposition and realization of the web service selection method based on uncertain QoS perception
It is most important.
Currently, the method for the Web service selection of uncertain QoS perception mainly has: the (< The 4th IEEE such as Cheng Wan
International Conference on Services Computing >, 2007:154-161) propose one kind to uncertain
The description method of QoS sets up a weight for discrete uncertain QoS attribute, and every record in Web service is calculated
The sum of weighted Q oS value be then ranked up, checked using U or kurtosis and skewness return to the lesser Web service of numerical value, so
After carry out services selection；(< The 9th International Conference on the Services such as Karim Benouare
Computing >, 2012:523-530) it proposes to apply to Skyline inquiry method in uncertain QoS, and propose not needing
The weight of user's offer QoS；(< Computers and Mathematics the with such as Chonghai Wang
Applications >, 2011,62 (7): 2812-2823) it proposes uncertain QoS attribute being expressed as continuous type, then carry out
Selection, three of the above method all only consider the QoS attribute description mode under uncertain condition, do not consider that user services in selection
When demand and selection to QoS attribute when attitudes toward risk.(< The 27th Chinese the Control such as Xiaodong FU
Conference >, 2008,271-275) it proposes the possibility degree method compared using interval number and approaches more attributes of ideal point
Decision-making technique establishes the Web service preference pattern based on uncertain QoS information, but QoS is considered as and is uniformly distributed by this method
Stochastic variable cannot fully demonstrate the stochastic behaviour of variety classes QoS.Fan little Qin, Jiang Changjun etc. (<Journal of Software>, 2009,
20 (3): 546-556) value of each index is measured using the mathematic expectaion and variance of stochastic variable, propose Web service respectively with
The measure of machine QoS index, but this method assumes that each QoS attribute has fixed distribution form, which is not inconsistent simultaneously
QoS diversity is closed, and the algorithm studied does not account for user for the demand of QoS, therefore the solution obtained is different
Surely meet the needs of users.(< The 8th IEEE International Symposium the on such as W.Wiesemann
Cluster Computing and the Grid >, 2008:226-233) plan model is devised, utilize venture worth
To response time and cost can the risk of the loss of energy quantify, to consider QoS randomness bring risk problem, still
Only assume that the risk measurement carried out under the premise of QoS has specific distribution is just reasonable, but not due to the probability distribution of QoS
Certain satisfaction is uniformly distributed equal distribution form, therefore all QoS attributes are all assumed that it is unreasonable for having specific distribution.On
Method is stated to a certain extent and can realize the selection to Web service, but limitation is also evident from.For example, based on pair
Uncertain QoS attribute sets up weight, and the sum of weighted Q oS value is calculated and then is ranked up the method for obtaining optimal service,
Attitudes toward risk when demand and user of the user to QoS attribute select service is had not focused on, leads to the service of final choice not
User demand can be met to the full extent and make the service risk of selection minimum.
The present invention will carry out Web service and brought risk is selected mutually to tie with the demand of user under QoS uncertain condition
It closes, proposes a kind of uncertain QoS perception web service selection method based on prospect theory, it is expected to make with the attribute that user provides
For reference point, risk income matrix and risk of loss matrix of each service QoS attribute relative to reference point are established, according to prospect
Theory obtains comprehensive prospect value, is then ranked up according to comprehensive prospect value size to service, selects and meet user demand most
Excellent service.Method is from user perspective, and QoS risk when not knowing in conjunction with QoS may be selected to be best suitable for user demand and risk
The smallest Web service.
Summary of the invention
The present invention provides a kind of uncertain QoS perception web service selection method based on prospect theory, for solving public affairs
Do not consider the problems of that user it is expected demand and attitudes toward risk in perception method under QoS uncertain condition.
The technical scheme is that a kind of uncertain QoS based on prospect theory perceives web service selection method, it is first
Reference point is first used as according to QoS attribute expectation of the user to service, obtains receipts of the QoS attribute of each service relative to reference point
Benefit and loss, establish risk income matrix and risk of loss matrix respectively；Secondly it for each QoS attribute of each service, examines
Consider attitudes toward risk of the user to the attitudes toward risk of income and loss and user to the probability size of QoS attribute value, according to
Prospect theory calculates separately the earned value for obtaining each QoS attribute, lost value, income probability right and loss probability weight,
And obtain the synthesis prospect value of each service；Finally service is ranked up according to the size of comprehensive prospect value, obtains optimal clothes
Business.
Specific step is as follows:
Step 1, the expectation according to candidate service set and user to service QoS attribute, calculation risk gain matrix and
Risk of loss matrix
1.1 notes M={ 1,2 ..., m }, N={ 1,2 ..., n }, H={ 1,2 ..., h }.If candidate service collection is combined into S
={ s1,s2,...,si,...,sm(i ∈ M), the QoS attribute set A={ a of service1,a2,....aj,...,an(j ∈ N), A'
And A " respectively indicates cost type attribute set and profit evaluation model attribute set, A' ∈ A, A " ∈ A, and A' ∪ A "=A,Mean Vector R=(r of the user to QoS attribute1,r2,...,rj,...rn)(j∈N).Remember F={ f1 j,
f2 j,...,fh jIndicate service QoS attribute ajValue state set, wherein ft jIndicate service QoS attribute ajT kind shape
State, t ∈ H.
Expectation r of the user to service QoS attributejAcquisition be divided into two kinds of situations:
(a) it when service, QoS attribute number are less, is directly provided by user to the expectation of service QoS attribute and value；
(b) when service, QoS attribute number are more, quantile (the two quantiles or quartile etc.) service of obtaining can be used
The desired value of QoS attribute.Select two quantiles as the expectation of service QoS attribute and value in the present invention, i.e., in t kind
Under state, service s is takeniQoS attribute be respectively worth in median as service QoS attribute expectation rj。
1.2 it is expected r by the service QoS attribute that step 1.1 obtainsjAs reference point, obtain each QoS attribute relative to ginseng
According to the income and loss of point.
Wherein rj=(rj 1,rj 2,...rj t,...rj h) it is user to service siQoS attribute ajExpectation, rj tIt indicates
User is to service s under t kind stateiQoS attribute ajExpectation value, j ∈ N, t ∈ H；aij tIt indicates under t kind state
Service siQoS attribute ajValue.
(a) work as aij t≥rj tWhen, then attribute value aij tRelative to reference point rj tIncome Gij tCalculation formula be
Lose Lij tFor
(b) work as aij t< rj tWhen, attribute value aij tRelative to reference point rj tIncome Gij tCalculation formula be
Lose Lij tFor
To sum up, income, the loss that service QoS attribute is obtained according to formula (1)-(4), can establish risk income square respectively
Battle array Gt=[Gij t]m×nWith risk of loss matrix Lt=[Lij t]|m×n。
Step 2, the risk income matrix G established according to step 1tWith risk of loss matrix Lt, consider user to income and
The different attitudes toward risks of loss calculate separately the prospect value of each QoS attribute in each service, construct prospect decision matrix and obtain
To the synthesis prospect value of each service.
2.1, according to prospect theory, calculate the income and lost value of each service QoS attribute.It is calculated and is received according to formula (5)
Beneficial Gij tValue Vij (+)t, formula (6), which calculates, loses Lij tValue Vij (-)t。
Vij (+)t=(Gij t)α (5)
Vij (-)t=-λ (- Lij t)β (6)
Wherein, the parameter alpha in formula (5) and formula (6) is the risk goal function of user, and β is the risk aversion of user
Coefficient reflects user to the different attitudes toward risks of income and loss respectively, and 0 < α <, 1,0 < β < 1, α and β are bigger, indicate to use
Family is more intended to take a risk；λ indicates that degree is evaded in the loss of user, and λ > 1, λ is bigger, indicates that the loss of user is evaded degree and got over
Greatly.The value of parameter alpha, β and λ is set by the method that the presenter Kahneman and Tversky of prospect theory are proposed.
2.2 calculate under t kind state, service siEach QoS attribute value probabilityRule of thumb distribution function (7) calculates probability.
Wherein*{aij 1,aij 2,...aij tIn indicate (aij 1,aij 2,...aij t) in be not more than aij tNumber.
2.3 calculate the income and loss probability weight of each service QoS attribute.Income G is calculated according to formula (8)ij t's
Probability right πij (+)t, formula (9), which calculates, loses Lij tProbability right πij (-)t.Probability right πij tIt is to assign Probability pj tOne
Weight.
Wherein formula (8) and γ in formula (9) and δ reflect that user treats the difference of yield risk and loss risk respectively
Attitude, 0 < γ <, 1,0 < δ < 1.When one timing of γ value, with the reduction of δ value, the weight assigned to loss can be higher, performance
To over-evaluate small probability event, attitudes toward risk is risk partiality；When the timing of δ value one, with the reduction of γ value, show as underestimating in,
Great possibility, attitudes toward risk are risk averse.The value of parameter γ and δ are by prospect theory presenter Kahneman and Tversky
The method of proposition is set.
The 2.4 service QoS attribute values and probability right being calculated according to step 2.1 and step 2.3 calculate each
Service siAttribute ajProspect value Vij:
The 2.5 prospect values being calculated according to step 2.4 construct prospect decision matrix V=[Vij]m×n。
Step 3, the prospect decision matrix V constructed according to step 2.4, calculate standardization prospect matrix V*:
Wherein VjMax=max | Vij|} (12)
Step 4, the attribute weight ω provided according to simple weighted principle and userj, before the synthesis for calculating each service
Scape value Ui, its calculation formula is
(a) when service, QoS attribute number are less, the weight to Service Properties is directly provided by user；
(b) when service, QoS attribute number are more, service is assigned using the methods of subjective weighting method, objective weighted model and is belonged to
The weight of property.
According to comprehensive prospect value UiSize all services are ranked up, UiIt is bigger, service siBetter.
The beneficial effects of the present invention are:
1, the Web service selection towards uncertain QoS
Due to the opening and dynamic of running environment, so that Web service QoS has inherent uncertainty.Because with
It determines that the QoS of value form expression cannot accurately reflect the actual performance of service, is unfavorable for selecting optimal service, to make
Obtaining user, there are the unappeasable risks of demand when selecting service.Therefore, method of the present invention considers QoS information
Uncertainty selects the optimal service for meeting user demand under QoS uncertain condition from numerous Web services.
2, prospect of the application theory realizes Web service selection
The present invention realizes the Web service selection under QoS uncertain condition based on prospect theory.Prospect theory phase
Have the advantages that itself is peculiar for other methods in known method.Demand of the user to QoS attribute is considered first, according to user
The QoS attribute expectation of offer is used as reference point, obtains income and loss of each service QoS attribute relative to reference point, then divides
Risk income matrix and risk of loss matrix are not established；Consider that user treats the attitude of risk, for each of each service
QoS attribute, calculated separately according to the attitudes toward risk of prospect theory and user the income for obtaining each QoS attribute and lost value,
Income and loss probability weight, and obtain the synthesis prospect value of each service；Finally selection obtains optimal service.
3, realize that Web service selection effectively embodies user demand, attitudes toward risk by prospect theory
Due to the complexity of user demand, so that the Web service selection under QoS uncertain condition becomes difficult.So
And known method not operatively demand by user to Web service QoS attribute, the risk state of user when selecting service
These condition elements are spent to comprehensively consider.Therefore the method for the invention is using user demand and consumer's risk attitude as necessary item
Part selects the Web service that can be met the needs of users to the full extent from user perspective.It is embodied in: not true in QoS
Determine in situation, considers demand of the user to service QoS attribute, the QoS attribute that user is provided is as reference point；In QoS attribute
Income, lost value calculate in, in the form of parameter alpha and β embody user to income, the attitudes toward risk of loss；Belong in QoS
Property income, loss probability weight calculating in, user is embodied in the form of parameter γ and δ to the probability of QoS attribute value
The attitudes toward risk of large and small probability.
To sum up, the uncertain QoS of the present invention based on prospect theory perceives web service selection method, to solve
Web service select permeability under QoS uncertain condition provides a kind of new approaches.Method fully takes into account user and selects Web service
Expectation demand and user attitudes toward risk, obtained services selection result can meet to the full extent user demand simultaneously
And risk when servicing selection is preferably minimized, therefore user can be assisted preferably to carry out services selection.
Detailed description of the invention
Fig. 1 is method flow diagram in the present invention.
Specific implementation method
The present invention will be further explained below with reference to the attached drawings and specific examples.
As shown in Figure 1, step 1, the expectation according to candidate service set and user to service QoS attribute, calculation risk
Gain matrix and risk of loss matrix
1.1 5 service samples of selection, candidate service collection are combined into S={ s1,s2,s3,s4,s5, QoS attribute set is A=
{a1,a2, each service and service QoS attribute set are as shown in table 1.Attribute a1(response time) is cost type attribute, attribute a2
(throughput) is profit evaluation model QoS attribute.
The service of table 1 and service QoS attribute
Continued
User is R={ r to the Mean Vector of service QoS attribute1,r2, each QoS attribute has 5 attribute value states.
Use two quantiles as the service desired value of QoS attribute.
User is desired for r to each QoS attribute1=(r1 1,r1 2,r1 3,r1 4,r1 5) and r2=(r2 1,r2 2,r2 3,r2 4,
r2 5).It services the expectation of QoS attribute and value is as shown in table 2.
Table 2 services the expectation of QoS attribute and value
1.2 it is expected r by the service QoS attribute that step 1.1 obtainsjAs reference point, obtain each QoS attribute relative to ginseng
According to the income and loss of point.
It is obtained according to formula (1)-(4) according to the reference point for servicing QoS attribute in the service QoS attribute and table 2 in table 1
To the income of service QoS attribute and loss and establish risk income matrix GtWith risk of loss matrix Lt, respectively such as table 3 and table 4
It is shown.
3 risk income matrix G of tablet
| 0 | 0 | 0.016 | 0.061 | 0.066 | 0.704 | 1.149 | 1.845 | 1.2483 | 0.7152 |
| 0 | 0.001 | 0 | 0 | 0 | 1.783 | 1.337 | 0 | 1.196 | 1.3706 |
| 0.09 | 0 | 0.026 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 0.082 | 0.024 | 0 | 0.0399 | 0.064 | 0 | 0 | 0.3849 | 0 | 0 |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
4 risk of loss matrix L of tablet
| -0.012 | -0.045 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 0 | 0 | -0.025 | -2.1 | 0 | 0 | 0 | -0.667 | 0 | 0 |
| 0 | 0 | 0 | 0 | -0.003 | -0.397 | -0.101 | 0 | -0.469 | -0.339 |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| -0.039 | -0.117 | -0.109 | -0.064 | -0.044 | -0.8064 | -0.46 | -0.193 | -0.5868 | -0.556 |
Step 2, the risk income matrix G established according to step 1tWith risk of loss matrix Lt, consider user to income and
The different attitudes toward risks of loss calculate separately the prospect value of each QoS attribute in each service, construct prospect decision matrix
About the parameter in following formula, it should be noted that prospect theory presenter Tversky etc. is by a large amount of
Decision-making entity carries out experiment test, and carries out regression analysis to obtained data, obtains and takes with the most consistent parameter of experimental result
Value α=β=0.88, λ=2.25, γ=0.61, δ=0.69, these values be considered as the person that can indicate aritrary decision substantially
The parameter value of Behavior preference.Abdellaoui and Xu etc. also by experiment parameter value problem is studied, obtained with
Parameter value similar in above-mentioned value.
2.1, according to prospect theory, calculate the income and lost value of each service QoS attribute.Income is calculated according to formula (5)
Gij tValue Vij (+)t, formula (6), which calculates, loses Lij tValue Vij (-)t, respectively as shown in table 5 and table 6.
5 income G of tableij tValue Vij (+)t
| 0 | 0 | 0.02628 | 0.0853 | 0.0915 | 0.7347 | 1.1304 | 1.7139 | 1.2155 | 0.7446 |
| 0 | 0.0023 | 0 | 0 | 0 | 1.6633 | 1.2914 | 0 | 1.1706 | 1.3197 |
| 0.1202 | 0 | 0.04029 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 0.1107 | 0.0375 | 0 | 0.0587 | 0.089 | 0 | 0 | 0.4316 | 0 | 0 |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
Table 6 loses Lij tValue Vij (-)t
| -0.046 | -0.147 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 0 | 0 | -0.0876 | -4.323 | 0 | 0 | 0 | -1.575 | 0 | 0 |
| 0 | 0 | 0 | 0 | -0.014 | -0.998 | -0.298 | 0 | -1.156 | -0.869 |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| -0.13 | -0.341 | -0.32 | -0.2 | -0.144 | -1.861 | -1.135 | -0.529 | -1.408 | -1.344 |
2.2 calculate under 5 kinds of states, service s1-s5Each QoS attribute value probability.
Empirically distribution function formula (7) calculates under 5 kinds of states, the probability p of each service QoS attributej t, such as
Shown in table 7.
7 Service Properties probability p of tablej t
| 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 |
| 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 |
| 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 |
| 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 |
| 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 | 0.2 |
2.3 calculate the income and loss probability weight of each service QoS attribute.Income G is calculated according to formula (8)ij tIt is general
Rate weight πij (+)t, formula (9), which calculates, loses Lij tProbability right πij (-)t, respectively as shown in table 8 and table 9.
8 income G of tableij tProbability right πij (+)t
| 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 |
| 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 |
| 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 |
| 0.3507 | 0.3507 | 0.3507 | 0.5692 | 0.5692 | 0.3507 | 0.3507 | 0.5692 | 0.5692 | 0.3507 |
| 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 | 0.3507 |
Table 9 loses Lij tProbability right πij (-)t
| 0.2997 | 0.2997 | 0.29967 | 0.2997 | 0.2997 | 0.2997 | 0.2997 | 0.2997 | 0.2997 | 0.2997 |
| 0.2997 | 0.2997 | 0.29967 | 0.2997 | 0.2997 | 0.2997 | 0.2997 | 0.2997 | 0.2997 | 0.2997 |
| 0.2997 | 0.2997 | 0.29967 | 0.5124 | 0.5124 | 0.2997 | 0.2997 | 0.5124 | 0.5124 | 0.2997 |
| 0.2997 | 0.2997 | 0.29967 | 0.2997 | 0.2997 | 0.2997 | 0.2997 | 0.2997 | 0.2997 | 0.2997 |
| 0.2997 | 0.2997 | 0.29967 | 0.2997 | 0.2997 | 0.2997 | 0.2997 | 0.2997 | 0.2997 | 0.2997 |
The 2.4 service QoS attribute values and probability right being calculated according to step 2.1 and step 2.3, according to formula
(10) each service s is calculatediAttribute ajProspect value Vij。
The 2.5 prospect values being calculated according to step 2.4 construct prospect decision matrix V=[Vij]m×n, such as 10 institute of table
Show.
10 prospect decision matrix V of table
| a1 | a2 | |
| s1 | 0.0134 | 1.9426 |
| s2 | -1.3208 | 1.4375 |
| s3 | 0.0522 | -0.9951 |
| s4 | 0.1038 | 0.1514 |
| s5 | -0.3399 | -1.881 |
Step 3, the prospect decision matrix V constructed according to step 2.5, establish standardization prospect matrix V after standardization*
According to formula (11)-(12), according to prospect decision matrix, standardization prospect matrix is established after standardization, such as table 11
It is shown.
The standardization prospect matrix V of table 11*
Step 4, the attribute weight vector ω provided according to simple weighted principle and userj, according to formula (13), meter
Calculate the synthesis prospect value of each service
The synthesis prospect value of each service is calculated are as follows: U1=0.5051, U2=-0.13, U3=-0.2364, U4=
0.0783, U5=-0.6128.Result is ranked up according to the size of the synthesis prospect value of each service are as follows: U1 >U4 >U2 >U3 >U5,
The ranking results for finally obtaining each service are s1f s4f s2f s3f s5.According to ranking results, selection service s1For optimal clothes
Business.
It is s with the optimal service that above-mentioned sample is calculated in expected utility theory, mean variance model4、s3.It utilizes
Greatest hope value of utility decision rule (EMV), minimum opportunity loss decision rule (EOL) and maximum utility value decision rule
(EUV) to service s4、s3Efficiency assessment is carried out, obtains s1f s4f s3.Therefore the optimal service that prospect theory obtains is better than it
The optimal service of remaining two methods selection.EMV and EOL is by analyzing (or the opportunity loss of the probability that is likely to occur and profit or loss
Value) calculate expected revenus value (or expected shortfall) to carry out service and compare and select；EUV lays particular emphasis on the implicit of service
Value or preference, and selected after giving quantization by utility index.
Now select 339 services as test sample, if candidate service set S={ s1,s2,...,s339It is test set
It closes, by above-mentioned method to each service s in test set SiCorresponding comprehensive prospect value is calculated, then basis
The size of comprehensive prospect value is ranked up, and obtains optimal service.
Set S={ s will be tested1,s2,...,s339It is divided into 30 groups of T1,T2,...,T30, managed with prospect theory, expected utility
30 groups of optimal services: T are obtained by, three kinds of methods of mean variance model1:s10、s1、s5；T2:s14、s2、s19；...；T30:s25、
s11、s28.Using EMV, EOL and EUV criterion, by the optimal service obtained by prospect theory respectively with by expected utility theory,
The optimal service that mean variance model obtains compares two-by-two, and statistics meets above three by the optimal service that prospect theory obtains
The number of criterion carries out efficiency assessment, show that the efficient optimal service obtained during the test by prospect theory is excellent
In the ratio of optimal service of remaining two methods selection be respectively 80%, 93.3%.
Prospect theory and expected utility theory, mean variance model are obtained in test set S based on above implementation steps
On comparison, can learn that prospect theory is more suitable for not knowing than expected utility theory, mean variance model by comparing
Under the conditions of Web service selection；Side proposed by the invention is sufficiently illustrated by the comparison of three criterion indexs in embodiment
The validity of method.
The above is only implementation method of the invention, but the present invention is not limited to above-described implementation method,
This field has the personnel of the relevant technologies, can make a variety of changes based on the present invention and under the premise of not departing from objective, this
Kind variation also should be regarded as protection scope of the present invention.
Claims (4)
1. a kind of uncertain QoS based on prospect theory perceives web service selection method, it is characterised in that: first according to user
Reference point is used as to the QoS attribute expectation of service, obtains income and loss of the QoS attribute relative to reference point of each service, point
Risk income matrix and risk of loss matrix are not established；Secondly for each QoS attribute of each service, consider user to income
With the attitudes toward risk of loss and user to the attitudes toward risk of the probability size of QoS attribute value, distinguished according to prospect theory
Earned value, lost value, income probability right and the loss probability weight of each QoS attribute is calculated, and obtains each clothes
The synthesis prospect value of business；Finally service is ranked up according to the size of comprehensive prospect value, obtains optimal service.
2. the uncertain QoS according to claim 1 based on prospect theory perceives web service selection method, feature exists
In: specific step is as follows for the method:
Expectation of the step 1 according to candidate service set and user to service QoS attribute, calculation risk gain matrix and risk damage
Lose matrix；
Remember M={ 1,2 ..., m }, N={ 1,2 ..., n }, H={ 1,2 ..., h }, if candidate service collection is combined into S={ s1,
s2,...,si,...,sm(i ∈ M), the QoS attribute set serviced in candidate service set S is A={ a1,a2,....aj,...,
an(j ∈ N), A' and A " respectively indicate cost type attribute set and profit evaluation model attribute set, A' ∈ A, A " ∈ A, and A' ∪ A "=
A,User is R=(r to the Mean Vector of QoS attribute1,r2,...,rj,...rn) (j ∈ N), F={ f1 j,
f2 j,...ft j,...,fh jIndicate QoS attribute ajValue state set, wherein ft jIndicate QoS attribute ajT kind state, t
∈H；
Wherein rj=(rj 1,rj 2,...rj t,...rj h) it is user to service siQoS attribute ajExpectation, rj tIt indicates in t kind
User is to service s under stateiQoS attribute ajExpectation value, j ∈ N, t ∈ H；aij tExpression services s under t kind statei's
QoS attribute ajValue；With user to service siQoS attribute ajExpectation rjAs reference point, each QoS attribute a is obtainedjRelatively
In the income and loss of reference point；
(a) work as aij t≥rj tWhen, then attribute value aij tRelative to reference point rj tIncome Gij tFor
Lose Lij tFor
(b) work as aij t< rj tWhen, attribute value aij tRelative to reference point rj tIncome Gij tFor
Lose Lij tFor
According to above-mentioned income Gij tWith loss Lij tRisk income matrix G is established respectivelyt=[Gij t]m×nWith risk of loss matrix Lt=
[Lij t]|m×n；
Step 2, the risk income matrix G established according to step 1tWith risk of loss matrix Lt, consider user to income and loss
Different attitudes toward risks calculate separately the prospect value of each QoS attribute in each service, construct prospect decision matrix and obtain each clothes
The synthesis prospect value of business；
According to prospect theory, service s is calculatediQoS attribute ajIncome Gij tValue Vij (+)tWith loss Lij tValue Vij (-)t
Vij (+)t=(Gij t)α
Vij (-)t=-λ (- Lij t)β
Wherein, α is the risk goal function of user, and β is the risk-aversion coefficient of user, reflects user to income and loss respectively
Different attitudes toward risks, 0 < α <, 1,0 < β < 1；λ indicates that degree, λ > 1 are evaded in the loss of user；
It calculates under t kind state, services siQoS attribute value probability
Rule of thumb distribution function calculates probabilityWherein*{aij 1,aij 2,...aij tIndicate
(aij 1,aij 2,...aij t) in be not more than aij tNumber；
Calculate service siQoS attribute ajIncome probability right πij (+)tWith loss probability weight πij (-)t
Wherein γ and δ reflects that user treats the attitude of yield risk and loss risk, 0 < γ <, 1,0 < δ < 1 respectively；
Calculate service siAttribute ajProspect valueAccording to prospect value VijStructure
Build prospect decision matrix V=[Vij]m×n, obtain standardization prospect matrixWherein VjMax=max | Vij|, according to
According to simple weighted principle and Service Properties weights omegaj, calculate the synthesis prospect value of each service
Step 3 is according to comprehensive prospect value UiSize all services are ranked up, UiIt is bigger, service siBetter.
3. the uncertain QoS according to claim 2 based on prospect theory perceives web service selection method, feature exists
In: the user is to service siQoS attribute ajExpectation rjIt is directly provided by user or is selected two quantiles, i.e., in t kind
Under state, service s is takeniQoS attribute be respectively worth in median be used as desired rj。
4. the uncertain QoS according to claim 2 based on prospect theory perceives web service selection method, feature exists
In: the Service Properties weights omegajIt is directly provided by user or carries out assignment using subjective weighting method, objective weighted model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810784014.5A CN109167806B (en) | 2018-07-17 | 2018-07-17 | Uncertain QoS (quality of service) perception Web service selection method based on foreground theory |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810784014.5A CN109167806B (en) | 2018-07-17 | 2018-07-17 | Uncertain QoS (quality of service) perception Web service selection method based on foreground theory |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109167806A true CN109167806A (en) | 2019-01-08 |
| CN109167806B CN109167806B (en) | 2021-04-09 |
Family
ID=64897697
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810784014.5A Active CN109167806B (en) | 2018-07-17 | 2018-07-17 | Uncertain QoS (quality of service) perception Web service selection method based on foreground theory |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109167806B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111325648A (en) * | 2020-02-20 | 2020-06-23 | 杭州电子科技大学 | Aging service combination optimization method and system based on hesitation ambiguity |
| CN114760585A (en) * | 2022-04-18 | 2022-07-15 | 中南大学 | Vehicle crowd sensing excitation method, system and equipment |
| CN115150405A (en) * | 2022-07-11 | 2022-10-04 | 天津理工大学 | Foreground theory framework-based edge calculation task unloading method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150161733A1 (en) * | 2013-09-25 | 2015-06-11 | Fmr Llc | Life Cycle Based Portfolio Construction Platform Apparatuses, Methods and Systems |
| CN107784394A (en) * | 2017-10-30 | 2018-03-09 | 长安大学 | Consider that the highway route plan of prospect theory does not know more attribute method for optimizing |
-
2018
- 2018-07-17 CN CN201810784014.5A patent/CN109167806B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150161733A1 (en) * | 2013-09-25 | 2015-06-11 | Fmr Llc | Life Cycle Based Portfolio Construction Platform Apparatuses, Methods and Systems |
| CN107784394A (en) * | 2017-10-30 | 2018-03-09 | 长安大学 | Consider that the highway route plan of prospect theory does not know more attribute method for optimizing |
Non-Patent Citations (2)
| Title |
|---|
| 付晓东，等: "QoS信息不确定情况下的Web服务选择", 《PROCEEDINGS OF THE 27TH CHINESE CONTROL CONFERENCE》 * |
| 任剑: "模糊环境下信息不完全的随机多准则决策方法研究", 《中国博士学位论文全文数据库（电子期刊）》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111325648A (en) * | 2020-02-20 | 2020-06-23 | 杭州电子科技大学 | Aging service combination optimization method and system based on hesitation ambiguity |
| CN114760585A (en) * | 2022-04-18 | 2022-07-15 | 中南大学 | Vehicle crowd sensing excitation method, system and equipment |
| CN114760585B (en) * | 2022-04-18 | 2024-04-16 | 中南大学 | Method, system and equipment for intelligent perception excitation of vehicle group |
| CN115150405A (en) * | 2022-07-11 | 2022-10-04 | 天津理工大学 | Foreground theory framework-based edge calculation task unloading method |
| CN115150405B (en) * | 2022-07-11 | 2023-08-01 | 天津理工大学 | Edge computing task unloading method based on foreground theoretical framework |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109167806B (en) | 2021-04-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109543737B (en) | Information system health degree evaluation method based on FAHP-FCA combined empowerment | |
| Qu et al. | CCCloud: Context-aware and credible cloud service selection based on subjective assessment and objective assessment | |
| CN109167806A (en) | A kind of uncertain QoS perception web service selection method based on prospect theory | |
| CN108399453A (en) | A kind of Electric Power Customer Credit Rank Appraisal method and apparatus | |
| CN103888543B (en) | Medical resource recommendation method and system based on Web services | |
| CN104850727A (en) | Distributed big data system risk evaluation method based on cloud barycenter theory | |
| CN111898839B (en) | Importance degree classification method and device for power users | |
| CN108171369A (en) | Short term combination forecasting method based on customer electricity differentiation characteristic | |
| CN111695830A (en) | Power resource allocation method, system and equipment | |
| CN106447403A (en) | User priority classification method in large-user direct power purchase environment | |
| CN106296315A (en) | Context aware systems based on user power utilization data | |
| Wang et al. | An unsupervised strategy for defending against multifarious reputation attacks | |
| CN114036570A (en) | Dynamic evaluation method and system for information sensitivity of smart power grid | |
| CN101217427B (en) | A network service evaluation and optimization method under uncertain network environments | |
| Yiping et al. | An improved multi-view collaborative fuzzy C-means clustering algorithm and its application in overseas oil and gas exploration | |
| Mahan et al. | A novel resource productivity based on granular neural network in cloud computing | |
| CN111932106A (en) | Effective and practical cloud manufacturing task and service resource matching method | |
| CN115525230A (en) | Storage resource allocation method and device, storage medium and electronic equipment | |
| US20220129318A1 (en) | Quota Request Resolution On Computing Platform | |
| CN110675069A (en) | Real estate industry client signing risk early warning method, server and storage medium | |
| CN116645197B (en) | Consumer financial product service construction method based on distributed automation | |
| Cui et al. | Customer churn prediction using improved FCM algorithm | |
| CN114065109A (en) | Optimal selection method for power system demand response aggregation users | |
| CN104778636B (en) | Method is recommended in the information service of apparel maintenance support | |
| Cabrerizo et al. | An adaptive feedback mechanism for consensus reaching processes based on individuals’ credibility |
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 |