CN108170404A - A kind of Web service combination verification method based on parameterized model - Google Patents
A kind of Web service combination verification method based on parameterized model Download PDFInfo
- Publication number
- CN108170404A CN108170404A CN201711420311.3A CN201711420311A CN108170404A CN 108170404 A CN108170404 A CN 108170404A CN 201711420311 A CN201711420311 A CN 201711420311A CN 108170404 A CN108170404 A CN 108170404A
- Authority
- CN
- China
- Prior art keywords
- model
- web service
- service combination
- qos
- decision process
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/10—Requirements analysis; Specification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/20—Software design
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of Web service combination verification method based on parameterized model, including：Web service combination process and QoS attributes are taken out, determines the component part of QoS attributes；The environmental condition of analyzing influence QoS attributes, and the Markovian decision process that the environmental condition is modeled as parameterizing；Limitation requirement constraint Web service combination process is introduced, and the process model building is become into the parametrization Markovian decision process with limitation requirement；QoS model attributes to be verified are become into temporal logic formula；Using can processing parameter model probabilistic model checking tool, verify whether finite state model meets QoS attributes to be verified in the environment of dynamic, and obtain quantitative verification result.The method of the present invention is solved under complicated open environment, the problem of Traditional Web services combined authentication method caused by the dependence between randomness, abstract service environment dynamic change is inaccurate, and effectively alleviate the state explosion problem in verification process.
Description
Technical field
The invention belongs to computer software engineering development technique field, more particularly to a kind of Web based on parameterized model
Services composition verification method.
Background technology
Web service combination is the skill different specific services to combine to complete composite services with better function
Art.With the fast development of Web service technology, functional attributes are similar and service quality (Quality of Service, i.e. QoS)
Different specific service quantity rapidly increase.In the case, according to service quality for user recommend different specific service with
So that the optimal method of the global qos value of entire Web service combination is referred to as the web service composition method of QoS perception.In addition,
Environment is dynamic change, and when environmental change, the qos value of specific service can also change therewith.Therefore, how in dynamic change
Environment in selection service so that overall situation QoS it is optimal, be under open environment QoS perception Web service combination research in it is important
Problem.
For probabilistic model checking technology as a kind of formalization verification method, its object is to finite state model is used to describe
Given probability system describes attribute to be verified using temporal logic formula, and finite state model and sequential logic is public
Whether input of the formula as model checking tools, verification probability system meet attribute to be verified, and quantitative result.
PRISM has probability behavior and random sexual behaviour as a probabilistic model checking tool being widely used available for analysis
System and carry out association attributes verification.Probabilistic model checking technology is applied to the related side of Web service combination research above
Method is known as Web service combination verification method.
PARAM is the tool used in this method, it be it is a can be with the probabilistic model checking work of processing parameter model
Tool is the expansion of PRISM.Difference lies in PARAM can be with processing parameter model by PARAM and PRISM.So with PRISM
Equally, PARAM is a tool for being used for analyzing probability system, can support the model of three types, discrete time Ma Erke
Husband's chain, the Markov chain and Markovian decision process of continuous time.The tool is by automatically analyzing established parameter
Change model and temporal logic formula, the rational function containing multiple variables will be obtained, these variables are exactly the quilt in modeling process
The variable of parametrization.The value of the variable is given according to domain, and brings rational function calculating into, so that it may attribute to be verified be obtained
Quantitative analysis results.
For the Web service combination process under open environment, first, the dynamic change of environment will influence different tools
The qos value of body service, so as to influence the selection of Web service.And the dynamic change of environment is random, uses the horse of parametrization
Er Kefu decision processes model the dynamic change of environment, can be with the dynamic change of real simulation environment, so as to improve out
Put the accuracy of Web service combination verification method under environment.Secondly, in practical Web service combination scene, abstract service it
Between be not independent from each other, difference service between there may be complementary situations.To solve the problems, such as this, present invention introduces
Limitation requires to constrain Web service combination process, while can also alleviate the state explosion problem in verification process.Finally, make
The parametrization Markovian decision process of environmental condition and the parametrization markov of the band limitation requirement of Web service combination process
Decision process interacts and is used as mono- input of PARAM, and the temporal logic formula of QoS attributes to be verified is defeated as another
Enter.It is verified using PARAM, the rational function with variable can be obtained.After the occurrence of given variable, patrolled with regard to sequential can be obtained
Collect the verification result of formula, you can obtain state transduction pathway corresponding with result, that is, optimal Web service combination side
Formula.Therefore, Web service combination process is verified using PARAM, the Web service that QoS is perceived under open environment can be solved
Relevant issues in combination research.
More mesh of the nonfunctional space of user are concentrated on to the research of the QoS Web service combination verification methods perceived before
Mark property and probabilistic processing, not in view of influence of the randomness of environment dynamic change to Web service combination process.Together
When, also seldom in view of the caused web service composition method inaccuracy that interdepends between abstract service under actual conditions
Problem.
Invention content
Above-mentioned the deficiencies in the prior art are directed to, the purpose of the present invention is to provide a kind of Web based on parameterized model
Services composition verification method, to solve not consider the uncertain and randomness of environment to Web service combination in the prior art
Web service composition method inaccuracy caused by interdepending between abstract service under the influence of process and actual conditions
Problem.Environmental model is modeled as parametrization Markovian decision process by the present invention, and appropriate has handled environmental condition difference shape
The randomness mutually converted between state.It requires to constrain in addition, Web service combination process model building is become parametrization and introduces limitation
Web service combination process is converted into the parametrization Markovian decision process with limitation requirement, solves Web service combination
Dependence Problem between abstract service in the process, while alleviate state explosion problem present in probabilistic model checking technology, it carries
The accuracy and validity of high method.
In order to achieve the above objectives, the technical solution adopted by the present invention is as follows：
A kind of Web service combination verification method based on parameterized model of the present invention, includes the following steps：
(1) Web service combination process and QoS attributes are taken out according to the characteristics of object to be studied；
(2) the QoS attributive analysis in step (1) influences the environmental condition of the QoS attributes, and the environmental condition is built
Mould is the Markovian decision process (PMDP) of parametrization；
(3) the QoS model attributes in step (1) are become into temporal logic formula；
(4) the Web service combination process model building in step (1) is become into the parametrization Markovian decision with limitation requirement
Process (Re-PMDP), and be allowed to interact with the Markovian decision process model parameterized in step (2)；
(5) by the parameter of the band limitation requirement in the Markovian decision process of the parametrization in step (2) and step (4)
Change the finite state model that Markovian decision process forms description probability system；Temporal logic formula in step (3) represents
System property to be verified using the probabilistic model checking technology of parametrization, verifies finite state model in the environment of dynamic
Whether meet QoS attributes to be verified, and be verified result.
Preferably, the step (1) specifically includes：
(11) task that object to be studied need to be completed is analyzed, defines one group of abstract service description system action；
(12) abstract service in analytical procedure (11), same abstract service is provided by different specific services, by this
The set of specific service is defined as one group of specific service of each abstract service；
(13) analysis result in step (11) and (12), Web service combination process is abstracted by object to be studied；
(14) according to features of the object to be studied, QoS attributes to be verified qos value corresponding with its is taken out.
Preferably, the step (2) specifically includes：
(21) the characteristics of qos value in step (14), the component part of QoS attributes is determined；
(22) component part of the QoS attributes in step (21), the actual conditions of the analyzing influence QoS attributes will
It is determined as environmental condition；
(23) environmental condition in step (22) is modeled as Markovian decision process (MDP), wherein, environmental condition
There is different states, different conditions correspond to the different conditions of Markovian decision process；Between environmental condition different conditions with
The process of machine conversion corresponds to the migration of different conditions in Markovian decision process；
(24) probability parameter that will be migrated between the different conditions of Markovian decision process in step (23), definition should
Probability is a variable, and gives a value according to the characteristics of environmental condition, by the Markov decisior process of the environmental model
The Markovian decision process of journey conversion parameter.
Preferably, the Markovian decision process of the parametrization in the step (2) by probabilistic model checking tool into
Row creates.
Preferably, the step (3) specifically includes：
(31) incentive structure of Web service combination process is determined according to the component part of QoS attributes in step (21)
(Reward)；
(32) it is public to be become sequential logic by the incentive structure in step (31) for the QoS model attributes in step (14)
Formula.
Preferably, the step (4) specifically includes：
(41) the Web service combination process in the particular state of environmental condition in step (23) and step (13), point
Analysis Web service combination selection under different conditions is serviced the variation of generation, and a label is defined in Markovian decision process
State, for the state under different environmental condition states, the migration of state also will be different, with this different to represent environment item
Influence of the variation of part to Web service combination process；With reference to the incentive structure that step (31) determines, by what is influenced by environmental model
Web service combination process model building becomes Markovian decision process；
(42) it according to the characteristics of the Web service combination process influenced in step (41) by environmental model, determines during being somebody's turn to do
It needs parameterized specific probability and gives rational value range, the Markovian decision process in step (41) is converted
Markovian decision process for parametrization；
(43) the Web service combination process in step (13) is analyzed the dependence during this between abstract service and is closed
System defines corresponding limit according to dependence and requires (Request), uses ginseng in the limitation requirement constraint step (42) of definition
The Markovian decision process of numberization, and it is converted into the parametrization Markovian decision process with limitation requirement.
Preferably, the parametrization Markovian decision process of the band limitation requirement in the step (4) passes through probabilistic model
Detection instrument is created.
Beneficial effects of the present invention：
Environmental model is modeled as the Markovian decision process of parametrization by the present invention using probabilistic model checking technology
(PMDP), the Web service combination process model building influenced by environmental model is become the parametrization markov with limitation requirement to determine
Plan process (Re-PMDP), and be allowed to interact with the PMDP of environmental model.It solves random because what is converted between environmental condition state
The inaccurate problem of Traditional Web services combined method caused by property；Meanwhile introduce limit value requirement concept to constrain PMDP,
Solve the problems, such as that web service composition method is inaccurate caused by the Dependence Problem between abstract service under true environment, and
Alleviate the state explosion problem that probabilistic model checking technology is brought.
Description of the drawings
Fig. 1 is the frame diagram of the present invention.
Fig. 2 is Web service combination process schematic.
Fig. 3 is the process schematic that environmental condition changes at random.
Fig. 4 is p rational function image schematic diagram when being 0.5.
Fig. 5 a be p_bad and p as variable when rational function schematic diagram.
Fig. 5 b be p_good and p as variable when rational function schematic diagram.
Rational function schematic diagram when Fig. 6 a are p_good and p_bad is variable.
Rational function schematic diagram when Fig. 6 b are p_good and p_normal is variable.
Rational function schematic diagram when Fig. 6 c are p_normal and p_bad is variable.
Rational function schematic diagram when Fig. 7 a are p_good and p is variable.
Rational function schematic diagram when Fig. 7 b are p_bad and p is variable.
Rational function schematic diagram when Fig. 7 c are p_normal and p is variable.
Specific embodiment
For the ease of the understanding of those skilled in the art, the present invention is made further with reference to embodiment and attached drawing
Bright, the content that embodiment refers to not is limitation of the invention.
A kind of Web service combination verification method based on parameterized model of the present invention coordinates shown in Fig. 1, with " online purchase
For object service " scene, specifically include as follows：
Step (1)：Web service combination process and QoS attributes are taken out according to the characteristics of object to be studied.
Specifically include following steps：
(11) task that the need of object to be studied are completed is analyzed, defines one group of abstract service description system action.In this reality
In example, online service system needs to complete following task：Select suitable shopping platform, such as Jingdone district, Taobao；Selection is closed
Suitable shop, as different shops, its price of same commodity is possible and different, and the discount possessed also can be different；Choosing
Payment platform is selected, now most are exactly Alipay platform and wechat payment platform；Select express delivery, the arrival of different express deliveries
Time is different with price.So under this example, one shares 4 abstract services, is defined as AS1, AS2, AS3, AS4。
(12) abstract service in analytical procedure (11), same abstract service is provided by different specific services, by this
The set of a little specific services is defined as one group of specific service of each abstract service.From step (11) it is found that one in this example
Four kinds of abstract services are defined altogether, are respectively：Sales platform, shop, payment platform, express delivery are defined as AS1, AS2, AS3, AS4。
Abstract service AS1It is middle there are two kinds of specific services, be CS1_1, CS1_2 respectively, be defined as AS1={ CS1_1, CS1_2 }.This two
Kind specific service expression provides the service of sales platform, and CS1_1 represents Taobao, and CS1_2 represents Jingdone district.In abstract service AS2
In again there are 3 kinds of specific services, i.e. AS2={ CS2_1, CS2_2, CS2_3 }.CS2_1 represents Bank of Communications, during CS2_2 is represented
Bank of state, CS2_3 represent Construction Bank.In abstract service AS3In have 2 specific services, AS3={ CS3_1, CS3_2 }.CS3_1
Alipay platform is represented, CS3_2 represents wechat payment platform.In abstract service AS4In have 4 specific services, AS4=CS4_1,
CS4_3, CS4_3, CS4_4 }.CS4_1 is represented along rich express delivery, and CS4_2 represents rhythm up to express delivery, and CS4_3 represents the logical express delivery of circle, CS4_
Lead to express delivery in 4 representatives.
(13) analysis result in step (11) and (12), Web service combination process is abstracted by object to be studied.
In this example, the abstract service and specific service of online shopping service system, Web service combination process such as Fig. 2 have been obtained.
(14) according to features of the object to be studied, QoS attributes to be verified qos value corresponding with its is taken out.In this example
In, using " total time for completing online shopping process is minimum ", corresponding qos value " is completed as QoS attributes to be verified
The total time of line shopping process ".
Step (2)：QoS attributive analysis in step (2) influences the environmental condition of the QoS attributes, and by the environment
Condition is modeled as the Markovian decision process (PMDP) of parametrization.
Specifically include following steps：
(21) the characteristics of qos value in step (14), the component part of QoS attributes is determined.In instances, according to
The characteristics of line shopping service, qos value " total time for completing online shopping process is minimum " are made of two parts：" call service
Time " and " time of express delivery delivery ".
(22) component part of the QoS attributes in step (21), the actual conditions of the analyzing influence QoS attributes will
It is determined as environmental condition.In instances, qos value is by " time for calling service " and " time of express delivery delivery " two parts group
Into.In the actual environment, " purchase volume of unit interval " can influence two above part simultaneously.When " purchase volume of unit interval "
During increase, the service request quantity in network environment can increase, and network environment is likely to enter congestion state, so as to influence " to adjust
With the time of service ".Meanwhile the increase of purchase volume can make increasing for express delivery quantity in logistics service, so as to influence AS4In it is specific
" time of express delivery delivery " of service.Therefore, " purchase volume of unit interval " is defined as environmental condition by us.
(23) environmental condition in step (22) is modeled as Markovian decision process (MDP).Wherein, environmental condition
There is different states, different conditions correspond to the different conditions of Markovian decision process.Between environmental condition different conditions with
The process of machine conversion corresponds to the migration of different conditions in Markovian decision process.In instances, it is the environmental condition " unit interval
Purchase volume " be defined as three states, i.e., purchase volume is very big, purchase volume is normal, purchase volume is smaller.In environmental model, respectively
Correspond to Good, tri- states of Normal, Bad.Each state convert when there are three types of may, i.e., with certain probability P 1 according to
Old oneself original state of holding, other two state is transformed into certain probability P 2 and probability P 3.P1, P2, P3 are referred to as
For environmental condition transition probability.Wherein, P1, P2, P3 should meet the constraint of the following formula：
P1+P2+P3=1 (1)
Its state transition graph is shown in Fig. 3.The process model building is become into MDP using PARAM language.
(24) probability parameter that will be migrated between the different conditions of Markovian decision process in step (23), definition should
Probability is a variable, and gives a value according to the characteristics of environmental condition.By the Markov decisior process of the environmental model
The Markovian decision process of journey conversion parameter.In instances, it by the transition probability parameters in environmental model, will parameterize
Variable-definition is p_env.The relationship of wherein P1, P2, P3 and p_env are as shown in table 1 below：
Table 1
| Probability title | The relationship of p_env and ambient condition transition probability |
| P1 | P1=p_env |
| P2 | P2=p_env |
| P3 | P3=1-2*p_env |
At this point, given p_env ∈ [0.1,0.4].By above procedure, the MDP of environmental model in step (23) is converted
For PMDP.
Step (3)：QoS model attributes in step (1) are become into temporal logic formula；
Specifically include following steps：
(31) incentive structure of Web service combination process is determined according to the component part of qos value in step (21)
(Reward).In instances, determine that the incentive structure of the Web service combination process under " online shopping service " scene is entitled
" time " includes " time for calling service " and " express delivery delivery period " two parts.However, during Web service combination not
It is that each specific service can have two above part simultaneously.For example, the specific service Taobao in abstract service 1 is come
It says, QoS attributes only have " time for calling service ", without " express delivery delivery period ".To handle such case, defining
When the incentive structure of " time ", when abstract service is 4, " express delivery delivery period " that part need to be set as 0.
(32) it is public to be become sequential logic by the incentive structure in step (31) for the QoS model attributes in step (14)
Formula.In instances, QoS attributes " total time for completing online shopping process is minimum " to be verified are modeled as temporal logic formula
It is as follows：
R{“time”}min=[F " completed "] (2).
Step (4)：Web service combination process model building in step (1) is known as the parametrization Ma Erke with limitation requirement
Husband's decision process (Re-PMDP), and be allowed to interact with PMDP models in step (2).
Specifically include following steps：
(41) the Web service combination process in the particular state of environmental condition in step (22) and step (13), point
Analysis Web service combination selection under different conditions is serviced the variation of generation.A label is defined in Markovian decision process
State, for the state under different environmental condition states, the migration of state also will be different, with this different to represent environment item
Influence of the variation of part to Web service combination process.With reference to the incentive structure that step (32) determines, by what is influenced by environmental model
Web service combination process model building becomes MDP.In instances, the variation of environmental condition " purchase volume of unit interval " state will be from
Two aspects influence Web service combination.For " time for calling service " this part, environmental model passes through shadow
" calling the service probability of success " is rung to influence the value of QoS.A variable " select " defined in Markovian decision process is used
The state transition of variable " select " come represent call service this action.And the conversion between " select " state has one
This probability is known as " call and service successful probability " by fixed probability.The different states of environmental condition by the probability come
Constrain the conversion between " select " variable states, such as Fig. 3.Specifically " probability of success for calling service " and " purchase of unit interval
Relationship between the amount of buying " state, such as the following table 2：
Table 2
| " purchase volume of unit interval " state | Environmental model state | Call the probability of success of service |
| It is less | Good | 0.9 |
| Normally | Normal | 0.5 |
| It is more | Bad | 0.4 |
For environmental condition in the state of difference, the probability of success for calling service is different.When service is called, if do not had
Successfully, by repeating the step for.So the probability of success is relatively low compared with the probability of success is higher, the possibility for having bigger repeatedly weighs
Multiple this calls the action serviced." calling service time " in incentive structure " time ", this part is exactly to calculate the action to hold
Capable number.In addition, for " express delivery delivery period " this part, " the purchase volume variation of unit interval " will influence " single
The express delivery delivery amount of position time ", so as to influence express delivery delivery period, such as the following table 3：
Table 3
| The purchase volume of unit interval | The express delivery delivery amount of unit interval | Express delivery delivery period |
| It is less | It is less | Comparatively fast |
| Normally | Normally | Normally |
| It is more | It is more | It is relatively slow |
Environmental condition " purchase volume of unit interval " influences qos value by the definition of incentive structure." online shopping takes
Only AS in business "4In specific service need to calculate, this part will be set as zero in other three abstract services.When abstract
It services as AS4When, " express delivery delivery period " such as the following table 4 of the different specific services under different ambient conditions：
Table 4
| Environmental model state | Good | Normal | Bad |
| Along rich express delivery | 2 (my god) | 2 (my god) | 3 (my god) |
| Rhythm reaches express delivery | 3 (my god) | 4 (my god) | 7 (my god) |
| In lead to express delivery | 3 (my god) | 3 (my god) | 5 (my god) |
| Jingdone district logistics | 2 (my god) | 2 (my god) | 2 (my god) |
" express delivery delivery period " this part is calculated by table 4 in incentive structure " time ".
(42) it according to the characteristics of the Web service combination process influenced in step (41) by environmental model, determines during being somebody's turn to do
It needs parameterized specific probability and gives rational value range, the MDP in step (41) is converted into PMDP.In example
In, when " purchase volume of unit interval " during Web service combination is smaller, the state of environmental model is Good, is also just meaned
It, calls the service probability of success larger.But " call and service successful probability " is more than because of this reason, it also can be by it
He influences factor.If for example, user is in elevator, poor signal, then call service successful probability can relative reduction.Institute
With in the state of same environmental model, call service the probability of success should not a definite value, and a model should be in
In enclosing.It is for appropriate processing such case, " probability of success of service " under different condition is parameterized as three changes
Measure p_good, p_normal, p_bad.Then, a range is given for each parameter, MDP is converted into PMDP.Given range
Such as the following table 5：
Table 5
| Call the probability of success of service | Range |
| p_good | [0.7-0.9] |
| p_normal | [0.4-0.6] |
| p_bad | [0.1-0.3] |
(43) the Web service combination process in step (13) is analyzed the dependence during this between abstract service and is closed
System defines corresponding limitation requirement (Request) according to dependence.Using in the limitation requirement constraint step (42) of definition
PMDP, and it is converted into Re-PMDP.In instances, abstract service AS1With AS4There is dependence.AS1It is selected in if
Specific service " Jingdone district ", then in AS4In can only just select " Jingdone district logistics ".And if in AS1It is middle to have selected " Taobao ", then
AS4In, other than " Jingdone district logistics " cannot select, other specific services can unrestricted choice.For appropriate processing
This dependence introduces limitation requirement and Web service combination model is constrained.So AS4Middle specific service CS4_4
It will be required to definite limitation, only in AS1In when having selected CS1_2, CS4_4 could be selected.
Step (5)：PMDP in step (2) and the Re-PMDP in step (3) is constituted into the finite of description probability system
State model；Temporal logic formula in step (4) represents system property to be verified, using probabilistic model checking technology, tests
It demonstrate,proves whether finite state model meets attribute to be verified in the environment of dynamic, and obtains quantitative verification result.In instances,
The rational function with variable that quantitative verification result is exactly.
After the completion of above-mentioned all steps, pass through the correctness and feasibility of two experimental verification present invention.The reality
The running environment tested be Ubuntu17.04LTS systems, Intel's Duo I3 processors, 4GB memories；Experimental tool is probability mould
Type detector PARAM, version number PARAM2.0.In experiment one, the basic status of environmental model is two, is Good respectively
And Bad.Parameterized variable there are three being had altogether in experiment, shown in value range table 6.On the basis of above, Web service group
The global QoS that the purpose of conjunction method finds Web service combination is optimal, in this experiment, that is, the reward structures defined
The value of " time " is minimum.So the temporal logic formula shown in above-mentioned formula (2) is verified.Obtained result such as 7 institute of table
Show.
Table 6
| Variable name | Meaning | Value |
| p_good | Probability when environmental condition is preferable | [0.7..0.9] |
| p_bad | Probability when environmental condition is poor | [0.1..0.3] |
| p | Ambient condition random transition probability | [0.1..0.9] |
Table 7
The rational function that can be seen that from more than experimental result is a three element complex.Three element complex not will pass through
Graphical representation comes out, in order to intuitively represent the experimental result of the present invention, by giving the value of a parameter (in its value
In the range of), and pass through probabilistic model checking technology and obtain the rational function of its other two parameter, design parameter numerical value is set such as
Table 8：
Table 8
From more than parameter setting it is found that the variation probability of setting ambient condition is 0.5.P_good be x-axis, p_bad y
Axis, z-axis be the present invention verify temporal logic formula as a result, this result is that a face, specific functional image such as Fig. 4.
As can be seen from the above results, when p is definite value, when p_good and p_bad is incremented by, the sequential logic of verification is public
The global QoS of the value of formula, i.e. Web service combination can also reduce therewith.The correctness of model of the present invention can be proved from this.p_
Good is represented in the case where environmental condition state is Good:Call and service successful probability, p_bad represent environmental condition as
In the case of Bad, call and service successful probability.No matter under what circumstances, call the successful probability of service higher, it is global
The value of QoS is just smaller.
On the basis of above, the value of fixed p_good and p_bad, design parameter sets as shown in table 9 with experimental result.Its
In, the rational function schematic diagram of the experimental result in table 9 is as shown in figure 5 a and 5b.Fig. 5 a represent p_good and fix, p_bad with
Rational function schematic diagram when p is as variable, Fig. 5 b represent p_bad and fix, and rational function when p_good and p is as variable shows
It is intended to.
Table 9
| p_good | p_bad | p | As a result |
| 9/10 | [1/10,3/10] | [1/10,9/10] | (-30)/(10*y*x-10*y-9*x) |
| [7/10,9/10] | 1/10 | [1/10,9/10] | (30)/(10*y*x-1*x+1) |
Analysis more than functional image can obtain two rules：First, with the increase of p, the overall situation of Web service combination
Qos value is smaller.P represents in environmental model under a certain state random transition to Good shape probability of states, when the probability is smaller,
It is smaller to illustrate that ambient condition is transferred to Good shape probability of states, is transferred to that Bad shape probability of states are bigger, and " time " value should at this time
It is larger.It is consistent so obtaining result with our expection.In addition, the value with p_good and p_bad increases, the value of global QoS
Into the trend successively decreased, equally it is consistent with expection.
More than, demonstrate the correctness of the web service composition method present invention introduces environmental model.In order to verify this method
Middle environmental model it is expansible, by two conditional extensions of environmental model be three states, other conditions are constant to be tested.This
When, in an experiment altogether there are four parameterized variable, value range is as shown in table 10 below.Then, to above-mentioned formula (1)
It is verified, it is as shown in table 11 to obtain result,
Table 10
| Variable name | Meaning | Value |
| p_good | Probability when environmental condition is preferable | [0.7..0.9] |
| p_normal | Probability when environmental condition is general | [0.4..0.6] |
| p_bad | Probability when environmental condition is poor | [0.1..0.3] |
| p | Ambient condition random transition probability | [0.1..0.9] |
Table 11
From the results, it was seen that it obtains the result is that a quaternary rational function.More intuitive experimental result in order to obtain,
It determines the variable of two parameters, and re-starts experiment.First, ambient condition transition probability p=0.3 is set, and design parameter is set
It puts and experimental result such as the following table 12, and draws the functional image of its result as shown in Fig. 6 a, Fig. 6 b and Fig. 6 c.Wherein, Fig. 6 a generations
Table p and p_normal is fixed, rational function schematic diagram when p_good and p_bad is variable.Fig. 6 b represent p and p_bad and fix,
Rational function schematic diagram when p_good and p_normal is variable.Fig. 6 c represent p and p_good and fix, p_normal and p_
Rational function schematic diagram when bad is variable.
Table 12
| p_good | P_bad | p_normal | As a result |
| [7/10,9/10] | [1/10,3/10] | 0.5 | (30)/(3*p_good+3*p_bad+2) |
| [7/10,9/10] | 0.2 | [4/10,6/10] | (150)/(15*p_good+20*p_normal+3) |
| 0.8 | [1/10,3/10] | [4/10,6/10] | (150)/(15*p_bad+20*p_normal+12) |
Then, for verification environment state probability p and the relationship of other parameter.One in other parameter is selected to become
Amount, verification model obtain binary rational function.Design parameter is set and experimental result such as table 13.Rational function image in table 13
As shown in Fig. 7 a, Fig. 7 b and Fig. 7 c.Wherein, Fig. 7 a represent p_bad and p_normal and fix, having when p_good and p is variable
Manage function schematic diagram.Fig. 7 b represent p_good and p_normal and fix, rational function schematic diagram when p_bad and p is variable.Figure
7c represents p_good and p_bad and fixes, rational function schematic diagram when p_normal and p is variable.
Table 13
| p_good | P_bad | p_normal | p | As a result |
| [7/10,9/10] | 0.2 | 0.5 | [1/10,4/10] | (30)/(10*p*p_good-8*p+5) |
| 0.8 | [1/10,3/10] | 0.5 | [1/10,4/10] | (30)/(10*p*p_bad-8*p+5) |
| 0.6 | 0.2 | [4/10,6/10] | [1/10,4/10] | (-3)/(2*p*p_normal-p-p_normal) |
To verify the scalability of the method for the present invention, the number of set environment condition is two other parameters settings and preceding phase
Together, using abstract service number and the number of specific service experiment 2 is carried out as variable.Concrete outcome is as follows shown in each table.Its
Middle table 14 represents and can verify that abstract service when being 4, supported specific service quantity and corresponding status number.
Table 14
| Abstract service quantity | Specific service quantity | Status number | Model state number with limitation requirement |
| 4 | 20 | 1347369 | 131369 |
| 4 | 21 | 1633641 | 151881 |
| 4 | 22 | 1963289 | 174425 |
| 4 | 23 | 2340489 | 199097 |
| 4 | 24 | 2769609 | 2239689 |
| 4 | 25 | It can not calculate | 2163313 |
Table 15
| Abstract service quantity | Specific service quantity | Status number | Model state number with limitation requirement |
| 5 | 9 | 531441 | 298161 |
| 5 | 10 | 888889 | 600889 |
| 5 | 11 | 1417249 | 884849 |
| 5 | 12 | 2171625 | 1411305 |
| 5 | 13 | It can not calculate | 2163313 |
Table 16
| Abstract service quantity | Specific service quantity | Status number | Model state number with limitation requirement |
| 6 | 4 | 43689 | 19113 |
| 6 | 5 | 156249 | 56249 |
| 6 | 6 | 447897 | 188697 |
| 6 | 7 | 1098057 | 521817 |
| 6 | 8 | 2396745 | 987721 |
| 6 | 9 | It can not calculate | 2683449 |
Table 17
| Abstract service quantity | Specific service quantity | Status number | Model state number with limitation requirement |
| 7 | 3 | 26241 | 9393 |
| 7 | 4 | 174761 | 76457 |
| 7 | 5 | 781249 | 281249 |
| 7 | 6 | 2687385 | 613785 |
| 7 | 7 | It can not calculate | 1981625 |
Table 18
| Abstract service quantity | Specific service quantity | Status number | Model state number with limitation requirement |
| 8 | 3 | 78729 | 43737 |
| 8 | 4 | 699049 | 305833 |
| 8 | 5 | It can not calculate | 1406249 |
Table 19
| Abstract service quantity | Specific service quantity | Status number | Model state number with limitation requirement |
| 9 | 3 | 236193 | 131217 |
| 9 | 4 | 2796201 | 1223337 |
| 9 | 5 | It can not calculate | 2531249 |
More than experimental result is provable, introduces limitation requirement to constrain Web service combination model, can largely reduce shape
State number alleviates the state explosion problem always existed in probabilistic model checking technology, and supports greater number of service.
There are many concrete application approach of the present invention, and the above is only the preferred embodiment of the present invention, it is noted that for
For those skilled in the art, without departing from the principle of the present invention, several improvement can also be made, this
A little improve also should be regarded as protection scope of the present invention.
Claims (7)
1. a kind of Web service combination verification method based on parameterized model, which is characterized in that include the following steps：
（1）Web service combination process and QoS attributes are taken out according to the characteristics of object to be studied；
（2）According to step（1）In QoS attributive analysis influence the environmental condition of the QoS attributes, and the environmental condition is modeled as
The Markovian decision process of parametrization；
（3）By step（1）In QoS model attributes become temporal logic formula；
（4）By step（1）In Web service combination process model building become band limitation require parametrization Markov decisior process
Journey, and be allowed to and step（2）The Markovian decision process model of middle parametrization interacts；
（5）By step（2）In parametrization Markovian decision process and step（4）In band limitation requirement parametrization horse
Er Kefu decision processes form the finite state model of description probability system；Step（3）In temporal logic formula represent it is to be tested
The system property of card, using the probabilistic model checking technology of parametrization, verify finite state model in the environment of dynamic whether
Meet QoS attributes to be verified, and be verified result.
2. the Web service combination verification method according to claim 1 based on parameterized model, which is characterized in that described
Step（1）It specifically includes：
（11）The task that object to be studied need to be completed is analyzed, defines one group of abstract service description system action；
（12）Analytical procedure（11）In abstract service, same abstract service is provided by different specific service, this is specific
The set of service is defined as one group of specific service of each abstract service；
（13）According to step（11）With（12）In analysis result, object to be studied is abstracted into Web service combination process；
（14）According to features of the object to be studied, QoS attributes to be verified qos value corresponding with its is taken out.
3. the Web service combination verification method according to claim 2 based on parameterized model, which is characterized in that described
Step（2）It specifically includes：
（21）According to step（14）In qos value the characteristics of, determine the component parts of QoS attributes；
（22）According to step（21）In QoS attributes component part, the actual conditions of the analyzing influence QoS attributes, by its really
It is set to environmental condition；
（23）By step（22）In environmental condition be modeled as Markovian decision process, wherein, environmental condition has different
State, different conditions correspond to the different conditions of Markovian decision process；Random transition between environmental condition different conditions
Process corresponds to the migration of different conditions in Markovian decision process；
（24）By step（23）The probability parameter migrated between the different conditions of middle Markovian decision process defines the probability
For a variable, and a value is given according to the characteristics of environmental condition, the Markovian decision process of the environmental model is turned
Change the Markovian decision process of parametrization.
4. the Web service combination verification method based on parameterized model according to any one in claim 1-3, special
Sign is, the step（2）In the Markovian decision process of parametrization created by probabilistic model checking tool.
5. the Web service combination verification method according to claim 3 based on parameterized model, which is characterized in that described
Step（3）It specifically includes：
（31）According to step（21）The component part of middle QoS attributes determines the incentive structure of Web service combination process；
（32）According to step（31）In incentive structure, by step（14）In QoS model attributes become temporal logic formula.
6. the Web service combination verification method according to claim 5 based on parameterized model, which is characterized in that described
Step（4）It specifically includes：
（41）According to step（23）The particular state and step of middle environmental condition（13）In Web service combination process, analyze
Web service combination selection is serviced the variation of generation under different conditions, and a flag state is defined in Markovian decision process,
For the state under different environmental condition states, the migration of state also will be different, with this different to represent environmental condition
Change the influence to Web service combination process；With reference to step（31）Determining incentive structure, the Web that will be influenced by environmental model
Services Composition process model building becomes Markovian decision process；
（42）According to step（41）In influenced by environmental model Web service combination process the characteristics of, determine should during need
Parameterized specific probability and given rational value range, by step（41）In Markovian decision process be converted to ginseng
The Markovian decision process of numberization；
（43）According to step（13）In Web service combination process, analyze the dependence between abstract service, root during this
Corresponding limitation requirement is defined according to dependence, uses the limitation requirement constraint step of definition（42）The markov of middle parametrization
Decision process, and it is converted into the parametrization Markovian decision process with limitation requirement.
7. the Web service combination verification method based on parameterized model according to claim 1 or 6, which is characterized in that institute
State step（4）In band limitation requirement parametrization Markovian decision process created by probabilistic model checking tool.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711420311.3A CN108170404B (en) | 2017-12-25 | 2017-12-25 | Web service combination verification method based on parameterized model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711420311.3A CN108170404B (en) | 2017-12-25 | 2017-12-25 | Web service combination verification method based on parameterized model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108170404A true CN108170404A (en) | 2018-06-15 |
| CN108170404B CN108170404B (en) | 2021-04-06 |
Family
ID=62520155
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711420311.3A Active CN108170404B (en) | 2017-12-25 | 2017-12-25 | Web service combination verification method based on parameterized model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108170404B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110780862A (en) * | 2019-10-12 | 2020-02-11 | 南京邮电大学 | Automatic service combination method based on key path spanning tree |
| CN110971683A (en) * | 2019-11-28 | 2020-04-07 | 海南大学 | Service combination method based on reinforcement learning |
| CN111857092A (en) * | 2020-06-22 | 2020-10-30 | 杭州群核信息技术有限公司 | Real-time error detection system and method for household parametric model |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080172714A1 (en) * | 2007-01-16 | 2008-07-17 | International Business Machines Corporation | Method for model based verification of security policies for web service composition |
| CN102868757A (en) * | 2012-09-28 | 2013-01-09 | 南京大学 | Dynamic Web service combination method based on quality of service (QoC) indexes |
| US20160299936A1 (en) * | 2015-04-08 | 2016-10-13 | Oracle International Corporation | Automated test for uniform web service interfaces |
| CN106600058A (en) * | 2016-12-13 | 2017-04-26 | 北京理工大学 | Prediction method for combinations of cloud manufacturing service quality of service (QoS) |
-
2017
- 2017-12-25 CN CN201711420311.3A patent/CN108170404B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080172714A1 (en) * | 2007-01-16 | 2008-07-17 | International Business Machines Corporation | Method for model based verification of security policies for web service composition |
| CN102868757A (en) * | 2012-09-28 | 2013-01-09 | 南京大学 | Dynamic Web service combination method based on quality of service (QoC) indexes |
| US20160299936A1 (en) * | 2015-04-08 | 2016-10-13 | Oracle International Corporation | Automated test for uniform web service interfaces |
| CN106600058A (en) * | 2016-12-13 | 2017-04-26 | 北京理工大学 | Prediction method for combinations of cloud manufacturing service quality of service (QoS) |
Non-Patent Citations (1)
| Title |
|---|
| 范小芹等: "随机QoS感知的可靠Web服务组合", 《软件学报》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110780862A (en) * | 2019-10-12 | 2020-02-11 | 南京邮电大学 | Automatic service combination method based on key path spanning tree |
| CN110780862B (en) * | 2019-10-12 | 2021-05-11 | 南京邮电大学 | Automatic service combination method based on key path spanning tree |
| CN110971683A (en) * | 2019-11-28 | 2020-04-07 | 海南大学 | Service combination method based on reinforcement learning |
| CN111857092A (en) * | 2020-06-22 | 2020-10-30 | 杭州群核信息技术有限公司 | Real-time error detection system and method for household parametric model |
| CN111857092B (en) * | 2020-06-22 | 2024-04-30 | 杭州群核信息技术有限公司 | Real-time error detection system and method for household parameterized model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108170404B (en) | 2021-04-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Dantan et al. | Worst-case and statistical tolerance analysis based on quantified constraint satisfaction problems and Monte Carlo simulation | |
| US20180308132A1 (en) | Configurable relevance service test platform | |
| CN108170404A (en) | A kind of Web service combination verification method based on parameterized model | |
| Shao et al. | Decision guidance methodology for sustainable manufacturing using process analytics formalism | |
| CN111552728B (en) | Data processing method, system, terminal and storage medium of block chain | |
| CN105303447A (en) | Method and device for carrying out credit rating through network information | |
| Brunnert et al. | Using architecture-level performance models as resource profiles for enterprise applications | |
| US11538095B2 (en) | Virtual marketplace for distributed tools in an enterprise environment | |
| JP7000339B2 (en) | Virtual market for decentralized tools in a corporate environment | |
| US8671391B2 (en) | Component model for analytic applications supporting parameterization | |
| CN105426177B (en) | A kind of method and system for establishing SaaS Application models | |
| US20230086659A1 (en) | Facilitating customization and proliferation of state models | |
| CN108830480B (en) | ERP purchasing calculation method and system | |
| der Landwehr et al. | Computer simulation as evaluation tool of information systems: Identifying quality factors of simulation modeling | |
| Saputra et al. | Utilization of Service Oriented Architecture to Customize Accounting Business Process in Open ERP System for Smart SME’s | |
| Maddodi et al. | Generating workload for erp applications through end-user organization categorization using high level business operation data | |
| Offutt et al. | An industrial study of applying input space partitioning to test financial calculation engines | |
| CN107645412B (en) | Web service combination multi-target verification method in open environment | |
| US20120253880A1 (en) | Simulating dependency of token arrival on process performance | |
| Matveev et al. | Simulation modelling for assessing the adequacy of decision support models with choosing a product offer | |
| WO2021046105A1 (en) | Systems and methods for assigning attribution weights to nodes | |
| Youn et al. | Deriving queuing network model for UML for software performance prediction | |
| Duttagupta et al. | Performance mimicking benchmarks for multi-tier applications | |
| Dhas et al. | Relative analysis of sizing methods in the sense of e-commerce system | |
| Agarwal et al. | Quality model for component-based E-commerce application |
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 |