CN106570080A - Multilevel semantic matching method for cloud manufacturing resource services - Google Patents

Abstract

The invention discloses a multilevel semantic matching method for cloud manufacturing resource services. According to the cloud service requirement description of a cloud user, a multilevel semantic matching model for the cloud manufacturing resource services is established based on a cloud manufacturing service resource ontology library, the matching degree of each level is computed and is compared with a set threshold, the service filtering is carried out, and the matching of the next level is carried out after the conditions are met, so that the calculated amount of the next level can be reduced after matching and filtering for many times, and an optimal service matching result set can be also obtained. The method solves the problems that the existing method lacks sematic processing and the service matching cannot be measured; and meanwhile, fast optimization from massive service resources can be realized for the cloud user to select and use the cloud manufacturing service resources conveniently, quickly and cheaply.

Description

A kind of cloud manufacturing recourses service multi-level semantic matching method

Technical field

The present invention relates to a kind of Advanced Manufacturing System, more particularly to a kind of multi-level semantic matches side of cloud manufacturing recourses service Method.

Background technology

In recent years, as various countries increasingly deepen to manufacturing attention degree, cloud computing, net instrument, Internet of Things etc. Emerging technology has obtained quick development, promotes manufacturing industry constantly to be changed from type of production to service type direction, there is provided Yi Zhongxian The manufacturing service type developmental pattern entered.Cloud manufacturing recourses Service Matching technology be realize service type developmental pattern key technology it One, it is referred under cloud manufacturing environment, according to the resource service demand of user individual, rely on cloud manufacturing recourses service platform with The cloud manufacturing service of variation, modularity and functionalization carries out intelligent Matching, with the optimised service demand that this meets user, reaches To efficient centralized management and the utilization of manufacturing recourses.It is after cloud service demand is proposed, how quick effectively intactly numerous from magnanimity In many Service Sources, searching of optimal matching out, is the basis for realizing cloud resource service.But it is traditional based on Keywords matching skill Art can not meet user's requirement, lack the semantical definition to users ' individualized requirement and resource service, the language to target requirement Justice description is inaccurate, lacks semantic processes, and resource service matching is unable to quantitative Analysis etc., so as to have impact on resource service matching Efficiency and accuracy.To meet the demand of the more preferable level of cloud user, realize demand resource effective fast search and accurate Match somebody with somebody, thus the correlation technique of research manufacturing recourses Service Matching has certain theoretical and realistic meaning.

The content of the invention

For the problem that prior art is present, the present invention provides a kind of cloud manufacturing recourses and services multi-level semantic matches side Method, both there is provided cloud manufacturing service resource and the semantic description and semantic processes of demand for services, it is further contemplated that each level of Service Source On semantic matches, carry out the matching measurement of semantic similarity, quick effectively the searching of cloud service demand resource be better achieved Rope is matched, the demand for services higher level so as to meet user.

What the above-mentioned technical problem of the present invention was mainly addressed by following technical proposals：

A kind of cloud manufacturing recourses of the present invention service multi-level semantic matching method, are retouched according to the cloud service demand of cloud user State, set up cloud manufacturing recourses ontology library；The service of cloud manufacturing recourses is set up on the basis of the cloud manufacturing recourses ontology library multi-level Semantic matches model, is calculated every layer of matching degree and is compared with setting threshold values, carry out service filtration, more than or equal to ability after setting threshold values Can match into next layer, so repeatedly matching not only can reduce amount of calculation after filtering, and can also obtain optimal service Matching result collection.

Further, the cloud manufacturing recourses ontology library of setting up is：Using Ontology Language, to manufacturing recourses from basic letter Breath, three aspects of function information and constraint information carry out Ontology description, set up resource ontology storehouse, so as to the clothes for cloud user Matching between business demand and cloud manufacturing service resource provides semantic information and supports.

Further, the cloud manufacturing recourses of setting up service multi-level semantic matches model, calculate every layer of matching degree and with Setting threshold values compares, and carries out service filtration, obtains optimal service matching result collection and be specially：

Essential information matching layer model is set up, function information matching layer model is set up, constraint information matching layer model is set up, Constraint information matching layer model is set up, and carries out matching operation successively to each layer, then the threshold value by the result for calculating with setting is entered Row compares, and only just matches into next layer more than or equal to after given threshold, specific as follows：

1.1) essential information matching

Define 1：Cloud manufacturing service resource RP, cloud service demand RR of cloud user.

Cloud manufacturing service resource essential information is described as：RP_BaseInf=<rp₁,rp₂,rp₃,...,rp_n>；The cloud of cloud user Demand for services essential information is described as：RR_BaseInf=<rr₁,rr₂,rr₃,...,rr_m>.The title of demand for services is derived from body Notional word, based on this, measured by similarity between Ontological concept, set up essential information Matching Model：

In formula：Match(RP,RR)_BaseInfRepresent the essential information of the demand for services of cloud manufacturing service resource and cloud user Matching degree；N, m represent resource service essential information notional word number (n>=m)；Sim(rp_i,rr_i) for notional word rp_iAnd rr_i's Semantic similarity；λ_i(0≤λ_i≤ 1) represent Sim (rp_i,rr_i) weight coefficient；Dis(rp_i,rr_i) for notional word rp_iAnd rr_i's Semantic distance；Weight_jFor Anknupfungskeregriff word rp_iAnd rr_iWeighted value on shortest path；R is Anknupfungskeregriff word rp_iAnd rr_iMost The side number of short path；I(parent(c_j)) for target father's conceptual information amount；I(c_j) for sub- conceptual information amount；N ' is general for target father The quantity of the direct sub- concept read；p(c_j) represent Ontological concept word c_jProbability；Concept root is body root node, and its probability is 1。

1.2) function information matching

It is main to be matched in terms of input, output, premise and effect four.

Input/output (I/O) is matched

Input/output parameters are the formalized descriptions to cloud manufacturing service resource, and corresponding Semantic Similarity Measurement is all Carry out for Ontological concept, thus the Computing Principle of essential information notional word matching can be adopted.

Define 2：The |input paramete of cloud manufacturing recourses service describing integrates as InPS；The cloud service requirement description of cloud user it is defeated Enter and integrate as InPD.Element InPS of the matching i.e. from InPS₁Start to InPS_jRespectively with InPD in element InPD_jMatched Degree is calculated, and finds the maximum InPS of semantic matching degree more afterwards by calculating_i, then matching output.Based on this, the input of foundation Information matches model is：

In formula：InPS_InjRepresent that cloud manufacturing service resource issues j-th parameter of input；InPD_InjRepresent cloud service demand J-th parameter of input；n_IRepresent all parametric component numbers of input.

In the same manner, output matching model can be obtained：

In formula：Sim(RP,RR)_outputRepresent the output semantic similitude of the demand for services of cloud manufacturing service resource and cloud user Degree, represents OutPS_OutjRepresent that Service Source issues j-th parameter of output；OutPD_OutjRepresent the jth of cloud service demand output Individual parameter；n_ORepresent all parametric component numbers of output.

Premise/effect (P/E) is matched

The property parameters of premise and effect (P/E) are not belonging to the concept in body domain, simply a kind of constraint expression formula, by individual Pronouns, general term for nouns, numerals and measure words object, grammatical term for the character predicate and parameter value parameter are constituted, and semantic description is<object,predicate, parameter>.Object and predicate are derived to class in resource ontology, attribute and example etc., base can be adopted In the computation model of Ontological concept information matches；It is for parameter is using the matching algorithm of numerical intervals, specific as follows：

In formula：Abbreviations of the pa for parameter；pa_RRFor the attribute value Operations of Interva Constraint of cloud users service needs；pa_RPFor The attribute value Operations of Interva Constraint of cloud manufacturing service resource；" | | " is parameter interval length of field, such as | 60,70 |=10.

Based on this, P/E (premise/effect) Matching Model is set up：

Therefore function information Matching Model is：

In formula：δ₁, δ₂, δ₃, δ₄(δ₁,δ₂,δ₃,δ₄∈ [0,1]) input, output, premise, four aspects of effect are represented respectively The attribute weight coefficient of function.

1.3) constraint information matching

The main time for including resource service, three constrained parameters of cost and credit worthiness, help cloud service platform each with this Side goes the service quality for weighing cloud resource.

Time, cost constraint matching

Time, cost this kind of index for resource service, mainly numerical value parameter, refer to values match algorithm, tool Body is as follows：

Meet condition, that is, time and cost that the time for servicing and cost are required less than or equal to party in request, that is, think It is 1 with degree, is otherwise 0.

Credit worthiness constrained matching

When credit worthiness constraint is evaluated, it is related to different resource and corresponding service-evaluating index, and in particular to soft Part resource, knowledge resource and human resourcess.Based on this, the computation model for setting up credit worthiness evaluation of estimate is：

In formula：I=1,2,3, software, knowledge and human resourcess are represented respectively；N represents Service Source S_iEvaluation index number Amount；After representing that cloud user obtains demand resource, the evaluation of estimate is given by j-th index of the resource；N_QRepresent that the service is commented The total degree of valency.Then, set up credit worthiness Matching Model：

In formula：V is the credit worthiness threshold values of cloud user's request setting.

Then constraint information Matching Model is：

Match(RR,RP)_QoS=ε₁Match(RR,RP)_T+ε₂Match(RR,RP)_C+ε₃Match(RR,RP)_Q (10)

In formula：ε₁, ε₂And ε₃The respectively weight coefficient of three binding target information.

1.4) comprehensive matching

It is ω to define the weight shared by the information matches of above three level_i(i=1,2,3), ω_i∈ [0,1], and ∑ ω_i =1.Son carries out comprehensive matching degree calculating to which according to the following formula：

Match(RR,RP)_{Comprehensive}=ω₁Match(RR,RP)_BaseInf+ω₂Match(RR,RP)_FunctionInf+ω₃Match(RR,RP)_QoS (11)

According to calculated comprehensive matching degree, again which is compared with setting threshold values, the resource clothes being filtrated to get Business set, as optimal service matching result collection, are available for user to select.

A kind of aforesaid cloud manufacturing recourses service multi-level semantic matching method, it is characterised in that：Described optimal service Matching result collection, is filtrated to get by matching, and according to the resource service collection of matching angle value size sequence.

The invention has the beneficial effects as follows：The present invention provides a kind of cloud manufacturing recourses and services multi-level semantic matching method, no It is only capable of being matched on grammer layer, and can be matched on semantic layer, for different information types, sets up different With function model, the quantitative matching degree calculated at all levels, the Service Matching of cloud manufacturing recourses is finally realized, money is improve The precision ratio and recall ratio of the matching of source service.

Description of the drawings

In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing Accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are only this Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, may be used also To obtain other accompanying drawings according to these accompanying drawings.

Fig. 1 is the Organization Chart of the present invention；

Fig. 2 is the matching flow chart of the present invention.

Specific embodiment

Below in conjunction with the accompanying drawings the preferred embodiments of the present invention are described in detail, so that advantages and features of the invention energy It is easier to be readily appreciated by one skilled in the art, apparent clearly defines so as to make to protection scope of the present invention.

Fig. 1 is that a kind of cloud manufacturing recourses service of the invention multi-level semantic matching method realizes framework.With reference to Fig. 1, cloud The multi-level semantic matching method of manufacturing recourses service realize framework mainly include three below in terms of：

(1) cloud service Requirement Acquisition：Based on B/S frameworks, cloud user submits to own resource to need by way of Web page Ask, cloud service platform extracts the key feature index (such as title, function, constraint etc.) of demand for services by resolver, is sent to Cloud service platform is stored.

(2) cloud manufacturing recourses ontology library：To realize the shared of Service Source in field, based on Ontology Language, cloud is manufactured Resource carries out semantic description in terms of essential information, function information and constraint information three, sets up resource ontology storehouse, so as to for clothes Business matching provides semantic information basis.

(3) cloud manufacturing service is resource matched：According to the semantic description of the demand for services of cloud user, based on cloud manufacturing recourses sheet Body storehouse, sets up cloud manufacturing recourses and services multi-level semantic matches model, mainly include essential information matching, function information matching and Constraint information is matched and comprehensive matching layer.Calculate every layer of matching degree and compare with setting threshold values, carry out service filtration, be more than or equal to Just match into next layer after setting threshold values, so repeatedly matching not only can reduce the matching primitives of next level after filtering Amount, is also obtained optimal service matching result collection.

Fig. 2 is the matching flow chart of the present invention, according to the demand for services of cloud user, by cloud manufacturing service resource according to corresponding Matched rule carry out Service Matching, concrete matching content and flow process are as follows：

Step1：Essential information is matched

Define 1：Cloud manufacturing service resource RP, cloud service demand RR of cloud user.

Cloud manufacturing service resource essential information is described as：RP_BaseInf=<rp₁,rp₂,rp₃,...,rp_n>；The cloud of cloud user Demand for services essential information is described as：RR_BaseInf=<rr₁,rr₂,rr₃,...,rr_m>.The title of demand for services is derived from body Notional word, based on this, measured by similarity between Ontological concept, set up essential information Matching Model：

In formula：Match(RP,RR)_BaseInfRepresent the essential information of the demand for services of cloud manufacturing service resource and cloud user Matching degree；N, m represent resource service essential information notional word number (n>=m)；Sim(rp_i,rr_i) for notional word rp_iAnd rr_i's Semantic similarity；λ_i(0≤λ_i≤ 1) represent Sim (rp_i,rr_i) weight coefficient；Dis(rp_i,rr_i) for notional word rp_iAnd rr_i's Semantic distance；Weight_jFor Anknupfungskeregriff word rp_iAnd rr_iWeighted value on shortest path；R is Anknupfungskeregriff word rp_iAnd rr_iMost The side number of short path；I(parent(c_j)) for target father's conceptual information amount；I(c_j) for sub- conceptual information amount；N ' is general for target father The quantity of the direct sub- concept read；p(c_j) represent Ontological concept word c_jProbability；Concept root is body root node, and its probability is 1。

Step2：Function information is matched

It is main to be matched in terms of input, output, premise and effect four.

Input/output (I/O) is matched

Input/output parameters are the formalized descriptions to cloud manufacturing service resource, and corresponding Semantic Similarity Measurement is all Carry out for Ontological concept, thus the Computing Principle of essential information notional word matching can be adopted.

Define 2：The |input paramete of cloud manufacturing recourses service describing integrates as InPS；The cloud service requirement description of cloud user it is defeated Enter and integrate as InPD.Element InPS of the matching i.e. from InPS₁Start to InPS_jRespectively with InPD in element InPD_jMatched Degree is calculated, and finds the maximum InPS of semantic matching degree more afterwards by calculating_i, then matching output.Based on this, foundation it is defeated Entering information matches model is：

In formula：InPS_InjRepresent that cloud manufacturing service resource issues j-th parameter of input；InPD_InjRepresent cloud service demand J-th parameter of input；n_IRepresent all parametric component numbers of input；Sim(RP,RR)_inputRepresent that notional word RP's and RR is defeated Enter semantic similarity, Match (InPS, InPD) represents the matching degree of InPS and InPD, Sim (InPS_Inj,InPD_Inj) represent InPS_InjAnd InPD_InjSemantic similarity.

In the same manner, output matching model can be obtained：

In formula：Sim(RP,RR)_outputRepresent the output semantic similitude of the demand for services of cloud manufacturing service resource and cloud user Degree, OutPS_OutjRepresent that Service Source issues j-th parameter of output；OutPD_OutjRepresent j-th ginseng of cloud service demand output Amount；n_ORepresent all parametric component numbers of output；Output parameter sets of the OutPS for cloud manufacturing recourses service describing, OutPD is cloud The output collection of the cloud service requirement description of user, Match (OutPS, OutPD) represent the matching degree of OutPS and OutPD, Sim (OutPS_Outj,OutPD_Outj) represent OutPS_OutjAnd OutPD_OutjSemantic similarity, Sim (RP, RR)_outputRepresent notional word The output semantic similarity of RP and RR.

Premise/effect (P/E, premise/effect) is matched

The property parameters of premise and effect (P/E) are not belonging to the concept in body domain, simply a kind of constraint expression formula, by individual Pronouns, general term for nouns, numerals and measure words object, grammatical term for the character predicate and parameter value parameter are constituted, and semantic description is<object,predicate, parameter>.Object and predicate are derived to class in resource ontology, attribute and example etc., base can be adopted In the computation model of Ontological concept information matches；It is for parameter is using the matching algorithm of numerical intervals, specific as follows：

In formula：Abbreviations of the pa for parameter；pa_RRFor the attribute value Operations of Interva Constraint of cloud users service needs；pa_RPFor The attribute value Operations of Interva Constraint of cloud manufacturing service resource；" | | " is parameter interval length of field, such as | 60,70 |=10, Sim (pa_RR, pa_RP) represent pa_RRWith pa_RPSemantic similarity.

Based on this, P/E Matching Models are set up：

α₁、α₂、α₃Represent；

Sim(RP,RR)_objectRepresent the distributive semantic similarity of notional word RP and RR, Sim (RP, RR)_predicateRepresent The grammatical term for the character semantic similarity of notional word RP and RR, Sim (RP, RR)_parameterRepresent the parameter value semanteme phase of notional word RP and RR Like degree；Sim(RP,RR)_P/ERepresent the premise or effect semantic similarity of notional word RP and RR.

Therefore function information Matching Model is：

In formula：δ₁, δ₂, δ₃, δ₄(δ₁,δ₂,δ₃,δ₄∈ [0,1]) input, output, premise, four aspects of effect are represented respectively The weight coefficient of functional attributes.Wherein, Match (RP, RR)_FunctionInfRepresent that the service of cloud manufacturing service resource and cloud user is needed The function information matching degree asked, Sim (RP, RR)_inputRepresent the input language of the demand for services of cloud manufacturing service resource and cloud user Adopted similarity, Sim (RP, RR)_outputThe output semantic similarity of the demand for services of cloud manufacturing service resource and cloud user is represented, Sim(RP,RR)_preconditionRepresent the premise semantic similarity of the demand for services of cloud manufacturing service resource and cloud user, Sim (RP,RR)_effectRepresent the effect semantic similarity of the demand for services of cloud manufacturing service resource and cloud user.

Step3：Constraint information is matched

The main time for including resource service, three constrained parameters of cost and credit worthiness, help cloud service platform each with this Side goes the service quality for weighing cloud resource.

Time, cost constraint matching

Time, cost this kind of index for resource service, mainly numerical value parameter, refer to values match algorithm, tool Body is as follows：

Meet condition, that is, time and cost that the time for servicing and cost are required less than or equal to party in request, that is, think It is 1 with degree, is otherwise 0.

Wherein, Match (RR, RP)_T/CFor the time or cost of cloud manufacturing service resource and the demand for services of cloud user Matching degree；Sim(T₁/C₁,T₂/C₂) for notional word time T₁With time T₂Semantic similarity or notional word cost C₁And C₂'s Semantic similarity.

Credit worthiness constrained matching

When credit worthiness constraint is evaluated, it is related to different resource and corresponding service-evaluating index, and in particular to soft Part resource, knowledge resource and human resourcess.Based on this, the computation model for setting up credit worthiness evaluation of estimate is：

In formula：I=1,2,3, software, three aspects of knowledge and human resourcess are represented respectively；N represents Service Source S_iEvaluate Index quantity；After representing that cloud user obtains demand resource, the evaluation of estimate is given by j-th index of the resource；N_QRepresenting should The total degree of service evaluation.Then, set up credit worthiness Matching Model：

In formula：V is the credit worthiness threshold values of cloud user's request setting.Q_RRFor the credit worthiness of cloud users service needs, Q_RPFor cloud The credit worthiness of manufacturing service resource.Q(S_i) for Service Source S_iCredit worthiness.

Then constraint information Matching Model is：

Match(RR,RP)_QoS=ε₁Match(RR,RP)_T+ε₂Match(RR,RP)_C+ε₃Match(RR,RP)_Q (10)

In formula：Match(RR,RP)_QoSRepresent the constraint information matching of the demand for services of cloud manufacturing service resource and cloud user Degree；Match(RR,RP)_TFor the time match degree of cloud manufacturing service resource and the demand for services of cloud user, Match (RR, RP)_CFor The cost matching degree of the demand for services of cloud manufacturing service resource and cloud user, Match (RR, RP)_QFor cloud manufacturing service resource and The prestige matching degree of the demand for services of cloud user, ε₁, ε₂And ε₃Respectively time, three binding target information of cost and credit worthiness Weight coefficient.

Step4：Comprehensive matching

It is ω to define the weight shared by the information matches of above three level_i(i=1,2,3), ω_i∈ [0,1], and ∑ ω_i =1.Comprehensive matching degree calculating is carried out to which, model is as follows：

Match(RR,RP)_{Comprehensive}=ω₁Match(RR,RP)_BaseInf+ω₂Match(RR,RP)_FunctionInf+ω₃Match(RR,RP)_QoS (11)

Wherein, Match (RR, RP)_{Comprehensive}Represent the synthesis of the demand for services of cloud manufacturing service resource and cloud user Matching degree, ω₁、ω₂、ω₃Essential information layer, function information layer and weight shared by constraint Information Level matching are represented respectively.

According to calculated comprehensive matching degree, again which is compared with setting threshold values, the resource clothes being filtrated to get Business set, as optimal service matching result collection, are available for user to select.

The above, the only specific embodiment of the present invention, but protection scope of the present invention is not limited thereto, any The change or replacement expected without creative work, should all be included within the scope of the present invention.Therefore, it is of the invention The protection domain that protection domain should be limited by claims is defined.

Claims (7)

1. a kind of cloud manufacturing recourses service multi-level semantic matching method, it is characterised in that：According to the cloud service demand of cloud user Description, sets up cloud manufacturing recourses ontology library；Cloud manufacturing recourses service multilamellar is set up on the basis of the cloud manufacturing recourses ontology library Secondary semantic matches model, is calculated every layer of matching degree and is compared with setting threshold values, carry out service filtration, obtain optimal service matching knot Fruit collects.

2. a kind of cloud manufacturing recourses according to claim 1 service multi-level semantic matching method, it is characterised in that：It is described Setting up cloud manufacturing recourses ontology library is：Using Ontology Language, to manufacturing recourses from essential information, function information and constraint information three Individual aspect carries out Ontology description, sets up resource ontology storehouse, so as to the demand for services for cloud user and cloud manufacturing service resource Between matching provide semantic information support.

3. a kind of cloud manufacturing recourses according to claim 1 service multi-level semantic matching method, it is characterised in that：It is described Set up cloud manufacturing recourses and service multi-level semantic matches model, calculate every layer of matching degree and compare with setting threshold values, serviced Filter, obtain optimal service matching result collection and be specially：

Essential information matching layer model is set up, function information matching layer model is set up, constraint information matching layer model is set up, is set up Comprehensive matching layer model, and carry out matching operation successively to each layer, then the threshold value by the result for calculating with setting is compared, only Have and just match into next layer more than or equal to after given threshold, finally according to the comprehensive matching degree that comprehensive matching layer model is calculated It is compared with given threshold, the resource service set being filtrated to get, as optimal service matching result collection.

4. a kind of cloud manufacturing recourses according to claim 3 service multi-level semantic matching method, it is characterised in that set up Essential information matching layer model is specially：

It is RP to define cloud manufacturing service resource, and the demand for services of cloud user is RR, sets up essential information Matching Model：

$\left\{\begin{array}{c}Match{(RP,RR)}_{BaseInf}=\underset{i=1}{\overset{m}{\&Sigma;}}{\mathrm{\&lambda;}}_i\&CenterDot;Sim({\mathrm{rp}}_i,{\mathrm{rr}}_i)\\ Sim({\mathrm{rp}}_i,{\mathrm{rr}}_i)=1-Dis({\mathrm{rp}}_i,{\mathrm{rr}}_i)\\ Dis({\mathrm{rp}}_i,{\mathrm{rr}}_i)=\underset{j=1}{\overset{r}{\&Sigma;}}{\mathrm{Weight}}_j\\ {\mathrm{Weight}}_j=\frac{1}{n^{\&prime;}}\frac{I\left(parent\right(c_j\left)\right)-I\left(c_j\right)}{I\left(parent\right(c_j\left)\right)+I\left(c_j\right)}\\ I\left(c_j\right)=-\log p\left(c_j\right)=-\log \frac{p\left(parent\right(c_j\left)\right)}{n^{\&prime;}}\\ p\left(root\right)=1\end{array}\right.---\left(1\right)$

In formula：Match(RP,RR)_BaseInfRepresent the essential information matching of the demand for services of cloud manufacturing service resource and cloud user Degree；N, m represent resource service essential information notional word number (n>=m)；Sim(rp_i,rr_i) for notional word rp_iAnd rr_iSemanteme Similarity；λ_i(0≤λi_i≤ 1) represent Sim (rp_i,rr_i) weight coefficient；Dis(rp_i,rr_i) for notional word rp_iAnd rr_iLanguage Adopted distance；Weight_jFor Anknupfungskeregriff word rp_iAnd rr_iWeighted value on shortest path；R is Anknupfungskeregriff word rp_iAnd rr_iIt is most short The side number in path；I(parent(c_j)) for target father's conceptual information amount；I(c_j) for sub- conceptual information amount；N ' is target father's concept Direct sub- concept quantity；p(c_j) represent Ontological concept word c_jProbability；Concept root is body root node, and its probability is 1.

5. a kind of cloud manufacturing recourses according to claim 4 service multi-level semantic matching method, it is characterised in that set up Function information matching layer model is specially：

Definition：The |input paramete of cloud manufacturing recourses service describing integrates as InPS, the input set of the cloud service requirement description of cloud user For InPD, input/output information matches model is set up, set up premise/effect Matching Model；

Input Matching Model：

$\begin{array}{c}Sim{(RP,RR)}_{input}=Match(InPS,InPD)\\ =\underset{j=1}{\overset{n_I}{\&Sigma;}}\frac{1}{n_I}Sim({\mathrm{InPS}}_{Inj},{\mathrm{InPD}}_{Inj})\end{array}---\left(2\right)$

Wherein, InPS_InjRepresent that cloud manufacturing service resource issues j-th parameter of input；InPD_InjRepresent the input of cloud service demand J-th parameter；n_IRepresent all parametric component numbers of input；

Output matching model：

$\begin{array}{c}Sim{(RP,RR)}_{output}=Match(OutPS,OutPD)\\ =\underset{j=1}{\overset{n_O}{\&Sigma;}}\frac{1}{n_O}Sim({\mathrm{OutPS}}_{Outj},{\mathrm{OutPD}}_{Outj})\end{array}---\left(3\right)$

Wherein, OutPS_OutjRepresent that Service Source issues j-th parameter of output；OutPD_OutjRepresent the output of cloud service demand J-th parameter；n_ORepresent all parametric component numbers of output；

Set up premise/effect Matching Model；

$\begin{array}{c}Sim{(RP,RR)}_{P/E}={\mathrm{\&alpha;}}_{1}Sim{(RP,RR)}_{object}+{\mathrm{\&alpha;}}_{2}Sim{(RP,RR)}_{predicate}\\ +{\mathrm{\&alpha;}}_{3}Sim{(RP,RR)}_{parameter}\end{array}---\left(5\right)$

Function information Matching Model is set up on the basis of input/output information matches model and premise/effect Matching Model：

$\begin{array}{c}Match{(RP,RR)}_{FunctionInf}={\mathrm{\&delta;}}_{1}Sim{(RP,RR)}_{input}+{\mathrm{\&delta;}}_{2}Sim{(RP,RR)}_{output}\\ +{\mathrm{\&delta;}}_{3}Sim{(RP,RR)}_{precondition}+{\mathrm{\&delta;}}_{4}Sim{(RP,RR)}_{effect}\end{array}---\left(6\right)$

Wherein, δ₁, δ₂, δ₃, δ₄(δ₁,δ₂,δ₃,δ₄∈ [0,1]) input, output, four aspect functions of premise and effect are represented respectively The weight coefficient of attribute.

6. a kind of cloud manufacturing recourses according to claim 4 service multi-level semantic matching method, it is characterised in that set up Constraint information matching layer model is specially：

Setup time, cost constraint Matching Model：

Set up credit worthiness evaluation of estimate computation model：

${\mathrm{QS}}_i=\underset{i=1}{\overset{3}{\&Sigma;}}\underset{j=1}{\overset{N}{\&Sigma;}}{Q}_i^j/N_Q---\left(8\right)$

Wherein, i=1,2,3, software, knowledge and human resourcess are represented respectively；N represents Service Source S_iEvaluation index quantity； After representing that cloud user obtains demand resource, the evaluation of estimate is given by j-th index of the resource；N_QRepresent the total of the service evaluation Number of times；

Credit worthiness Matching Model is set up on the basis of credit worthiness evaluation of estimate computation model：

$Match{(RR,RP)}_Q=Sim(Q_{RR},Q_{RP})=\left\{\begin{array}{cc}1& Q\left(S_i\right)\&GreaterEqual;\mathrm{\&upsi;}\\ 0& Q\left(S_i\right)<\mathrm{\&upsi;}\end{array}\right.---\left(9\right)$

Wherein, υ is the credit worthiness threshold values of cloud user's request setting；

Constraint information layer Matching Model is set up on the basis of time, cost constraint Matching Model and credit worthiness Matching Model：Match (RR,RP)_QoS=ε₁Match(RR,RP)_T+ε₂Match(RR,RP)_C+ε₃Match(RR,RP)_Q (10)

Wherein, ε₁, ε₂And ε₃The respectively weight coefficient of three constrained parameters indexs.

7. a kind of cloud manufacturing recourses according to claim 6 service multi-level semantic matching method, it is characterised in that set up Comprehensive matching model is specially：

Definition essential information matching, function information matching, weight shared by constraint information matching are ω_i(i=1,2,3), ω_i∈[0, , and Σ ω 1]_i=1, comprehensive matching model such as formula (11) is set up,

Match(RR,RP)_{Comprehensive}=ω₁Match(RR,RP)_BaseInf+ω₂Match(RR,RP)_FunctionInf+ω₃Match (RR,RP)_QoS(11)。