CN103888543B - Medical resource recommendation method and system based on Web services - Google Patents

Abstract

The invention discloses a medical resource recommendation method and system based on Web services. The medical resource recommendation method comprises the steps that demand information sent by clients is received, the demand information is matched with preset medical service resources packaged to be the Web services, a plurality of candidate Web services meeting the needs of the clients are determined, and a candidate Web service set is generated; a plurality of resource attribute values included by each candidate Web service are determined, and a service quality matrix of the candidate Web service set is constructed; the comprehensive service quality evaluation value of each candidate Web service is calculated according to the service quality matrix and an equalization service quality evaluation method, and the candidate Web service with the largest comprehensive service quality evaluation value is recommended to the clients. The information technology is utilized in the method and system, information estrangement between service supply and demands is eliminated, and equalization allocation of the service resources can be achieved as far as possible on the premise of meeting the individualized needs of different clients.

Description

The medical resource of sing on web service recommends method and system

Technical field

The present invention relates to the technical field of medical services resource distribution, particularly to the recommendation side of a kind of medical services resource Method and system.

Background technology

At present, medical services resource is distributed phenomenon unbalanced, that imbalance between supply and demand is nervous and is generally existed in medical treatment & health field: On the one hand, a lot of clients blindly pursue high-quality Service Source, cause partial service resource (such as Grade A hospital) to be in super negative Lotus operating condition, demand for services occurs congested and waits, service quality keeps falling, and client self has also paid unnecessary one-tenth This (time, price etc.)；On the other hand, partial service resource (community hospital) although service quality still can, but due to client Service Source information is not known about, causes this Service Source vacancy rate higher, it is impossible to be fully used.

Because medical services resource can not occur increasing substantially at short notice, close most so only demand assignment being given Suitable and non-optimal medical resource, just can make customer demand as much as possible be met, keep higher service quality simultaneously.

Along with the development of information technology, the medical treatment resource information of Different hospital is packaged by Web service and issues, logical Cross the Dynamic Matching between customer demand and medical services resource, recommend disclosure satisfy that the medical services money of its demand to user Source.Based on this, as how Web service is medium, eliminate the information estrangement between service supply and demand, it is achieved medical services resource equal Deng changing allotment, make full use of existing medical services resource, meet the individual demand of different client to greatest extent, have become as The problem that medical services industry needs solution badly.

Summary of the invention

In view of this, it is contemplated that the medical resource proposing the service of a kind of sing on web recommends method and system, to utilize Information technology, eliminates the information estrangement between service supply and demand, on the premise of the individual demand that disclosure satisfy that different client, to the greatest extent Realize the equalization allotment of medical services resource possibly.

First aspect, the medical resource that the invention discloses the service of a kind of sing on web is recommended method, is comprised the steps: Receive the demand information that client sends；Described demand information is carried out with the preset medical services resource being encapsulated as Web service Coupling, is determined for compliance with multiple candidate's Web services of customer demand, generates candidate's Web service set；Determine that each candidate Web takes Multiple source attribute values included by business, build the service quality matrix of described candidate's Web service set；According to described Service Quality Moment matrix and equalization QoS evaluating method, calculate the Integrated Services Quality evaluation of estimate of each described candidate's Web service, and And, by candidate's Web service recommendation maximum for Integrated Services Quality evaluation of estimate to client.

Second aspect, the invention discloses the medical resource commending system of a kind of sing on web service, including: receiver module, Candidate's Web service set generation module, service quality matrix generation module and recommending module.Receiver module is used for receiving client and sends out The demand information sent；Candidate's Web service set generation module is for being encapsulated as Web service by described demand information with preset Medical services resource is mated, and is determined for compliance with multiple candidate's Web services of customer demand, generates candidate's Web service set；Clothes Business mass matrix generation module, for determining the multiple source attribute values included by each candidate's Web service, builds described candidate The service quality matrix of Web service set；Recommending module is for commenting according to described service quality matrix and equalization service quality Valency method, calculates the Integrated Services Quality evaluation of estimate of each described candidate's Web service, and, by Integrated Services Quality evaluation of estimate Maximum candidate's Web service recommendation is to client.

The present invention is based on the coupling between client requirement information and medical services resource, by building equalization service choosing Select process, eliminate the unbalanced phenomena in Service Matching, on the one hand find for client and the most suitably service, improve expiring of client Meaning degree；On the other hand the utilization ratio of existing Service Source is improved as far as possible, it is achieved the load balancing of Service Source.

Accompanying drawing explanation

The accompanying drawing of the part constituting the present invention is used for providing a further understanding of the present invention, and the present invention's is schematic real Execute example and illustrate for explaining the present invention, being not intended that inappropriate limitation of the present invention.In the accompanying drawings:

Fig. 1 is the flow chart of steps of the medical resource recommendation embodiment of the method for sing on web of the present invention service；

Fig. 2 is that the medical resource of sing on web of the present invention service is recommended, in one embodiment of method, to determine candidate's Web service The flow chart of steps of service quality evaluation value；

Fig. 3 is that the medical resource of sing on web of the present invention service recommends one preferred embodiment of method；

Fig. 4 is that the medical resource of sing on web of the present invention service is recommended in method preferred embodiment, the step of service type coupling Rapid flow chart；

Fig. 5 is that the medical resource of sing on web of the present invention service is recommended in method preferred embodiment, the step of service name coupling Rapid flow chart；

Fig. 6 is that the medical resource of sing on web of the present invention service is recommended in method preferred embodiment, the step of service describing coupling Rapid flow chart；

Fig. 7 is a service scenarios schematic diagram faced by medical resource recommendation method of sing on web of the present invention service；

When Fig. 8 is customer demand number change, use two kinds of distinct methods (equalization methods and unequalization method) Service availability compares schematic diagram；

When Fig. 9 is customer demand number change, use two kinds of distinct methods (equalization methods and unequalization method) The demand sequence deadline compares schematic diagram；

When Figure 10 a is customer demand number change, use two kinds of distinct methods (equalization methods and unequalization method) Unequalization service quality (EnQoS) total value compare schematic diagram；

When Figure 10 b is customer demand number change, use two kinds of distinct methods (equalization methods and unequalization method) Integrated Services Quality (FuQoS) total value compare schematic diagram；

When Figure 11 is candidate service number change, use two kinds of distinct methods (equalization methods and unequalization method) Service availability compares schematic diagram；

When Figure 12 is candidate service number change, use two kinds of distinct methods (equalization methods and unequalization method) The demand sequence deadline compares schematic diagram；

When Figure 13 a is candidate service number change, use two kinds of distinct methods (equalization methods and unequalization method) Unequalization service quality (EnQoS) total value compare schematic diagram；

When Figure 13 b is candidate service number change, use two kinds of distinct methods (equalization methods and unequalization method) Integrated Services Quality (FuQoS) total value compare schematic diagram；

When Figure 14 is the change of equalization weight, use the clothes of two kinds of distinct methods (equalization methods and unequalization method) Business utilization rate compares schematic diagram；

When Figure 15 is the change of equalization weight, use the need of two kinds of distinct methods (equalization methods and unequalization method) The deadline is asked to compare schematic diagram；

When Figure 16 a is the change of equalization weight, use two kinds of distinct methods (equalization methods and unequalization method) Unequalization service quality (EnQoS) total value compares schematic diagram；

When Figure 16 b is the change of equalization weight, use two kinds of distinct methods (equalization methods and unequalization method) Integrated Services Quality (FuQoS) total value compares schematic diagram.

Figure 17 a is the structured flowchart of the medical resource commending system embodiment of sing on web of the present invention service；

Figure 17 b is in the medical resource commending system embodiment of sing on web of the present invention service, it is recommended that modular structure block diagram；

Figure 17 c is in the medical resource commending system embodiment of sing on web of the present invention service, and service quality matrix generates mould The structured flowchart of block.

Detailed description of the invention

It should be noted that in the case of not conflicting, the embodiment in the present invention and the feature in embodiment can phases Combination mutually.Describe the present invention below with reference to the accompanying drawings and in conjunction with the embodiments in detail.

The flow chart of steps of embodiment of the method, bag is recommended with reference to the medical resource that Fig. 1, Fig. 1 are sing on web of the present invention service Include following steps:

Receiving step S110, receives the demand information that client sends；

Candidate's Web service set generation step S120, by demand information and the preset medical services being encapsulated as Web service Resource is mated, and is determined for compliance with multiple candidate's Web services of customer demand, generates candidate's Web service set；

Service quality matrix generation step S130, determines the multiple source attribute values included by each candidate's Web service, structure Build the service quality matrix of candidate's Web service set；

Recommendation step S140, according to service quality matrix and equalization QoS evaluating method, calculates each candidate Web The Integrated Services Quality evaluation of estimate of service, and, give visitor by candidate's Web service recommendation maximum for Integrated Services Quality evaluation of estimate Family.

The present embodiment is based on the coupling between client requirement information and medical services resource, by building based on equalization The selection course of service quality evaluation, eliminates the unbalanced phenomena in Service Matching, and it is the most suitable on the one hand to find for client Service, improves the satisfaction of client；On the other hand the utilization ratio of existing Service Source is improved as far as possible, it is achieved Service Source Load balancing.

With reference to shown in Fig. 2, in one more specifically embodiment, when performing recommendation step S140, each candidate Web takes The Integrated Services Quality evaluation of estimate of business is determined as follows,

Step S1401, calculates each equalization service quality attribute value and this attribute respective weights of this candidate's Web service Product, and by all product addition, it is thus achieved that the equalization service quality evaluation value of this candidate's Web service；

Step S1402, calculates each unequalization service quality attribute value power corresponding with this attribute of this candidate's Web service The product of weight, and by all product addition, it is thus achieved that the unequalization service quality evaluation value of this candidate's Web service；

Step S1403, by equalization service quality evaluation value and the first multiplication, obtains the first product, and, by non- Equalization service quality evaluation value and the second multiplication, obtain the second product；First coefficient is the equalization of this candidate's Web service Change coefficient；Second coefficient is the unequalization coefficient of this candidate's Web service；Calculate the first product and second sum of products, as The Integrated Services Quality evaluation of estimate of candidate's Web service；

Wherein, the weight of each equalization service quality attribute value is arranged according to expertise, corresponding weight and It is 1；The weight of each unequalization service quality attribute value is arranged according to expertise, corresponding weight and be 1；First Coefficient and the second coefficient and be 1, if emphasizing the equalization of service resource allocation, then the first coefficient is more than the second coefficient, as Fruit does not emphasize the equalization of service resource allocation, then the first coefficient is less than the second coefficient.

The preferred embodiment of method is recommended to carry out the medical resource of sing on web service of the present invention below in conjunction with Fig. 3 to Fig. 6 Explain.

With reference to Fig. 3, the medical resource of sing on web service of the present invention recommends the preferred embodiment of method to comprise the steps:

Step 1: receive customer demand and (inputted by network or telephone counseling etc., and be converted to order form service requirement information Table.

(1) ST represents client requirement information

ST=<Category, Name, Profile, Inputs, Outputs, Constraint>

Wherein, Category represents the service type belonging to service；Name represents the title of service；Profile represents clothes The description information of business；Inputs is the input variable set of service；Outputs represents the output variable set of service； The Constraint i.e. client quality of service requirement to required service, such as medical level, price, deadline etc., belong to hard Constraints.

Step 2: the preset medical services resource being encapsulated as Web service is stored in data base 1, to its traversal, solves Analyse and obtain its each parameter information.Specifically include following steps:

Traversal package is the medical services resource of Web service, resolves its semantic letter according to its element structure describing language Breath.

(1) CS represents the Web service information of candidate

CS=<Category, Name, Profile, Inputs, Outputs, QoS>

Wherein, Category represents the service type belonging to candidate service；Name represents the title of candidate service； Profile represents the description information of candidate service；Inputs is the input variable set of candidate service；Outputs represents candidate The output variable set of service；The current service quality of the i.e. candidate service of QoS.

Step 3: according to client requirement information and the parameters of candidate's Web service, carry out service type, service name, First coupling of service describing, obtains first candidate service set.

Specifically, when carrying out the coupling of service type, qualified Web service is stored in data base 2, carries out During the coupling of service name, qualified Web service is stored in data base 3；When carrying out the coupling of service describing, will symbol The Web service of conjunction condition is stored in data base 4；Then, take the common factor of these three data base, be stored in data base 5.

Matching degree is mainly weighed according to the semantic distance between service concept, by two concepts in body network The beeline of respective nodes weighs their semantic relevancy, computational methods as shown in Equation 1:

$Sim(C_{ST},C_{CS})=\frac{\mathrm{\&alpha;}}{{\mathrm{\&beta;}}^{dis(C_{ST},C_{CS})}+\mathrm{\&alpha;}-1}---\left(1\right)$

Wherein, ST represents that client requirement information, CS represent candidate's Web service set, C_ST, C_CSRepresent ST's Yu CS respectively Ontological concept (as follows)；α is correction parameter；β is that semantic relevancy change controls parameter, typically takes natural constant e；In formula Semantic distance dis (C_ST,C_CS), according to the different relations between Ontological concept, shown in computational methods such as formula (2):

$dis(C_{ST},C_{CS})=\left\{\begin{array}{cc}0& C_{ST}=C_{CS}\\ \underset{i}{\&Sigma;}{w}_i& Others\end{array}\right.---\left(2\right)$

Wherein, i represents when two concept non-equivalences, finds out and connects the shortest path of two concepts in body；Represent the On i paths, the relation weighted value sum on all limits, is corresponding semantic distance.

Below, service type, service name, the coupling of service describing are done and described in detail further.

(1) service type coupling.

With reference to Fig. 4, resolve the service type semantic information in ST Yu CS, travel through in CS according to the service type demand in ST Respectively service and calculate its semantic distance, if kth candidate service and ST service type semantic distance Sim_category (ST.Category,CS_k.Category) less than matching degree threshold value u1, then it is deposited in data base 2.Then proceed to calculate K+1 candidate service semantic distance, the like.Data base 2 has all candidate Web clothes meeting service type coupling Business.

(2) service name coupling.

With reference to Fig. 5, resolve the service name semantic information in ST Yu CS, travel through in CS according to the service name demand in ST Each service.If the kth service name in CS storehouse and service name semantic matches degree of association Sim in ST_name(ST.Name, CS_k.Name) less than matching degree threshold value u2, then data base 3 it is deposited into.Then proceed to calculate+1 candidate service semanteme phase of kth Guan Du, the like.Data base 3 has all candidate service meeting service name.

(3) service describing coupling

With reference to Fig. 6, resolve the service describing semantic information in ST Yu CS, travel through in CS according to the service describing demand in ST Each service, if the k-th service describing in CS storehouse and service describing semantic matches degree of association Sim in ST_profile (ST.Profile,CS_k.Profile) less than matching degree threshold value u3, then data base 4 it is deposited into.Then proceed to calculate kth+1 Individual candidate service semantic relevancy, the like.Data base 4 has all candidate service meeting service describing.Based on right After service type, service name, service describing coupling, gained candidate service takes its common factor and is stored in data base 5, i.e. by data base 2, the candidate's Web service in data base 3, data base 4 takes common factor, stores first candidate's Web service set in data base 5.

Step 4: update the most relevant, by QoS (the Quality of of medical services to the running status of service according to it Service) attribute is divided into two classes:

(1) the QoS attribute that equalization is relevant: this kind of QoS attribute directly reflects the current operating conditions of medical services, each Cycle all can change, and directly influences the equalization effect that follow-up service selects, so being referred to as the QoS that equalization is relevant Attribute, mainly includes the response time of service, current service ability, reliability, availability etc..

(2) the QoS attribute that equalization is unrelated: this kind of QoS attribute mainly reflects the fixed attribute of medical services, does not reflect doctor Treating the current operating conditions of service, the frequency that its change occurs is relatively low, will not select to produce impact on follow-up service, so It is referred to as the QoS attribute that equalization is unrelated, mainly includes medical level, service price, maximum service ability, reputation degree etc..

From data base 5, read candidate's Web service one by one, equalization QoS attribute is updated, obtains each resource The currently available state of service.Including,

● update current service ability Capacity of service_cur；

Current service ability Capacity of medical services_curDetermine whether these medical services can immediately begin to as visitor Family services.As shown in Equation 3, Capacity_curRepresent the current service ability of these medical institutions；Capacity_maxRepresent this doctor Treat the maximum service ability of mechanism, the most relevant to the medical resource of these medical institutions (such as sick bed quantity, doctor's quantity etc.)； SumPatient_curRepresent the sufferer number that these medical institutions are currently treated.Because after each cycle, SumPatient_cur All it may happen that change (patient discharge the most treated, or new patient enter the waiting list of service), thus Cause Capacity_curChange the most therewith.

Capacity_cur=Capacity_max-SumPatient_cur (3)

Wherein, Capacity_curValue be broadly divided into following several situation: if 1. SumPatient_cur≤ Capacity_max, then Capacity_cur＞ 0, represents that this service patient for newly arriving can provide service immediately；If 2. SumPatient_cur＞ Capacity_max, then Capacity_cur=0, represent that new client need waits, or select other Medical services.

● update the response time Rtime of service；

(the patient discharge the most treated, or the newly because waiting list of each medical services constantly changes Patient enter waiting list), so the response time of this service also can change therewith.As shown in Equation 4, it is assumed that at present These medical services have the medical resource of k unit, i.e. Capacity_max=k；Rtime represents the response time point of this service； CurTime represents current time point；TimePatient_iRepresent that i-th medical resource needs to continue as its sufferer and provides treatment Time, and 1≤i≤k.Because after each cycle, TimePatient_iAll it may happen that change, cause Rtime also with Change.

Rtime=CurTime+Min{TimePatient₁,TimePatient₂,......,TimePatient_k} (4)

If there being TimePatient_iNumerical value is 0, then Rtime=CurTime, indicates that medical resource is idle, can be with horse Upper offer services, and response time is 0；Without TimePatient_iNumerical value is 0, then response time needs to count according to formula 4 Calculating, until there being medical resource to vacate, being only possible to provide service.

● update the waiting time Wcycle of service；

Wcycle=Rtime-CurTime (5)

As shown in Equation 5, Wcycle represents that sufferer selects the time that these service needs wait, namely response time with work as The difference of front time.As Wcycle=0, show that the waiting list of this service does not has demand, if selecting this service, permissible Obtain response immediately；When in the waiting list that certain services, demand number increases, service response time also can be the most elongated.

● update the deadline Ftime of service；

For certain medical services, if accept a new sufferer, its deadline mainly by this demand etc. Treat that the treatment time of time and patient determines.As shown in Equation 6, Rtime represents that the response time of new patient is accepted in this service Point；Ftime represents the deadline after this service undertaking new patient；TimePatient_newRepresent the required acceptance of new patient The medical services time, it is generally dependent on the coincident with severity degree of condition of patient and the service level of doctor.

Ftime=Rtime+TimePatient_new (6)

● update CSAT Satisfaction of service；

The factor affecting CSAT mainly has the waiting time Wcycle of customer demand, treatment time TimePatient_new.CSAT is with demand for services Wcycle and TimePatient_newIncrease and decline.In addition, The factor affecting CSAT also has a lot, credit worthiness Reputation etc. such as serviced.Therefore the meter of CSAT Calculation formula is:

$Satisfaction=\frac{\mathrm{\&alpha;}\&times;\mathrm{Re}putation}{\mathrm{\&beta;}\&times;Wcycle+\mathrm{\&gamma;}\&times;{\mathrm{TimePatient}}_{new}}---\left(8\right)$

Wherein Reputation represents the credit worthiness of this service, this attribute be in the regular period user to these medical services Evaluate feedback, the most stable, therefore using this attribute as the standing part affecting CSAT；α, β and γ are respectively Represent the factor weight affecting CSAT, and α ＞ 0；β ＞ 0；γ ＞ 0；Alpha+beta+γ=1.

● update other equalizations QoS attribute of service；The calculating of other attributes can be according to its feature by similar Process is derived from.

Step 5: from data base 5, reads candidate's Web service, one by one according to demand for the hard constraint of some QoS attribute Restriction is screened, the higher limit of such as price or response time, the lower limit etc. of reputation.Such as, for single service Speech, if meeting this condition, then it represents that this candidate service is available, otherwise will filter out this service.Ultimately form new time Select set of service, put in data base 6.

Step 6: from data base 6, reads candidate's Web service one by one, constructs the equalization service of candidate service set Quality (QoS) matrix.QoS attribute in service matrix is divided into two classes: unequalization quality of service attribute (EnQoS attribute) and equal Deng change quality of service attribute (EqQoS attribute).

$Q=\left(\begin{array}{ccccccc}{q}_{1,1}& q_{1,2}& \mathrm{...}& q_{1,k}& q_{1,k+1}& \mathrm{...}& q_{1.m}\\ q_{2,1}& q_{2,2}& \mathrm{...}& q_{2,k}& q_{2,k+1}& \mathrm{...}& q_{2,m}\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ q_{n,1}& q_{n,2}& \mathrm{...}& q_{n,k}& q_{n,k+1}& \mathrm{...}& q_{n,m}\end{array}\right)$

The transverse dimensions of service quality matrix Q represents that the QoS attribute of candidate service, longitudinal dimension represent the sequence of candidate service Number；q_ijRepresent the jth service quality attribute value of i-th candidate's Web service；q_i,1q_i,2...q_i,kRepresent i-th candidate Web clothes The equalization quality of service attribute of business, quantity is k；q_i,k+1q_i,k+2...q_i,mRepresent unequalization of i-th candidate's Web service Quality of service attribute, quantity is m-k.

Step 7: service quality (QoS) matrix of candidate service set is normalized.

Because different QoS attribute type is different, span is the most different, so needing to return before calculating qos value One change processes.For medical services, QoS attribute is broadly divided into profit evaluation model and cost type.Profit evaluation model attribute refers to that property value is more The best big attribute, such as medical level, maximum service ability etc.；Cost type attribute refers to the attribute that property value is the smaller the better, Such as service price, response time, deadline etc..

Formula (7) and formula (8) are used for being standardized each value in matrix Q, such that it is able to obtain normalized Service quality matrix Q'.

Wherein, vector N={n is used₁,n₂,......,n_i,n_i+1,......n_m(1≤i≤m) be used for distinguishing this service The type of each QoS attribute, wherein n_iValue can be 1 or 0；Work as n_iWhen=1, represent that this attribute is profit evaluation model, i.e. QoS It is worth the biggest service quality the best；Work as n_iWhen=0, represent that this attribute belongs to the least service quality of cost type, i.e. qos value the best； Vector C={c₁,c₂,......,c_i,c_i+1,......c_m(1≤i≤m) when representing each QoS property value standardization of this service Maximum standard value；q_ijRepresent the jth service quality attribute value of i-th candidate's Web service；Represent service quality square The meansigma methods of n candidate service jth QoS attribute in battle array Q.

$Q^{\&prime;}=\left(\begin{array}{ccccccc}{v}_{1,1}& v_{1,2}& \mathrm{...}& v_{1,k}& v_{1,k+1}& \mathrm{...}& v_{1.m}\\ v_{2,1}& v_{2,2}& \mathrm{...}& v_{2,k}& v_{2,k+1}& \mathrm{...}& v_{2,m}\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ v_{n,1}& v_{n,2}& \mathrm{...}& v_{n,k}& v_{n,k+1}& \mathrm{...}& v_{n,m}\end{array}\right)$

Wherein, v_ijRepresent the jth service quality attribute value of the i-th candidate's Web service after normalization.v_i,1v_{i, 2}...v_i,kRepresenting the equalization service quality attribute value of the i-th candidate's Web service after normalization, quantity is k；v_{i,k+ 1}v_i,k+2...v_i,mRepresenting the unequalization service quality attribute value of the i-th candidate's Web service after normalization, quantity is m-k.

Step 8: from data base 6, reads candidate's Web service one by one, according to equalization QoS judgement schematics, calculates every The QoS comprehensive evaluation value of individual candidate service, until by complete for all candidate service coupling.Finally, QoS comprehensive evaluation value is selected Big service is recommended.

In the QoS evaluation model of equalization, not only to consider this service unequalization QoS evaluation of estimate, in addition it is also necessary to consider The equalization QoS evaluation of estimate of this service.The unequalization QoS evaluation of estimate of service typically keeps stable, for reflecting this service General Properties；The equalization QoS evaluation of estimate of service is as servicing the change of current state and changing, and is used for reflecting this service Current state.

Calculating i-th candidate service equalization service quality evaluation value formula is:

Wherein, EqQoS (CS_i) represent i-th candidate's Web service equalization service quality evaluation value, v_i,jRepresent i-th The jth equalization QoS property value of individual service,Represent the weight shared by jth equalization quality of service attribute,And

The formula calculating i-th candidate service unequalization service quality evaluation value is:

$EnQoS\left({\mathrm{CS}}_i\right)=\underset{j=1}{\overset{m-k}{\&Sigma;}}{v}_{i,k+j}\&times;{\mathrm{\&psi;}}_{k+j}---\left(12\right)$

Wherein, EnQoS (CS_i) represent i-th candidate's Web service unequalization service quality evaluation value；v_i,k+jRepresent The jth unequalization QoS property value of i-th candidate's Web service, ψ_k+jRepresent shared by jth unequalization quality of service attribute Weight, 0≤ψ_k+j≤ 1, and ψ_k+1+ψ_k+2+…+ψ_m=1.

The Integrated Services Quality comprehensive evaluation value formula calculating i-th candidate service is:

FuQoS(CS_i)=a × EqQoS (CS_i)+b×EnQoS(CS_i) (13)

Wherein, FuQoS (CS_i) represent i-th candidate's Web service Integrated Services Quality evaluation of estimate；, EqQoS (CS_i) table Show the equalization service quality evaluation value that i-th services；EnQoS(CS_i) represent that the unequalization service quality of i-th service is commented It is worth；A represents the weight that equalization service quality evaluation value is shared in a model, and b represents unequalization service quality evaluation value Weight shared in a model, and a+b=1, a ＞ 0, b ＞ 0.

In addition, it can include:

Step 9: after service system completes M demand the numerical value of sets itself M as required (client can), need The effect of services selection is evaluated feedback.Then, according to feedback result, determine how to adjust equalization QoS model Weight coefficient, realizes desired equal services degree.If it is intended to equalization intensity is high, just increase the value of a, reduce b Value；If contrary equalization degree to be weakened, it is reduced by the value of a, increases the value of b.

The effect assessment of equalization services selection mainly uses three indexs:

(A) Service Source utilization rate

$Use\_Ratio=\frac{\underset{i=1}{\overset{m}{\&Sigma;}}{\mathrm{QoES}}_{i,M-Capacity}}{\underset{i=1}{\overset{m}{\&Sigma;}}{\mathrm{QoES}}_{i,M-Capacity}+\underset{j=m+1}{\overset{n}{\&Sigma;}}{\mathrm{QoES}}_{j,M-Capacity}}---\left(14\right)$

Wherein, Use_Ratio represents that system processes the Service Source utilization rate of this demand sequence, equal to " being used service Service ability and " with the ratio of " service ability of all candidate service and "；Molecule is in candidate service set, processes this need The maximum service ability sum of selected service when seeking sequence (assuming that quantity is m)；Denominator is all candidate service resource collections In, the maximum service ability sum of all services (assuming that quantity is n).

(B) deadline of demand sequence

Finish_Time=Max{Patient_1,Ftime,Patient_2,Ftime,......,Patient_n,Ftime} (15)

Wherein, Finish_Time represents the final time point that whole demand sequence has been processed, complete equal to each demand Maximum in the one-tenth time；Patient_i,FtimeRepresent " deadline " (1≤i≤n) of i-th demand in demand sequence.

(C) the QoS total value of selected service

$Sum\_QoS=\underset{i=1}{\overset{n}{\&Sigma;}}{\mathrm{QoS}}_i---\left(16\right)$

Wherein, Sum_QoS represented whole demand sequence each service qos value summation；QoS_iRepresent and needed Seek the service QoS (1≤i≤n) of i-th demand in sequence.

Technique effect is analyzed:

In order to compare unequalization method and the equalization algorithm performance difference when processing demand sequence, design three kinds of realities Test scene: 1. candidate's medical services number is fixed, along with the increase of demand for services number, the performance difference of algorithms of different, including The deadline of demand sequence, the overall qos value of institute's recommendation service, and the utilization rate of candidate service resource；2. the need serviced Ask number to fix, along with the increase of candidate's medical services number, the performance difference of algorithms of different, including demand sequence when completing Between, the overall qos value of institute's recommendation service, and the utilization rate of candidate service resource；3. candidate's medical services and demand for services Number is the most fixing, changes the weight coefficient in equalization QoS model, the performance difference of algorithms of different, complete including demand sequence One-tenth time, the overall qos value of institute's recommendation service, and the utilization rate of candidate service resource.

Case scene is as shown in Figure 7.

Assuming that need one demand sequence of sequential processing: sufferer demand randomly generates, including the medical time demand of sufferer, And the QoS evaluation model of sufferer；According to the demand of sufferer, existing medical services resource is evaluated and selects, then This demand is placed in most suitable service waiting list；Finally, according to CSAT and the number of service resource utilization According to analysis, equalization QoS model is carried out feedback adjustment, thus continues to optimize the utilization rate of resource.Concrete Setup Experiments is such as Under:

● the stochastic generation of candidate service

Stochastic generation medical institutions type: from-1, and randomly choose a number in 0,1}, as the type of medical institutions, its In-1 represent larger medical mechanism, 0 represents medium-sized medical institutions, and 1 represents small medical mechanism.

Stochastic generation medical institutions maximum service ability: according to medical institutions' type, generates corresponding maximum service ability, If larger medical mechanism, then its maximum service ability is an integer of stochastic generation between [30,50]；If it is medium-sized Medical institutions, then its maximum service ability is an integer of stochastic generation between [20,30]；If small medical mechanism, Then its maximum service ability is an integer of stochastic generation between [10,20].

The unequalization QoS attribute of stochastic generation service: the Price of these medical services is stochastic generation between [10,15] An integer；Quality is an integer of stochastic generation between [Isosorbide-5-Nitrae], wherein Quality={1 (typically), 2 (good), 3 (preferably), 4 (the best) }；Reputation is integer, wherein a Reputation={1 of stochastic generation between [1,5] (poor), 2 (typically), 3 (high), 4 (higher), 5 (the highest) }.The unequalization QoS property value of these medical services is at stochastic generation Afterwards, it is changeless.

Stochastic generation service equalization QoS attribute: the unequalization QoS attribute of service mainly include Wcycle with Satisfaction, after program brings into operation, its value is dynamically change during program is run. The span of Satisfaction is { 1 (being unsatisfied with), 2 (typically), 3 (satisfactions), 4 (relatively satisfactory), 5 (the most satisfied) }.

● the stochastic generation of demand sequence

Stochastic generation sufferer demand: assuming that randomly generate the demand for services queue that number is N, each demand is comprised Patient's number is stochastic generation between [50,200] integer, and suppose that treatment time of each patient is 1 time Cycle.

The hard constraint condition of stochastic generation demand: the patient in each demand has following hard requirement, the upper limit of Price For an integer of stochastic generation between [10,15]；The lower limit of Quality is an integer of stochastic generation between [1,4]；Clothes Lower limit is stochastic generation between [1,5] integer of business Reputation；All of attribute is required for head before the computation First it is normalized.

The QoS evaluation model of stochastic generation unequalization algorithm: unequalization algorithm uses unequalization QoS model to comment Valency Service Source, its formula is EnQoS=w1*price+w2*capacity+w3*quality+w4*reputation, wherein Weight w1 corresponding to each attribute, w2, w3, w4 stochastic generation, and w1 > 0, w2 > 0, w3 > 0, w4 > 0, w1+w2+w3+w4=1.

The QoS evaluation model of stochastic generation equalization algorithm: equalization algorithm uses comprehensive QoS model to evaluate service money Source, its formula is FuQoS=a*EqQoS+b*EnQoS=a* (w5/ (1+Wcycle)+w6*Satisfaction)+b* (w1* Price+w2*capacity+w3*quality+w4*reputation), wherein equalization attribute is corresponding weight w5, w6 are random Generate, and w5 > 0, w6 > 0, w5+w6=1；A and b represents equalization quality of service attribute and unequalization quality of service attribute Weight in evaluation model, is set according to the needs of case study, and a ＞ 0, b ＞ 0, a+b=1；As a ＞ b, Show that service evaluation emphasizes equalization quality of service attribute；As a ＜ b, show that service evaluation emphasizes that unequalization service quality belongs to Property.

● the matching process of service sequences

According to practical situation, new demand is continuously added in demand queue in a steady stream, it is assumed that every 1 time cycle with Machine generates a demand, is then forwarded to service system and goes to process.Different algorithms uses different QoS evaluation models to select Medical services process this demand.

It is respectively adopted unequalization (EnQoS model) and equalization (FuQoS model) selects corresponding medical services to provide Source, according to the emulation experiment of design, verifies the algorithm proposed, is concentrated mainly on the analysis of three indexs and compares: Service Source utilization rate, whether checking equalization algorithm can beneficially improve the service efficiency of service；Deadline, checking is all Deng change algorithm while increasing operation rate, if there will be the significantly prolongation of demand deadline；Service quality, checking equalization Change while improving service availability, if there will be being greatly reduced of service quality.

Experiment 1: the number change of demand for services

This experiment supposes that candidate service number is fixed, and demand for services number is continuously increased, and compares the poor performance of algorithms of different Different, including utilization rate, the deadline of demand sequence of candidate service resource, and the overall QoS mass of institute's recommendation service.This In to arrange the value of demand for services number be 10,20,30,40,50,60,70,80,90,100,150,200,300；Candidate service Number value is 100, keeps constant；The weighted value of equalization QoS index is 0.6, and the weighted value of unequalization QoS index is 0.4, keep constant.

As shown in Figure 8, along with being continuously increased of demand for services number, the Service Source utilization rate under unequalization algorithm begins It is maintained at about 10% eventually；Service Source utilization rate under equalization algorithm is obviously improved, and along with demand for services Quantity increases and presents increasing trend.This shows, equalization algorithm has clear superiority on lifting service availability.

As it is shown in figure 9, being continuously increased along with demand for services number, the demand for services deadline under equalization algorithm is bright Aobvious less than the deadline under unequalization algorithm；After demand for services quantity is more than 100, the deadline of two kinds of algorithms is poor Value expands rapidly.This shows, equalization algorithm has clear superiority in terms of reducing the demand sequence deadline.

As as-shown-in figures 10 a and 10b, along with the increase of demand for services number, no matter use EnQoS model or FuQoS Model, the QoS total value of two kinds of algorithms all occurs in that the trend of sharp increase.Compare from customer perspective, although unequal Change algorithm and be slightly above equalization algorithm, but the difference between them is the most little.Compare from comprehensive angle, unequalization algorithm QoS total value and equalization calculate between gap it is obvious that and along with the increase of demand for services quantity is increasing.This table Bright, equalization algorithm, in the case of chosen quality of service not being greatly lowered, substantially increases the overall profit of Service Source By rate.

Experiment 2: the number change of candidate service

This experiment supposes that the demand number of service is fixed, along with the increase of candidate service number, the poor performance of algorithms of different Different, including deadline, the overall QoS mass of institute's recommendation service of demand sequence, and the utilization rate of candidate service resource.This In to arrange the number of candidate service be 20,40,60,80,100,120,140,160,180,200；Demand for services is 50, keeps Immobilize；The weighted value of equalization QoS index is 0.6, and the weighted value of unequalization QoS index is 0.4, keeps fixing not Become.

As shown in figure 11, along with increasing of Service Source, the Service Source utilization rate of equalization algorithm apparently higher than non-all Deng changing algorithm, when Service Source number is less than 50, service availability is all more than 50%.Under unequalization Service Matching algorithm Service availability is the most on the low side, even if in the case of 5 times of candidate service number of resources of demand for services number, Service Source Utilization rate also only has more than 20%.This shows, equalization algorithm has clear superiority on lifting service availability.

As shown in figure 12, along with the increase of candidate service number of resources, the deadline of two kinds of algorithms does not has obvious ripple Dynamic, under equalization algorithm, the deadline of demand for services sequence is generally far below unequalization algorithm.This shows, equalization is calculated Method has clear superiority in terms of the sequence deadline of demaning reduction.

As shown in Figure 13 a and Figure 13 b, along with the increase of candidate service number, no matter use EnQoS model or FuQoS Model, the QoS total value of two kinds of algorithms all occurs in that the trend slightly increased.From customer perspective, unequalization algorithm EnQoS total value is slightly above equalization algorithm, but gap is little.From comprehensive angle, the FuQoS summation of equalization algorithm Value is substantially better than unequalization algorithm.This shows, equalization algorithm in the case of chosen quality of service not being greatly lowered, Substantially increase the overall utilization rate of Service Source.

Equalization weight change in experiment 3:QoS model

This experiment supposes that the number of candidate service and demand for services is the most fixing, changes in equalization QoS model (FuQoS) Weight coefficient, the performance difference of algorithms of different, including deadline, the overall QoS mass of institute's recommendation service of demand sequence, with And the utilization rate of candidate service resource.Here the weighted value arranging equalization QoS index takes 0.1,0.2,0.3,0.4 respectively, 0.5,0.6,0.7；Demand for services is 50, and candidate service number is 100, all keeps constant.

As shown in figure 14, along with the increase of equalization QoS Attribute Weight weight values, the Service Source utilization rate under equalization algorithm Present the trend of rising appreciably, rather than the Service Source utilization rate of equalization algorithm changes not quite relatively.This shows, equalization algorithm Promoting, there is on service availability clear superiority, and the weighted value of equalization can be adjusted as required by, thus flexibly The utilization rate of ground allotment integrity service resource.

As shown in figure 15, when equalization QoS attribute weight is less, the deadline difference of demand for services under two kinds of algorithms Less.When equalization QoS attribute weight increases, the demand for services deadline of equalization algorithm declines rapidly.This shows, all The deadline of demand for services can be effectively reduced Deng change algorithm.

As shown in fig. 16 a and 16b, from customer perspective, along with the increase of equalization weight, unequalization algorithm EnQoS value summation is slightly above equalization algorithm.From comprehensive angle, along with the increase of equalization weight, equalization algorithm The change of FuQoS value is little, but unequalization algorithm FuQoS value is but remarkably decreased.This shows, selected by unequalization algorithm The Service Source selecting out, performs poor in equal services index.Equalization algorithm be not greatly lowered selected clothes In the case of business quality, substantially increase the overall utilization rate of Service Source.

Second aspect, the invention discloses the embodiment of the medical resource commending system of a kind of sing on web service, with reference to figure 17a.Including: receiver module 171, candidate's Web service set generation module 172, service quality matrix generation module 173 and recommendation Module 174.Receiver module 171 is for receiving the demand information that client sends；Candidate's Web service set generation module 172 is used for Described demand information is mated with the preset medical services resource being encapsulated as Web service, is determined for compliance with customer demand Multiple candidate's Web services, generate candidate's Web service set；Service quality matrix generation module 173 is used for determining each candidate Multiple source attribute values included by Web service, build the service quality matrix of described candidate's Web service set；Recommending module 174, for according to described service quality matrix and equalization QoS evaluating method, calculate each described candidate's Web service Integrated Services Quality evaluation of estimate, and, by candidate's Web service recommendation maximum for Integrated Services Quality evaluation of estimate to client.

The present embodiment is based on the coupling between client requirement information and medical services resource, by building based on equalization The selection course of service quality evaluation, eliminates the unbalanced phenomena in Service Matching, and it is the most suitable on the one hand to find for client Service, improves the satisfaction of client；On the other hand the utilization ratio of existing Service Source is improved as far as possible, it is achieved Service Source Load balancing.

With reference to Figure 17 b, it is further preferred that above-mentioned recommending module 174 includes: the first computing unit 1741, second calculates Unit 1742 and service quality evaluation value acquiring unit 1743.

Wherein, the first computing unit 1741, for calculating each equalization service quality attribute value of this candidate's Web service With the product of this qualitative attribute respective weights, and by all product addition, it is thus achieved that the equalization evaluation of estimate of this candidate's Web service. Second computing unit 1742, for calculating each unequalization service quality attribute value and this attribute pair of this candidate's Web service Answer the product of weight, and by all product addition, it is thus achieved that the unequalization service quality evaluation value of this candidate's Web service.Service Quality evaluation value acquiring unit 1743, for by equalization service quality evaluation value and the first multiplication, obtaining the first product, And, by unequalization service quality evaluation value and the second multiplication, obtain the second product；First coefficient is this candidate Web The equalization coefficient of service；Second coefficient is the unequalization coefficient of this candidate's Web service；Calculate the first product and the second product Sum, as the Integrated Services Quality evaluation of estimate of candidate's Web service.

Wherein, each equalization service quality attribute value of this candidate's Web service is corresponding weight, this candidate's Web service Corresponding weight, the first coefficient and the second coefficient of each unequalization service quality attribute value is and presets；Each is impartial The weight changing service quality attribute value is arranged according to expertise, corresponding weight and be 1；Each unequalization Service Quality The weight of amount property value is arranged according to expertise, corresponding weight and be 1；First coefficient and the second coefficient and be 1.

Further, with reference to Figure 17 c, service quality matrix generation module 173 farther includes: service quality matrix generates Unit 1731, normalized unit the 1732, the 3rd computing unit the 1733, the 4th computing unit 1734 and the 5th computing unit 1735。

Service quality matrix generates single 1731 for generating the service quality matrix of candidate's Web service set

$Q=\left(\begin{array}{ccccccc}{q}_{1,1}& q_{1,2}& \mathrm{...}& q_{1,k}& q_{1,k+1}& \mathrm{...}& q_{1.m}\\ q_{2,1}& q_{2,2}& \mathrm{...}& q_{2,k}& q_{2,k+1}& \mathrm{...}& q_{2,m}\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ q_{n,1}& q_{n,2}& \mathrm{...}& q_{n,k}& q_{n,k+1}& \mathrm{...}& q_{n,m}\end{array}\right)$

Wherein, q_ijRepresent the jth quality of service attribute of i-th candidate's Web service；q_i,1q_i,2...q_i,kRepresent i-th The equalization quality of service attribute of candidate's Web service, quantity is k；q_i,k+1q_i,k+2...q_i,mRepresent i-th candidate's Web service Unequalization quality of service attribute, quantity is m-k.

Normalized unit 1732 is used for being normalized service quality matrix Q, the clothes after normalized Business mass matrix is:

$Q^{\&prime;}=\left(\begin{array}{ccccccc}{v}_{1,1}& v_{1,2}& \mathrm{...}& v_{1,k}& v_{1,k+1}& \mathrm{...}& v_{1.m}\\ v_{2,1}& v_{2,2}& \mathrm{...}& v_{2,k}& v_{2,k+1}& \mathrm{...}& v_{2,m}\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ v_{n,1}& v_{n,2}& \mathrm{...}& v_{n,k}& v_{n,k+1}& \mathrm{...}& v_{n,m}\end{array}\right)$

Wherein, v_ijRepresent the jth quality of service attribute of the i-th candidate's Web service after normalization.v_i,1v_i,2...v_i,k Representing the equalization quality of service attribute of the i-th candidate's Web service after normalization, quantity is k；v_i,k+1v_i,k+2...v_i,mTable Showing the unequalization quality of service attribute of the i-th candidate's Web service after normalization, quantity is m-k；

By equation below, 3rd computing unit 1733 is for determining that the equalization service quality of each candidate's Web service is commented It is worth:

Wherein, EqQoS (CS_i) represent i-th candidate's Web service equalization service quality evaluation value, v_i,jRepresent i-th The jth equalization QoS property value of individual service,Represent the weight shared by jth equalization quality of service attribute,And

4th computing unit 1734 for determining the unequalization Service Quality stating each candidate's Web service by equation below Amount evaluation of estimate:

$EnQoS\left({\mathrm{CS}}_i\right)=\underset{j=1}{\overset{m-k}{\&Sigma;}}{v}_{i,k+j}\&times;{\mathrm{\&psi;}}_{k+j}$

Wherein, EnQoS (CS_i) represent i-th candidate's Web service unequalization service quality evaluation value；v_i,k+jRepresent The jth unequalization QoS property value of i-th candidate's web services, ψ_k+jRepresent shared by jth unequalization quality of service attribute Weight, 0≤ψ_k+j≤ 1, and ψ_k+1+ψ_k+2+…+ψ_m=1.

5th computing unit 1735 for determining the Integrated Services Quality evaluation of each candidate's Web service by equation below Value:

FuQoS(CS_i)=a × EqQoS (CS_i)+b×EnQoS(CS_i)

Wherein, FuQoS (CS_i) represent i-th candidate's Web service Integrated Services Quality evaluation of estimate；EqQoS(CS_i) represent The equalization service quality evaluation value of i-th service；EnQoS(CS_i) represent the unequalization service quality evaluation that i-th services Value；A represents the weight that equalization service quality evaluation value is shared in a model, and b represents that unequalization service quality evaluation value exists Weight shared in model, and a+b=1, a ＞ 0, b ＞ 0.

It should be noted that the principle of the medical resource commending system of sing on web service is similar to above-mentioned recommendation method, phase Being referred to described above in place of pass, the present invention does not repeats them here.

The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all essences in the present invention Within god and principle, any modification, equivalent substitution and improvement etc. made, should be included within the scope of the present invention.

Claims (8)

1. the medical resource of a sing on web service recommends method, it is characterised in that comprise the steps:

Receiving step, receives the demand information that client sends；

Candidate's Web service set generation step, by described demand information and the preset medical services resource being encapsulated as Web service Mate, be determined for compliance with multiple candidate's Web services of customer demand, generate candidate's Web service set；

Service quality matrix generation step, determines the multiple source attribute values included by each candidate's Web service, builds described time Select the service quality matrix of Web service set；

Recommendation step, according to described service quality matrix and equalization QoS evaluating method, calculates each described candidate Web The Integrated Services Quality evaluation of estimate of service, and, give visitor by candidate's Web service recommendation maximum for Integrated Services Quality evaluation of estimate Family；

In described recommendation step, the Integrated Services Quality evaluation of estimate of each described candidate's Web service is determined as follows:

Calculate each equalization service quality attribute value of this candidate's Web service and the product of this qualitative attribute respective weights, and By all product addition, it is thus achieved that the equalization service quality evaluation value of this candidate's Web service；

Calculate each unequalization service quality attribute value of this candidate's Web service and the product of this attribute respective weights, and will All product addition, it is thus achieved that the unequalization service quality evaluation value of this candidate's Web service；

By described equalization service quality evaluation value and the first multiplication, obtain the first product, and, by described unequalization Service quality evaluation value and the second multiplication, obtain the second product；Described first coefficient is the equalization of this candidate's Web service Coefficient；Described second coefficient is the unequalization coefficient of this candidate's Web service；Calculate described first product and described second product Sum, as the Integrated Services Quality evaluation of estimate of described candidate's Web service；

Wherein, each equalization service quality attribute value of this candidate's Web service is corresponding weight, this candidate's Web service are each Corresponding weight, the first coefficient and the second coefficient of unequalization service quality attribute value is and presets；Each equalization takes The weight of business qualitative attribute value is arranged according to expertise, corresponding weight and be 1；Each unequalization service quality belongs to The weight of property value is arranged according to expertise, corresponding weight and be 1；First coefficient and the second coefficient and be 1；Wherein

Described equalization service quality attribute value directly reflects that the current operating conditions of medical services, each cycle can become Change, directly influence the equalization effect that follow-up service selects, including the response time serviced, current service ability, reliability And availability；

The fixed attribute of described unequalization service quality attribute value reflection medical services, does not reflect the current operation of medical services State, follow-up service will not be selected to produce impact by the frequency that its change occurs, including medical level, service price, maximum clothes Business ability and reputation degree.

The medical resource of sing on web the most according to claim 1 service recommends method, it is characterised in that

In described service quality matrix generation step, the service quality matrix of described candidate's Web service set is

$Q=\left(\begin{array}{ccccccc}{q}_{1,1}& q_{1,2}& \mathrm{...}& q_{1,k}& q_{1,k+1}& \mathrm{...}& q_{1.m}\\ q_{2,1}& q_{2,2}& \mathrm{...}& q_{2,k}& q_{2,k+1}& \mathrm{...}& q_{2,m}\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ q_{n,1}& q_{n,2}& \mathrm{...}& q_{n,k}& q_{n,k+1}& \mathrm{...}& q_{n,m}\end{array}\right)$

Wherein, q_ijRepresent the jth service quality attribute value of i-th candidate's Web service；q_i,1 q_i,2 ... q_i,kRepresent i-th The equalization service quality attribute value of candidate's Web service, quantity is k；q_i,k+1 q_i,k+2 ... q_i,mRepresent i-th candidate Web The unequalization service quality attribute value of service, quantity is m-k；

Further, described method also includes being normalized described service quality matrix Q, the Service Quality after normalized Moment matrix is

$Q^{\&prime;}=\left(\begin{array}{ccccccc}{v}_{1,1}& v_{1,2}& \mathrm{...}& v_{1,k}& v_{1,k+1}& \mathrm{...}& v_{1.m}\\ v_{2,1}& v_{2,2}& \mathrm{...}& v_{2,k}& v_{2,k+1}& \mathrm{...}& v_{2,m}\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ v_{n,1}& v_{n,2}& \mathrm{...}& v_{n,k}& v_{n,k+1}& \mathrm{...}& v_{n,m}\end{array}\right)$

Wherein, v_ijRepresent the jth service quality attribute value of the i-th candidate's Web service after normalization；v_i,1 v_i,2 ... v_i,kRepresenting the equalization service quality attribute value of the i-th candidate's Web service after normalization, quantity is k；v_i,k+1 v_i,k+2 ... v_i,mRepresenting the unequalization service quality attribute value of the i-th candidate's Web service after normalization, quantity is m-k.

The medical resource of sing on web the most according to claim 2 service recommends method, it is characterised in that

The equalization service quality evaluation value of described each candidate's Web service is determined by equation below:

Wherein, EqQoS (CS_i) represent i-th candidate's Web service equalization service quality evaluation value, v_i,jRepresent i-th service Jth equalization service quality attribute value,Represent the weight shared by jth equalization quality of service attribute, And

The unequalization service quality evaluation value of described each candidate's Web service is determined by equation below:

$EnQoS\left({\mathrm{CS}}_i\right)=\underset{j=1}{\overset{m-k}{\&Sigma;}}{v}_{i,k+j}\&times;{\mathrm{\&psi;}}_{k+j}$

Wherein, EnQoS (CS_i) represent i-th candidate's Web service unequalization service quality evaluation value；v_i,k+jRepresent i-th The jth unequalization service quality attribute value of candidate's web services, ψ_k+jRepresent shared by jth unequalization quality of service attribute Weight, 0≤ψ_k+j≤ 1, and ψ_k+1+ψ_k+2+…+ψ_m=1；

The service quality evaluation value of described each described candidate's Web service is determined by equation below:

FuQoS(CS_i)=a × EqQoS (CS_i)+b×EnQoS(CS_i)

Wherein, FuQoS (CS_i) represent i-th candidate's Web service Integrated Services Quality evaluation of estimate；EqQoS(CS_i) represent i-th The equalization service quality evaluation value of individual service；EnQoS(CS_i) represent the unequalization service quality evaluation value that i-th services； A represents the weight that equalization service quality evaluation value is shared in a model, and b represents that unequalization service quality evaluation value is at model Weight shared by, and a+b=1, a ＞ 0, b ＞ 0.

The medical resource of sing on web the most according to claim 3 service recommends method, it is characterised in that

By adjusting FuQoS (CS_iThe value of a and b in), can regulate the equalization degree in medical services resource selection:

If the equalization intensity of resource distribution need to be improved, then increase the value of a, reduce the value of b；

If the equalization intensity of resource distribution need to be reduced, then reduce the value of a, increase the value of b.

The medical resource of sing on web the most according to any one of claim 1 to 4 service recommends method, it is characterised in that

After described recommendation step, also include:

Parameter tuning step, after completing the recommendation to multiple customer demands, according to recommend evaluate, to described pre-set each Weight that each equalization service quality attribute value of candidate's Web service is corresponding and each unequalization service quality attribute value pair Weight, the first coefficient of each candidate's Web service and the second coefficient answered are adjusted.

The medical resource of sing on web the most according to claim 5 service recommends method, it is characterised in that

In described candidate's Web service set generation step, described candidate's Web service set obtains the most as follows:

According to client requirement information, in preset medical services resource, carry out service type, service name, service describing First coupling, obtains first service set；

Reading the candidate's Web service in described first service set one by one, the service quality updating each candidate's Web service belongs to Property；

In described first service set, screening meets candidate's Web service of described customer demand rigid constraint, forms the second clothes Business set, described second service collection is combined into candidate's Web service set.

7. the medical resource commending system of a sing on web service, it is characterised in that including:

Receiver module, for receiving the demand information that client sends；

Candidate's Web service set generation module, for by described demand information and the preset medical services being encapsulated as Web service Resource is mated, and is determined for compliance with multiple candidate's Web services of customer demand, generates candidate's Web service set；

Service quality matrix generation module, for determining the multiple source attribute values included by each candidate's Web service, builds institute State the service quality matrix of candidate's Web service set；

Recommending module, for according to described service quality matrix and equalization QoS evaluating method, calculates each described time Select the Integrated Services Quality evaluation of estimate of Web service, and, the candidate Web service recommendation maximum by Integrated Services Quality evaluation of estimate To client；

Described recommending module includes:

First computing unit, for calculating each equalization service quality attribute value and this qualitative attribute of this candidate's Web service The product of respective weights, and by all product addition, it is thus achieved that the equalization service quality evaluation value of this candidate's Web service；

Second computing unit, for calculating each unequalization service quality attribute value and this attribute pair of this candidate's Web service Answer the product of weight, and by all product addition, it is thus achieved that the unequalization service quality evaluation value of this candidate's Web service；

Service quality evaluation value acquiring unit, for by described equalization service quality evaluation value and the first multiplication, obtaining First product, and, by described unequalization service quality evaluation value and the second multiplication, obtain the second product；Described One coefficient is the equalization coefficient of this candidate's Web service；Described second coefficient is the unequalization coefficient of this candidate's Web service； Calculate described first product and described second sum of products, as the Integrated Services Quality evaluation of estimate of described candidate's Web service；

Wherein, each equalization service quality attribute value of this candidate's Web service is corresponding weight, this candidate's Web service are each Corresponding weight, the first coefficient and the second coefficient of unequalization service quality attribute value is and presets；Each equalization takes The weight of business qualitative attribute value is arranged according to expertise, corresponding weight and be 1；Each unequalization service quality belongs to The weight of property value is arranged according to expertise, corresponding weight and be 1；First coefficient and the second coefficient and be 1；Wherein

Described equalization service quality attribute value directly reflects that the current operating conditions of medical services, each cycle can become Change, directly influence the equalization effect that follow-up service selects, including the response time serviced, current service ability, reliability And availability；

The fixed attribute of described unequalization service quality attribute value reflection medical services, does not reflect the current operation of medical services State, follow-up service will not be selected to produce impact by the frequency that its change occurs, including medical level, service price, maximum clothes Business ability and reputation degree.

The medical resource commending system of sing on web the most according to claim 7 service, it is characterised in that

Described service quality matrix generation module farther includes: service quality matrix signal generating unit, normalized unit, Three computing units, the 4th computing unit and the 5th computing unit；Wherein

Service quality matrix signal generating unit is for generating the service quality matrix of candidate's Web service set

$Q=\left(\begin{array}{ccccccc}{q}_{1,1}& q_{1,2}& \mathrm{...}& q_{1,k}& q_{1,k+1}& \mathrm{...}& q_{1.m}\\ q_{2,1}& q_{2,2}& \mathrm{...}& q_{2,k}& q_{2,k+1}& \mathrm{...}& q_{2,m}\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ q_{n,1}& q_{n,2}& \mathrm{...}& q_{n,k}& q_{n,k+1}& \mathrm{...}& q_{n,m}\end{array}\right)$

Wherein, q_ijRepresent the jth quality of service attribute of i-th candidate's Web service；q_i,1 q_i,2 ... q_i,kRepresent that i-th is waited Selecting the equalization quality of service attribute of Web service, quantity is k；q_i,k+1 q_i,k+2 ... q_i,mRepresent i-th candidate's Web service Unequalization quality of service attribute, quantity is m-k；

Normalized unit is used for being normalized described service quality matrix Q, the Service Quality after normalized Moment matrix is:

$Q^{\&prime;}=\left(\begin{array}{ccccccc}{v}_{1,1}& v_{1,2}& \mathrm{...}& v_{1,k}& v_{1,k+1}& \mathrm{...}& v_{1.m}\\ v_{2,1}& v_{2,2}& \mathrm{...}& v_{2,k}& v_{2,k+1}& \mathrm{...}& v_{2,m}\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ .& .& .& .& .& .& .\\ v_{n,1}& v_{n,2}& \mathrm{...}& v_{n,k}& v_{n,k+1}& \mathrm{...}& v_{n,m}\end{array}\right)$

Wherein, v_ijRepresent the jth quality of service attribute of the i-th candidate's Web service after normalization；v_i,1 v_i,2 ... v_i,k Representing the equalization quality of service attribute of the i-th candidate's Web service after normalization, quantity is k；v_i,k+1 v_i,k+2 ... v_i,m Representing the unequalization quality of service attribute of the i-th candidate's Web service after normalization, quantity is m-k；

3rd computing unit for determining the equalization service quality evaluation of described each candidate's Web service by equation below Value:

Wherein, EqQoS (CS_i) represent i-th candidate's Web service equalization service quality evaluation value, v_i,jRepresent i-th service Jth equalization service quality attribute value,Represent the weight shared by jth equalization quality of service attribute, And

4th computing unit states the unequalization service quality evaluation of each candidate's Web service by equation below for being determined Value:

$EnQoS\left({\mathrm{CS}}_i\right)=\underset{j=1}{\overset{m-k}{\&Sigma;}}{v}_{i,k+j}\&times;{\mathrm{\&psi;}}_{k+j}$

Wherein, EnQoS (CS_i) represent i-th candidate's Web service unequalization service quality evaluation value；v_i,_k+jRepresent i-th The jth unequalization service quality attribute value of candidate's web services, ψ_k+jRepresent shared by jth unequalization quality of service attribute Weight, 0≤ψ_k+j≤ 1, and ψ_k+1+ψ_k+2+…+ψ_m=1；

5th computing unit for determining the Integrated Services Quality evaluation of described each described candidate's Web service by equation below Value:

FuQoS(CS_i)=a × EqQoS (CS_i)+b×EnQoS(CS_i)

Wherein, FuQoS (C_iS) the Integrated Services Quality evaluation of estimate of i-th candidate's Web service is represented；EqQoS(CS_i) represent i-th The equalization service quality evaluation value of individual service；EnQoS(CS_i) represent the unequalization service quality evaluation value that i-th services； A represents the weight that equalization service quality evaluation value is shared in a model, and b represents that unequalization service quality evaluation value is at model Weight shared by, and a+b=1, a ＞ 0, b ＞ 0.