CN106961356A - Web service choosing method and its device based on dynamic QoS and subjective and objective weight - Google Patents

Abstract

The present invention relates to a method and device for selecting Web services based on dynamic QoS and subjective and objective weights. The method includes: establishing an interval QoS model according to the fuzziness of user QoS requirements and the fluctuation range of candidate service QoS values; calculating each basic The similarity of candidate services under the service; combined with the user's subjective preference to calculate the comprehensive weight of QoS indicators; according to the comprehensive weight and similarity of QoS indicators, obtain the recommendation degree of each candidate service, and sort all candidate services according to the recommendation degree; according to the ranking results Select the service that meets the user's QoS requirements. The present invention has small calculation complexity and high feasibility, and combines the subjective and objective preference vectors to calculate the comprehensive weight of the QoS index, which avoids the one-sidedness of only using the subjective weighting mode or the objective weighting mode, and makes the determination of the weight more reasonable; the degree of recommendation is used as a measure The degree to which the candidate service meets the user's needs comprehensively considers the objective QoS data and the user's subjective preference information, which improves the accuracy of service selection.

Description

Web service selection method and device based on dynamic QoS and subjective and objective weights

technical field

The invention belongs to the field of computer application technology, and in particular relates to a method and device for selecting Web services based on dynamic QoS and subjective and objective weights.

Background technique

With the rapid development and commercial application of cloud computing technology, more and more data resources, computing resources and application resources rely on the Internet to provide applications in the form of Web services. In recent years, the number of Web services deployed on the Internet has exploded, and there are more and more Web services with the same functional attributes but different non-functional attributes. Therefore, the traditional method of selecting services based on functional attributes can no longer meet the needs of users. How to select services based on non-functional attributes of services, such as Quality of Service (QoS), has become a hot spot in the academic circle. A service selection method considering the uncertainty of QoS data was proposed by Shen Limin et al. According to the distribution of service QoS data, reasonable and accurate QoS constraints are customized; a QoS attribute cloud model is established to make uncertain decisions on the QoS data of Web services , to filter out the Web services whose QoS data performance fluctuates greatly, narrow down the search space for Web service selection on the basis of ensuring service reliability, and improve the efficiency of service selection; standardize the QoS data of the filtered service, and calculate the user's evaluation of the service The QoS aggregation value of the service; calculate the recommendation degree of the service according to the QoS evaluation and quantity of the service user, and then rank the services to complete the service recommendation. This method only uses the cloud model to filter services with QoS uncertainty, and assumes that the user has given a specific weight value of the QoS index, and the user's subjectivity is relatively large.

The existing QoS-based service selection methods have the following disadvantages: (1) The uncertainty of QoS attribute values of Web services is not considered. This uncertainty manifests itself in two ways. One aspect is that Web services run in a dynamic network environment, and any factors such as network bandwidth, location, time, etc. will affect the QoS attribute values of Web services, so the accuracy of service selection based on precise QoS attribute values is not high. Another aspect is that due to the ambiguity of human thinking patterns, users are more inclined to use interval numbers rather than precise values to express their needs for services. Considering the uncertainty of QoS should be taken as the premise of service selection and service combination, and most existing models have not considered the uncertainty of QoS data. (2) The determination of service QoS attribute weight is unreasonable. At present, there are mainly two methods of determining the weight, subjective weighting mode and objective weighting mode. In the subjective weighting mode, the QoS weight of the service is completely determined by the user. Although this reflects the individual needs of the user, it has great subjective randomness, and most users do not have relevant knowledge in the service field. The weight value places an additional burden on the user. Relatively speaking, users are more inclined to use preference order to express their needs, such as service price>response time>reliability, etc. The objective empowerment model is determined by objective data. Although the result has a strong numerical theoretical basis, it ignores the user's subjective preference. Therefore, it is more in line with the actual situation of service selection to consider both the user's subjective preference and objective data when determining the QoS weight.

Contents of the invention

Aiming at the deficiencies in the prior art, the present invention provides a method and device for selecting Web services based on dynamic QoS and subjective and objective weights, so as to improve the accuracy of service selection.

According to the design scheme provided by the present invention, a method for selecting web services based on dynamic QoS and subjective and objective weights includes the following content:

According to the ambiguity of user QoS requirements and the fluctuation range of candidate service QoS values, an interval QoS model is established;

Calculate the similarity of candidate services under each basic service;

Calculate the comprehensive weight of QoS indicators in combination with the user's subjective preference;

According to the comprehensive weight and similarity of QoS indicators, the recommendation degree of each candidate service is obtained, and all candidate services are sorted by the recommendation degree;

Select the service that meets the user's QoS requirements according to the sorting results.

As mentioned above, according to the influence of QoS attributes on the quality of Web services, the QoS attributes of Web services are defined as a four-dimensional vector Qos=(T, E, A, R), and intervals are used to represent the fluctuation range of QoS values, where T is Response time; E is service reputation; A is service availability; R is service reliability.

As mentioned above, the calculation of the similarity of candidate services under each basic service includes the following content: Assume that the combined service WSC is composed of m basic services WS _i , recorded as WSC={WS ₁ ,WS ₂ ,...WS _m }, each A basic service WS _i has n candidate services, denoted as WS _i ={s ₁ , s ₂ ,…s _n }, and k QoS indicators are denoted as set QCS; using the interval similarity formula, each candidate service can be obtained The similarity of QoS attributes; the n candidate services under each basic service are listed as an n×k similarity matrix V _i to obtain m matrices,

, where p _ij represents the similarity of the QoS index j of the i-th candidate service of the basic service.

As mentioned above, using the multi-attribute decision-making method approaching the ideal point, the objective weight of the basic service QoS index is solved by establishing a goal programming model.

Preferably, solving the objective weight of the QoS index of the basic service includes the following content: the weight vector of the QoS index of the basic service is W=<w ₁ ,w ₂ ,...w _k >, p _ij represents the ith candidate of the basic service The similarity of the QoS index j of the service; let the positive ideal point be 1, then the weighted distance between the candidate service i and the positive ideal point is: Establish a multi-objective decision-making model: solve a weight vector W, so that Transform the multi-objective optimization problem into single-objective optimization, namely: Calculate the objective weight vector W _l of the basic service l with respect to the QoS index by solving.

As mentioned above, the comprehensive weight of the QoS index is calculated in combination with the user's subjective preference, including the following content: combined with the user's subjective preference, the importance weight of the basic service is obtained by using the method based on the offset similarity measurement; according to the objective weight and importance weight of the QoS index, Calculate the comprehensive weight of the QoS indicator.

Preferably, the calculation of the comprehensive weight of the QoS index includes the following content: the objective weight of the QoS index W _l =<w _l1 ,w _l2 ,...w _lk >, 1≤l≤m, which is listed as the objective weight matrix W, namely :

, according to the importance weight vector U=<u ₁ ,u ₂ ,…u _m > of this series of basic services, the comprehensive QoS index weight Ω=<Ω ₁ ,Ω ₂ ,…Ω _k > considering user preference is calculated by formula .

Furthermore, the comprehensive QoS index weight Ω=<Ω ₁ ,Ω ₂ ,...Ω _k > considering user preference is obtained through formula calculation, and the calculation formula is:

As mentioned above, all candidate services are sorted by recommendation degree, which specifically includes the following content: the recommendation degree of candidate services for each basic service is arranged in order from large to small, and stored in the form of a service link list, and by setting the recommendation degree Threshold to filter.

A Web service selection device based on dynamic QoS and subjective and objective weights, comprising: a QoS model building module, a similarity acquisition module, a comprehensive weight calculation module, a service sorting module and a service selection module,

The QoS model building module is used to build an interval QoS model according to the ambiguity of user QoS requirements and the fluctuation range of candidate service QoS values;

A similarity acquisition module is used to calculate the similarity of each QoS attribute of each candidate service by adopting an interval similarity formula according to the interval QoS model;

The comprehensive weight calculation module is used to use the multi-attribute decision-making method approaching the ideal point to solve the objective weight of the basic service QoS index through the establishment of a multi-objective programming model; combined with the user's subjective preference, the method based on the offset similarity measurement is used to obtain the importance of the basic service. Calculate the comprehensive weight of QoS indicators according to the objective weight and importance weight of QoS indicators;

The service sorting module is used to obtain the recommendation degree of each candidate service according to the QoS index comprehensive weight obtained by the comprehensive weight calculation module and the similarity degree obtained by the similarity degree acquisition module, and sort all candidate services through the recommendation degree;

The service selection module selects services that meet the user's QoS requirements according to the sorting results of the service sorting module.

Beneficial effects of the present invention:

Aiming at the problems existing in the existing QoS-based service selection method, the present invention expresses the fluctuating range of candidate service QoS values and the ambiguity of user QoS requirements by designing an interval QoS model, and expresses the range of service QoS values in a dynamic network environment through intervals The fluctuating range and the ambiguity of the user's QoS requirements expression; the interval similarity calculation method is used to measure the closeness of the QoS attributes provided by the candidate service to the user's request, and based on the interval similarity, the multi-attribute decision-making method (TOPSIS ), by establishing a multi-objective programming model to solve the objective weight of the QoS index of the basic service, the calculation complexity is small, and the feasibility is high. And the user preference vector is used to represent the user's preference for the QoS index, which does not require the user to give an accurate weight value, which reduces the burden on the user. Combined with the objective preference vector, the comprehensive weight of the QoS index is calculated, which avoids the one-sidedness of only using the subjective or objective weighting model, and makes the determination of the weight more reasonable. The degree of recommendation is used to measure the extent to which candidate services meet user needs, and the objective QoS data and user subjective preference information are considered comprehensively, which improves the accuracy of service selection.

Description of drawings:

Fig. 1 is a schematic flow chart of the method of the present invention;

FIG. 2 is a schematic diagram of a service recommendation linked list in an embodiment of the present invention;

Fig. 3 is a schematic diagram of the device of the present invention.

detailed description:

In order to make the purpose, technical solution and advantages of the present invention more clear and understandable, the present invention will be further described in detail below in conjunction with the accompanying drawings and technical solutions. It should be noted that, in the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined arbitrarily with each other.

In the present invention: Quality of Service (QoS): represents the non-functional attributes of Web services, including response time, reputation, availability, reliability, etc., and is another important criterion for evaluating service quality. Response time: refers to the time it takes from the user sending a service request to obtaining the service, which is affected by the network operation status, so it is dynamic. Credibility: Refers to the credibility of the service. The credibility of the service is mainly determined by the user's evaluation after use. Due to the differences and dynamics of the service environment, different users may have different evaluations when using the same service. Availability: The probability that the Web service is called normally. The availability of services also changes dynamically with good or bad network conditions. Reliability: The probability that a Web service can correctly respond to user requests. In addition to the service quality of the service itself, the network environment will also affect the reliability of the service. Subjective weighting mode: It means that the QoS weight of the service is completely determined subjectively by the user. Objective weighting mode: means that the QoS weight of the service is determined by objective data. User preference vector: refers to the vector obtained by sorting the QoS indicators from large to small according to the degree of user preference, representing the user's preference for different QoS indicators. Recommendation degree: the comprehensive evaluation value of candidate services obtained after comprehensive consideration of objective data and user subjective preferences.

Embodiment, referring to shown in Fig. 1, a kind of Web service selection method based on dynamic QoS and subjective and objective weight, comprises the following content:

101. Establish an interval QoS model according to the ambiguity of user QoS requirements and the fluctuation range of candidate service QoS values;

102. Calculate the similarity of candidate services under each basic service;

103. Calculate the comprehensive weight of the QoS index in combination with the user's subjective preference;

104. According to the comprehensive weight and similarity of the QoS indicators, the recommendation degree of each candidate service is obtained, and all candidate services are sorted by the recommendation degree;

105. Select a service meeting the user's QoS requirement according to the sorting result.

Firstly, the QoS uncertainty of web service in dynamic environment is analyzed, and the dynamic change of QoS attribute value of web service is represented by interval QoS model. Due to the ambiguity of the user's thinking mode, the user is more inclined to use an approximate range rather than a precise value to express their service requirements. Therefore, this embodiment also uses an interval QoS model to express the user's QoS requirements. The concept of similarity is used to measure the closeness of the candidate service QoS attribute interval to the user demand value interval. The greater the similarity, the more the candidate service meets the user's requirements on this QoS attribute, and the higher the probability of being selected. After the similarity of each QoS attribute of each candidate service is obtained, combined with the corresponding weight, the recommendation value of each candidate service can be obtained, so as to sort and select the candidate services, and consider the user's subjective preference and objective The actual situation of the data; to solve the defects existing in the traditional QoS-based service selection method, including establishing an interval QoS model to represent the dynamic change of the QoS attribute value, to overcome the influence of QoS uncertainty; to use the multi-attribute decision-making method approaching the ideal point to solve The objective weight of the QoS index is combined with the subjective preference of the user to determine the comprehensive weight, which makes the determination of the QoS weight more reasonable.

The method can also include the following features:

Further, according to the influence of QoS attributes on the quality of Web services, the QoS attributes of Web services are defined as a four-dimensional vector Qos=(T, E, A, R), and intervals are used to represent the fluctuation range of QoS values, where T is Response time; E is service reputation; A is service availability; R is service reliability.

The QoS attributes of Web services have many aspects, such as response time, service cost, throughput, availability, security, etc., and they evaluate the quality of Web services from different angles. The invention defines the QoS attribute of the Web service as a four-dimensional vector Qos=(T, E, A, R), and uses intervals to represent the fluctuation range of its values. Response time T: the time it takes for the user to send a service request to obtain the service. The response time is affected by the network operation status and has dynamic variability; it is described as an interval of where _μT is the mean value, is the variance. Credibility E: The credibility of the service. The credibility of the service is mainly determined by the evaluation of the user after use; due to the differences and dynamics of the service environment, different users may have different evaluations when using the same service; its range in the form of Availability A: The probability that the Web service is called normally; the availability of the service will also change dynamically with the quality of the network environment; its interval form is Reliability R: the probability that a web service can correctly respond to user requests; in addition to the service quality of the service itself, the network environment will also affect the reliability of the service, such as network congestion causing service information to be lost or invalid during transmission; its interval form for

μ _M in the QoS interval, The calculation process of M={T,E,A,R} is as follows:

The relevant definitions of interval numbers are as follows:

Definition 1: Record Q=[x ^- ,x ⁺ ]=[x|x ^- ≤x≤x ⁺ ], where x ^- ,x ⁺ ∈R, R is a set of real numbers, and Q is called an interval number, if x ^- ＝x ⁺ , then Q degenerates into a real number. In this paper, interval numbers refer to positive interval numbers, that is, 0≤x ^- ≤x ⁺ .

Definition 2: If there are two interval numbers Q _x , Q _y , where Q _x ＝[x ^- ,x ⁺ ], Q _y ＝[y ^- ,y ⁺ ], record l _x ＝x ⁺ -x ^- ,l _y ＝y ⁺ -y ^- , says:

is the similarity of Q _x ≥ Q _y .

According to the definition, it can be proved that p(Q _x ≥ Q _y ) has the following properties:

(1) 0≤p(Q _x ≥Q _y )≤1, p(Q _x ≥Q _y )=1 if and only if y ^{+ ≤x - ,} p(Q _x ≥ Q _y )=0.

(2) Complementarity: p(Q _x ≥ Q _y )+p(Q _y ≥ Q _x )=1.

(3) If and only if x ⁺ +x ^- ≥ y ⁺ +y ^- , there is In particular, if and only if x ⁺ +x ^- ＝y ⁺ +y ^- , we have

(4) Transitivity: if and then there is

Let the user's demand for a certain QoS attribute be the interval number The QoS attribute interval number of a candidate service is p means similarity. For benefit-type QoS indicators, the larger the interval value, the better; for cost-type QoS indicators, the smaller the interval value, the better. but

Based on the above properties, it can be seen that the similarity is used to measure the closeness of the QoS attributes provided by the candidate service to the user's request. The greater the similarity, the more the candidate service meets the user's requirements on this QoS attribute, and the higher the probability of being selected.

As mentioned above, the calculation of the similarity of candidate services under each basic service includes the following content: Assume that the combined service WSC is composed of m basic services WS _i , recorded as WSC={WS ₁ ,WS ₂ ,...WS _m }, each A basic service WS _i has n candidate services, denoted as WS _i ={s ₁ , s ₂ ,…s _n }, and k QoS indicators are denoted as set QCS; using the interval similarity formula, each candidate service can be obtained The similarity of QoS attributes; the n candidate services under each basic service are listed as an n×k similarity matrix V _i to obtain m matrices,

Among them, p _ij represents the similarity of the QoS index j of the ith candidate service of the basic service.

As mentioned above, by using the multi-attribute decision-making method approaching the ideal point, the objective weight of the basic service QoS index is solved by establishing a multi-objective programming model.

Preferably, solving the objective weight of the QoS index of the basic service includes the following content: the weight vector of the QoS index of the basic service is W=<w ₁ ,w ₂ ,...w _k >, p _ij represents the ith candidate of the basic service The similarity of the QoS index j of the service; let the positive ideal point be 1, then the weighted distance between the candidate service i and the positive ideal point is: The closer the candidate service is to the positive ideal point, the better, that is The smaller the value, the better the candidate service. Establish a multi-objective decision-making model: solve a weight vector W, so that because And there is no distinction between each candidate service, so the multi-objective optimization problem is transformed into single-objective optimization, namely: Calculate the objective weight vector W _l of the basic service l with respect to the QoS index by solving.

As mentioned above, the comprehensive weight of the QoS index is calculated in combination with the user's subjective preference, including the following content: combined with the user's subjective preference, the importance weight of the basic service is obtained by using the method based on the offset similarity measurement; according to the objective weight and importance weight of the QoS index, Calculate the comprehensive weight of the QoS indicator.

Suppose the user's preference vector is pref, and the elements in the vector are the QoS indicators of Web services. The user's preference for the QoS indicators decreases from front to back according to the order of the elements in the vector. For example, pref=<T, E, A, R> indicates that the priority order of the user's QoS index preference is response time, reputation, availability, and reliability.

Objective preference vector: refers to the QoS index preference of basic services. Knowing the objective weight vector W _l of the QoS index of the basic service l =<w _l1 ,w _l2 ,...w _lk >, the corresponding objective preference vector is The weight value of each QoS index satisfies That is, it is obtained by reordering the QoS indicators from large to small according to the calculated weights.

It is known that there are m basic services in the composite service, then according to the previous section, the QoS objective weight vector W _l of the m basic services, 1≤l≤m, and the corresponding objective preference vector pref _l , 1≤l≤ m. The user's subjective preference vector is pref. It can be known from experience that if the objective preference vector of a basic service is close to or even equal to the user's subjective preference vector, it means that the basic service is more in line with the user's psychological expectations, and it should occupy a larger proportion in the determination of the comprehensive QoS weight. The present invention refers to this ratio as the importance weight of the basic service. Then the importance weights of m basic services can form a vector, which is recorded as vector U=<u ₁ , u ₂ ,... u _m >. By using an offset-based similarity measurement method, the similarity is described by the offset degree of the QoS index elements in the objective preference vector relative to the corresponding elements in the subjective preference vector. Due to the ordered nature of the preference vector, the offsets at different positions account for decreasing proportions in the similarity measure. For example, assuming that the user’s subjective preference vector is pref=<response time T, reputation E, availability A, reliability R>, the objective preference vector is

pref _l = <availability A, response time T, reliability R, reputation E>, it can be seen that relative to the objective preference vector, the response time T shifts backward by 1 digit, and the reputation E shifts backward by 2 digits, while Since the user attaches more importance to the response time T than the reputation E, the offset degree of the response time T is more important than the reputation E when calculating the similarity measure.

Preferably, the calculation of the comprehensive weight of the QoS index includes the following content: the objective weight of the QoS index W _l =<w _l1 ,w _l2 ,...w _lk >, 1≤l≤m, which is listed as the objective weight matrix W, namely :

, according to the importance weight vector U=<u ₁ ,u ₂ ,…u _m > of this series of basic services, the comprehensive QoS index weight Ω=<Ω ₁ ,Ω ₂ ,…Ω _k > considering user preference is calculated by formula .

Furthermore, the comprehensive QoS index weight Ω=<Ω ₁ ,Ω ₂ ,...Ω _k > considering user preference is obtained through formula calculation, and the calculation formula is:

As mentioned above, all candidate services are sorted by recommendation degree, which specifically includes the following content: the recommendation degree of candidate services for each basic service is arranged in order from large to small, and stored in the form of a service link list, and by setting the recommendation degree Threshold to filter.

The similarity matrix V _i of each basic service has been calculated, 1≤i≤m, multiplied by the weight of the comprehensive QoS index, and the recommendation matrix X _i corresponding to each basic service is obtained, namely:

For composite service WSC={WS ₁ ,WS ₂ ,…WS _m }, each basic service WS _i (1≤i≤m) has n candidate services s _i (1≤i≤n), each The recommendation degree of the candidate service is x _i , which are arranged in descending order and stored in the form of a linked list. As shown in Figure 2, when using the service recommendation linked list for service selection, a recommendation degree threshold can be set, and candidate services with low recommendation degrees can be eliminated according to this threshold, thereby reducing the candidate service space and further improving the accuracy of service selection.

The embodiment of the present invention also provides a Web service selection device based on dynamic QoS and subjective and objective weights, as shown in FIG. 204 and service selection module 205,

The QoS model building module 201 is used to build an interval QoS model according to the ambiguity of user QoS requirements and the fluctuation range of candidate service QoS values;

The similarity acquisition module 202 is used to calculate the similarity of each QoS attribute of each candidate service by using the interval similarity formula according to the interval QoS model;

The comprehensive weight calculation module 203 is used to solve the objective weight of the basic service QoS index by using the multi-attribute decision-making method approaching the ideal point by establishing a multi-objective programming model; and combining the user's subjective preference with a method based on offset similarity measurement to obtain the basic service Importance weight: Calculate the comprehensive weight of QoS indicators according to the objective weight and importance weight of QoS indicators;

The service sorting module 204 is used to obtain the recommendation degree of each candidate service according to the QoS index comprehensive weight obtained by the comprehensive weight calculation module and the similarity degree obtained by the similarity degree acquisition module, and sort all candidate services through the recommendation degree;

The service selection module 205 selects services that meet the user's QoS requirements according to the sorting results of the service sorting module.

The interval QoS model is used to represent the fluctuation range of the QoS value of candidate services and the QoS requirements of users, which overcomes the uncertain influence of QoS in the dynamic network environment and the ambiguity of user demand expression; it uses the multi-attribute decision-making method approaching the ideal point to solve the problem The objective weight of the QoS index is calculated, the time complexity is small, and the calculation efficiency is high; the user preference vector is defined to represent the user's subjective preference, and the user only needs to give the preference ranking of the QoS index, which reduces the burden on the user and is more suitable for Actual situation: By comprehensively using QoS objective data and user subjective preference information, the recommendation degree is defined to measure the pros and cons of candidate services, which improves the accuracy of service selection.

Those of ordinary skill in the art can understand that all or part of the steps in the above method can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium, such as: a read-only memory, a magnetic disk or an optical disk, and the like. Optionally, all or part of the steps in the above embodiments can also be implemented using one or more integrated circuits. Correspondingly, each module/unit in the above embodiments can be implemented in the form of hardware, or can be implemented in the form of software function modules. The form is realized. The present invention is not limited to any specific combination of hardware and software.

The above description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the application. Therefore, the present application will not be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (10)

1. a kind of Web service choosing method based on dynamic QoS and subjective and objective weight, it is characterised in that include following content：

According to the ambiguity of user's QoS demand and the fluctuation range of candidate service qos value, interval QoS model is set up；

Calculate the similarity of candidate service under each basic service；

QoS index comprehensive weight is calculated with reference to user's subjective preferences；

According to QoS index comprehensive weight and similarity, the recommendation degree of each candidate service is obtained, by degree of recommendation to all candidates Service is ranked up；

The service for meeting user's QoS demand is chosen according to ranking results.

2. the Web service choosing method according to claim 1 based on dynamic QoS and subjective and objective weight, it is characterised in that Influence according to QoS attributes to Web service quality, by the QoS attribute definitions of Web service be four dimensional vector Qos=(T, E, A, R the fluctuation range of qos value), and using interval is represented, wherein, T is the response time；E is credit degree of service；A is available for service Property；R is service reliability.

3. the Web service choosing method according to claim 1 based on dynamic QoS and subjective and objective weight, it is characterised in that The similarity of the candidate service under each basic service is calculated, following content is included：Assuming that composite services WSC is by m basic clothes Be engaged in WS_iConstitute, be designated as WSC={ WS₁,WS₂,…WS_m, each basic service WS_iThere is n candidate service, be designated as WS_i={ s₁, s₂,…s_n, k QoS index is designated as set QCS；Using similarity between intervals formula, each QoS category of each candidate service is obtained The similarity of property；N candidate service under each basic service is classified as to n × k similarity matrix V_i, obtain m square Battle array,

$V_i=\left[\begin{array}{cccc}{p}_{11}& p_{12}& ...& p_{1k}\\ p_{21}& p_{22}& ...& p_{2k}\\ \begin{array}{c}.\\ .\\ .\end{array}& \begin{array}{c}.\\ .\\ .\end{array}& \begin{array}{c}.\\ .\\ .\end{array}& \begin{array}{c}.\\ .\\ .\end{array}\\ p_{n1}& p_{n2}& ...& p_{nk}\end{array}\right],(1\&le;i\&le;m)$ ,

Wherein, p_ijRepresent the QoS index j of i-th of candidate service of basic service similarity.

4. the Web service choosing method according to claim 1 based on dynamic QoS and subjective and objective weight, it is characterised in that Using the multiple attributive decision making method for approaching ideal point, by setting up Multiobjective programming models, basic service QoS index is solved objective Weight.

5. the Web service choosing method according to claim 4 based on dynamic QoS and subjective and objective weight, it is characterised in that Basic service QoS index objective weight is solved, following content is included：The QoS index weight vectors of basic service are W=<w₁, w₂,...w_k>, p_ijRepresent the QoS index j of i-th of candidate service of basic service similarity；It is 1 to make Positive ideal point, then Weighted distance between candidate service i and Positive ideal point is：Set up multiobjective decision-making mould Type：Solve a weight vectors W so thatMulti-objective optimization question is converted into single goal excellent Change, i.e.,：The objective weight vector for calculating basic service l on QoS index by solving W_l。

6. the Web service choosing method according to claim 1 based on dynamic QoS and subjective and objective weight, it is characterised in that QoS index comprehensive weight is calculated with reference to user's subjective preferences, following content is included：With reference to user's subjective preferences, using based on inclined The method of shifting amount measuring similarity obtains the weights of importance of basic service；According to QoS index objective weight and weights of importance, Calculate QoS index comprehensive weight.

7. the Web service choosing method according to claim 6 based on dynamic QoS and subjective and objective weight, it is characterised in that QoS index comprehensive weight is calculated, following content is included：QoS index objective weight W_l=<w_l1,w_l2,...w_lk>, 1≤l≤m will It is classified as objective weight matrix W, i.e.,：

$W=\left[\begin{array}{c}{W}_1\\ W_2\\ \begin{array}{c}.\\ .\\ .\end{array}\\ W_m\end{array}\right]=\left[\begin{array}{cccc}{w}_{11}& w_{12}& ...& w_{1k}\\ w_{21}& w_{22}& ...& w_{2k}\\ ...& ...& ...& ...\\ w_{m1}& w_{m2}& ...& w_{mk}\end{array}\right]$

According to the weights of importance of serial basic service vector U=<u₁,u₂,…u_m>, calculated by formula and obtain considering user The comprehensive QoS index weight Ω of preference=<Ω₁,Ω₂,…Ω_k>。

8. the Web service choosing method according to claim 7 based on dynamic QoS and subjective and objective weight, it is characterised in that By formula calculate obtain consider user preference comprehensive QoS index weight Ω=<Ω₁,Ω₂,…Ω_k>, computing formula is：

$\mathrm{\&Omega;}=U*W=<u_1,u_2,...u_m>*\left[\begin{array}{c}{W}_1\\ W_2\\ .\\ .\\ .\\ W_m\end{array}\right]=u_1{W}_1+u_2{W}_2+...u_m{W}_m=<{\mathrm{\&Omega;}}_1,{\mathrm{\&Omega;}}_2,...{\mathrm{\&Omega;}}_k>.$

9. the Web service choosing method according to claim 1 based on dynamic QoS and subjective and objective weight, it is characterised in that All candidate services are ranked up by degree of recommendation, specifically comprising following content：The candidate service of each basic service is pushed away Degree of recommending is arranged according to descending order, and is stored with service chaining sheet form, and is carried out by setting recommendation degree threshold value Screening.

10. a kind of Web service selecting device based on dynamic QoS and subjective and objective weight, it is characterised in that include：QoS model is built Module is chosen in formwork erection block, similarity acquisition module, comprehensive weight computing module, service ranking module and service,

QoS model sets up module, is set up for the ambiguity according to user's QoS demand and the fluctuation range of candidate service qos value Interval QoS model；

Similarity acquisition module, for calculating each of each candidate service using similarity between intervals formula according to interval QoS model The similarity of individual QoS attributes；

Comprehensive weight computing module, the multiple attributive decision making method of ideal point is approached for utilizing by setting up Multiobjective programming models Solve basic service QoS index objective weight；And user's subjective preferences are combined using the method based on offset measuring similarity Obtain the weights of importance of basic service；According to QoS index objective weight and weights of importance, QoS index comprehensive weight is calculated；

Service ranking module, QoS index comprehensive weight and similarity for being obtained according to comprehensive weight computing module obtain mould The similarity that block is obtained, obtains the recommendation degree of each candidate service, and all candidate services are ranked up by degree of recommendation；

Module is chosen in service, and the service for meeting user's QoS demand is chosen according to the ranking results of service ranking module.