CN112511346A - Web service combination method based on credibility screening - Google Patents

Web service combination method based on credibility screening Download PDF

Info

Publication number
CN112511346A
CN112511346A CN202011319892.3A CN202011319892A CN112511346A CN 112511346 A CN112511346 A CN 112511346A CN 202011319892 A CN202011319892 A CN 202011319892A CN 112511346 A CN112511346 A CN 112511346A
Authority
CN
China
Prior art keywords
qos
particles
particle
web service
population
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202011319892.3A
Other languages
Chinese (zh)
Inventor
韩敏
刘锋
钟凯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Dalian University of Technology
Original Assignee
Dalian University of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Dalian University of Technology filed Critical Dalian University of Technology
Priority to CN202011319892.3A priority Critical patent/CN112511346A/en
Publication of CN112511346A publication Critical patent/CN112511346A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/02—Standardisation; Integration
    • H04L41/0246—Exchanging or transporting network management information using the Internet; Embedding network management web servers in network elements; Web-services-based protocols
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00—Computing arrangements based on specific mathematical models
    • G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0803—Configuration setting
    • H04L41/0823—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/50—Network service management, e.g. ensuring proper service fulfilment according to agreements

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Algebra (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Analysis (AREA)
  • Computational Mathematics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Health & Medical Sciences (AREA)
  • Pure & Applied Mathematics (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

A Web service combination method based on credibility screening belongs to the field of computer network services. Firstly, according to the historical running record of the Web service, the QoS credibility of the Web service is calculated by adopting a Bayesian learning theory and time factor method, and the Web service is screened by using a skyline theory according to the credibility, so that the credible Web service is obtained and used as a candidate service set in a service combination process. Then, a multi-objective optimization model is constructed for the QoS attributes. And finally, solving the optimization model by adopting an improved multi-target particle swarm algorithm to obtain a credible combination scheme meeting the user requirements. The calculation of the QoS attribute credibility value is real-time and reliable; screening out the dominated candidate services by using a skyline theoretical method, so that the requirements of users are better met, and the obtained candidate service set is credible; and an improved multi-target particle swarm algorithm is adopted to solve the Web service combination scheme, so that a better Web service combination scheme is solved while the solving speed is ensured.

Description

Web service combination method based on credibility screening
Technical Field
The invention belongs to the field of computer network services, and particularly relates to a Web service combination method based on credibility screening.
Background
Web services are open, loosely coupled, highly integrated, comprehensive computer software that has broad applications in social production and life. However, a single Web service has limited functions, and cannot meet diversified demands of users, and meanwhile, in order to utilize existing resources to the maximum extent, a Web service combination technology is gradually gaining attention. The Web service combination technology is a new Web service with value added function combined by different Web services through workflow, and the Web service combination technology based on workflow mainly selects Web services with proper functions (called subtasks in the service combination process) to combine on the basis of the designed workflow to obtain an optimal combination scheme.
With the increase of the number of Web services in a network, a large number of Web services with the same or similar functional attributes but different non-functional attributes appear, so that people select a Web Service, not only considering the functional attributes, but also paying more and more attention to the non-functional attributes of the Web Service, and Quality of Service (QoS) in the non-functional attributes is one of the most concerned contents, thereby deriving a QoS-based Web Service combination technology. The technology mainly selects proper Web services to be combined according to the attribute values of the QoS, so that the combined Web services meet the QoS requirements of users. However, sometimes the QoS attribute values of Web services that people obtain are not necessarily reliable because: on one hand, the network environment is dynamically changed, and the QoS attribute value of the Web service may be changed; on the other hand, a publisher of a Web service may publish QoS attribute values that do not match the actual values. This situation can result in the final combined Web service having QoS attributes that are not as desirable or even that do not meet the user's requirements. Therefore, it is of great significance to study the reliability of QoS.
The invention provides a method for calculating a credible QoS index, which is found by searching documents in the prior art, and is named as a credible Web service combination optimization method based on QoS (quality of service) (with the publication number of CN107070704A and the publication number of 2017.08.18). However, the disadvantages of this method are: on the one hand, a Web service with low reliability may cause that the combined Web service cannot reach the expected QoS attribute value, and on the other hand, the combined Web service is inconsistent with the behavior that the user prefers to select a Web service with high reliability. Therefore, the invention provides a Web service combination method based on credibility screening, which adopts a Bayesian learning theory and time factors to calculate the credibility of QoS, and deletes the Web service with low credibility according to a skyline theory to obtain a credible Web service as a candidate service set of a subtask in the combination process. Compared with a method combining reliability and QoS, the method provided by the invention is more reliable for constructing reliable combined service and can meet the requirements of users more easily.
Disclosure of Invention
In order to solve the problems, the invention provides a Web service combination method based on credibility screening, which is used for solving a credible Web service combination scheme, the credibility of the Web service combination scheme is calculated by utilizing a Bayesian learning theory and a time factor, and credible Web services are screened out by utilizing a skyline theory according to the calculated credibility and serve as a candidate service set of subtasks in the Web service combination process. Meanwhile, in order to solve the Web service combination scheme better, an improved multi-target particle swarm algorithm is adopted for solving, and the method can solve the Web service combination scheme with better quality while maintaining the solving efficiency.
In order to achieve the purpose, the technical scheme of the invention is as follows:
(1) aiming at a certain QoS attribute of Web service, calculating the credibility value of the attribute by adopting a Bayesian learning theory and a time factor method;
(2) calculating the credibility value of each QoS attribute of each candidate service in the candidate service set, and deleting the dominated candidate service by adopting a skyline theory according to the credibility value to obtain a credible candidate service set as a solving space for subsequent optimization;
(3) and constructing a multi-target model according to each QoS attribute, optimizing the multi-target model by adopting an improved multi-target particle swarm optimization, and solving a Web service combination scheme.
Further, the step of calculating the credibility value of a certain QoS attribute of the Web service in step (1) includes the following steps:
1) and calculating the trust R according to the release value q and the historical operation value record q' of a certain QoS attribute of the Web service, wherein the calculation formula is as follows:
Figure BDA0002792546830000021
2) selecting the historical running record of the attribute in a certain period of time, and obtaining the trust level set of the attribute through the formula (1)
Figure BDA0002792546830000022
Wherein
Figure BDA0002792546830000023
Representing the degree of trust of the i-th run, diRepresenting the number of days from the current time at the i-th run;
3) assuming degree of distrust
Figure BDA0002792546830000024
And calculating the total trust degree and the total distrust degree of the attribute in the period of time according to the following calculation formula:
Figure BDA0002792546830000025
Figure BDA0002792546830000026
wherein, lambda is a time factor, lambda is more than 0 and less than or equal to 1, and represents the importance degree of time to the trust; rλ,qRepresenting the overall trust of the attribute q; i denotes the ith run, i ∈ [1,2, …, n]；
Figure BDA0002792546830000027
Indicating the overall degree of distrust for the attribute;
Figure BDA0002792546830000028
representing the degree of distrust of the i-th run, diIndicating the number of days from the current time at the i-th run.
4) According to Bayes learning theory, calculating reliability value rep of the attributeqThe calculation formula is as follows:
Figure BDA0002792546830000031
further, the screening the trusted candidate service set in step (2) includes the following steps:
1) according to the reliability value calculation method in the step (1), the reliability value of each QoS attribute of each candidate service in the subtask candidate service set is calculated to obtain the QoS attribute reliability set of the candidate service
Figure BDA0002792546830000032
Wherein the content of the first and second substances,
Figure BDA0002792546830000033
representing candidate services wsiThe set of QoS trustworthiness of (a),
Figure BDA0002792546830000034
representing candidate services wsiQ (a) to (b)nConfidence value of each QoS.
2) According to the credibility set, sequentially comparing the candidate service with other candidate services by using a skyline theory, and deleting the dominated candidate service from the candidate service set, wherein the dominated rule is as follows: suppose candidate services wsmAnd wsnWhen s ismThe confidence values of all QoS attributes are not better than wsnConfidence values corresponding to QoS attributes, and at least one QoS attribute exists, the confidence value being completely inferior to wsnConfidence values corresponding to QoS attributes, wsmIs dominated by;
3) and after all the candidate services are compared, obtaining a credible candidate service set.
Further, the constructing the multi-target model in the step (3) includes the following steps:
1) for the condition that the QoS attribute dimensions are not consistent, a normalization method is adopted to convert the QoS attribute value into a [0,1] range;
2) calculating the aggregation value of each QoS of the combined service as a target function according to the calculation formula of the QoS aggregation function under different workflow models, and setting QoS constraint;
3) and taking the credible candidate service set as a solving space to construct a multi-objective optimization model.
Further, the improved multi-target particle swarm algorithm in the step (3) comprises the following steps:
1) setting algorithm parameters: the method comprises the steps of (1) including population scale, maximum iteration times, formula parameters and the like;
2) initializing a population: initializing a population by adopting a good point set theory;
3) initializing a memory set: calculating the fitness of each particle according to a fitness function, and selecting non-dominated particles to be added into a storage set through non-dominated sorting;
4) initializing individual optimal and global optimal: taking the fitness value of each particle as the initial individual optimum of each particle, and selecting the optimal particle from the storage set according to the weight sorting of the target function as the global optimum;
5) initializing the iteration number to be 1;
6) update the velocity and position of the particle: calculating to obtain new particles through the following speed updating formula and position updating formula;
vj,t＝ωvj,t+c1r1(pj,best-xj,t-1)+c2r2(gbest-xj,t-1) (5)
xj,t＝xj,t-1+vj,t (6)
wherein j is the order of the particlesColumns; omega is the inertial weight; v. ofj,tThe particle velocity of the particle j when the iteration number is t; x is the number ofj,tThe position of the particle j when the iteration number is t; p is a radical ofj,bestIs optimal for the individual of particle j; gbestThe global optimization is achieved; c. C1Is an individual accelerating factor, c2Is a global acceleration factor; r is1And r2Is a random number between 0 and 1.
7) And (3) population updating: mixing the previous generation particle population with the newly obtained particles to obtain a temporary population, selecting the particles with population scale quantity as a new population, wherein the selection rule is as follows: firstly, selecting non-dominant particles from a temporary population according to non-dominant sorting, stopping selection if the number of the selected particles reaches the population scale, and selecting the non-dominant particles from the rest particle populations according to the non-dominant sorting again if the number of the selected particles does not reach the population scale until the number of the selected particles reaches the population scale;
8) updating individual optimum: calculating the fitness of each particle in the new population according to the fitness function, comparing the fitness with the current individual optimum, and selecting a more optimal particle as a new individual optimum;
9) updating the storage set: selecting non-dominant particles from the new population by non-dominant sorting, mixing the non-dominant particles with particles in the storage set, and selecting the non-dominant particles as a new storage set according to the non-dominant sorting;
10) updating global optimum: sorting according to the weight sum of the objective functions of the particles, and selecting the optimal particles as a new global optimum;
11) updating parameters: dynamically updating parameters in a speed updating formula and a position updating formula;
12) adding 1 to the iteration number, judging whether the iteration number reaches the maximum iteration number, and if not, circulating to the step 6); if the result is reached, the next step is carried out;
13) and outputting the storage set to obtain a series of combination schemes which meet the requirements of users.
Compared with the prior art, the technical scheme provided by the invention has the advantages that: calculating the QoS attribute reliability value by adopting a Bayesian learning theory and time factor combined method, so that the calculation of the reliability value is real-time and reliable; screening out the dominated candidate services by using a skyline theoretical method, so that the requirements of users are better met, and the obtained candidate service set is credible; and an improved multi-target particle swarm algorithm is adopted to solve the Web service combination scheme, so that a better Web service combination scheme is solved while the solving speed is ensured.
Drawings
FIG. 1 is a schematic diagram of a Web services portfolio workflow, according to one embodiment of the present invention;
FIG. 2 is a diagram illustrating credibility screening according to one embodiment of the present invention;
FIG. 3 is a schematic diagram of an improved multi-target particle swarm algorithm according to an embodiment of the invention.
Detailed Description
The following is a detailed description of embodiments of the invention in conjunction with the technical solutions (and drawings).
Fig. 1-3 illustrate an embodiment of a Web service composition method based on credibility screening according to the present invention.
FIG. 1 is a schematic diagram of a Web services composition workflow, according to one embodiment of the invention. The embodiment is composed of 6 subtasks through different workflow models, wherein the ellipse represents the subtask, including AS1,AS2,AS3,AS4,AS5,AS6The workflow includes a sequential structure, a selection structure, a loop structure and a parallel structure, which are marked by the characters in the figures. As shown, each subtask has a set of candidate services for selecting an appropriate Web service, represented by a rectangular pattern in the figure, such as [ WS ]1,1,WS1,2,…,WS1,n]Is a subtask AS1And so on.
Table 1 below is an aggregation function table for QoS under different working flows according to an embodiment of the present invention. The present embodiment employs four QoS attributes, respectively cost (Ps), response time (Ti), reliability (Av) and availability (Re), and the calculation in the table is an aggregation function for each QoS under different operating flows. Where Pi represents the probability of selecting a branch in the selection structure workflow and k represents the number of loops in the loop structure workflow.
TABLE 1 aggregation function for QoS under different working flows
Figure BDA0002792546830000051
According to the Web service composition workflow of fig. 1, the candidate service set of each subtask is subjected to credibility screening, and fig. 2 shows a credibility screening diagram according to an embodiment of the present invention, which includes the following steps:
1) suppose a subtask AS1Candidate service WS1,1Has a historical operating value of cost (Ps) of
Figure BDA0002792546830000052
Where i denotes the ith run, i ∈ [1,2, …, n]，diTable ith running value is from the current calculation days; the release value of the cost (Ps) of the candidate service is Ps, and the trust degree of each operation of the candidate service is calculated
Figure BDA0002792546830000053
As shown in the following calculation formula
Figure BDA0002792546830000054
Deriving a set of degrees of trust for the cost (Ps) of the candidate service
Figure BDA0002792546830000055
2) Setting distrust
Figure BDA0002792546830000056
Calculating the overall confidence level R of the selected historical running record in the period of timeλ,PsAnd overall distrust
Figure BDA0002792546830000057
The calculation formula is as follows:
Figure BDA0002792546830000058
Figure BDA0002792546830000061
wherein, λ is a time factor, which represents the importance degree of time to the confidence level.
3) Calculating the credibility value of the cost (Ps) of the candidate service according to Bayesian learning theory
Figure BDA0002792546830000062
The calculation formula is as follows:
Figure BDA0002792546830000063
4) respectively calculating the sub-tasks AS according to the steps 1) to 3) repeatedly1Obtaining a confidence matrix of confidence values of the four QoS attributes of each candidate service in the candidate service set
Figure BDA0002792546830000064
The following calculation formula is shown:
Figure BDA0002792546830000065
wherein each behavior of the matrix is a set of credibility of four QoS of a candidate service;
Figure BDA0002792546830000066
Figure BDA0002792546830000067
respectively representing candidate services WS1,nA confidence value of cost (Ps), a confidence value of response time (Ti), a confidence value of reliability (Av) and a confidence value of availability (Re); WS1,nRepresenting subtasks AS1The nth candidate service.
5) Selecting candidate services WS1,iComparing with the reliability set of other candidate services if WS1,iIf the service is dominated by at least one candidate service, deleting the service from the candidate service set;
6) sequentially comparing other candidate services according to the method of the step 5), and finally obtaining a credible candidate service set, namely a subtask AS1A set of trusted candidate services;
7) and sequentially screening candidate service sets of other subtasks according to the steps 1) to 6).
After a trusted candidate service set of each subtask is obtained, a multi-objective optimization model is constructed, assuming that users pay attention to combined service cost and response time, and constraint conditions are set for reliability and availability, the steps of constructing the multi-objective optimization model are as follows:
1) the cost and response time of the composite service is calculated according to the workflow of FIG. 1, as follows:
Figure BDA0002792546830000068
Figure BDA0002792546830000069
wherein WS1,i,WS2,i,WS3,i,WS4,i,WS5,i,WS6,iAnd respectively selecting from the credible candidate service sets after the 6 subtasks are screened.
2) The present embodiment sets reliability and availability as constraints for each subtask, as follows:
s.t:Av(ASi)≥Av0,Re(ASi)≥Re0 (14)
wherein, ASiSelecting a subtask from a trusted candidate service set; av0，Re0A set reliability threshold and an availability threshold.
3) According to the calculation formula, a multi-objective optimization model is constructed as follows:
Figure BDA0002792546830000071
the model represents the AS at the selected subtaskiMake the cost f of the combined service in case of satisfying the reliability and availability constraintsPs(CS) and fTi(CS) is as small as possible.
F in the multi-objective optimization model to be constructedPs(CS) and fTi(CS) as a fitness function, a credible candidate service set after each subtask is screened in the figure 1 as a solving space, constraint conditions as the limitation of selecting candidate services, and an improved multi-target particle swarm algorithm is adopted for solving to obtain a combination scheme of a Web service combination, wherein the improved multi-target particle swarm algorithm schematic diagram shown in figure 3 according to one embodiment of the invention comprises the following steps:
1) setting algorithm parameters: population size N, maximum number of iterations tmaxMinimum value ω of inertial weight ωminAnd maximum value ωmaxIndividual acceleration factor c1Initial value of c1,initAnd a final value c1,finGlobal acceleration factor c2Initial value of c2,initAnd a final value c2,fin；
2) Initializing a population: initializing a population by adopting a good point set theory;
3) initializing a memory set: calculating the fitness of each particle according to a fitness function, and selecting non-dominated particles to be added into a storage set through non-dominated sorting;
4) initializing individual optimal and global optimal: taking the fitness value of each particle as the initial individual optimal p of each particlebestSelecting the optimal particles from the memory set as global optimal g according to the weight and the sequence of the particle objective functionbest；
5) Initializing the iteration number to be 1;
6) calculating to obtain new particles by the following speed updating formula and position updating formula:
vj,t＝ωvj,t+c1r1(pj,best-xj,t-1)+c2r2(gbest-xj,t-1) (16)
xj,t＝xj,t-1+vj,t (17)
wherein j is the sequence of particles; omega is the inertial weight; v. ofj,tThe particle velocity of the particle j when the iteration number is t; x is the number ofj,tThe position of the particle j when the iteration number is t; p is a radical ofj,bestOptimal for particle j individuals; gbestThe global optimization is achieved; c. C1Is an individual accelerating factor, c2Is a global acceleration factor; r is1And r2Is a random number between 0 and 1.
7) And (3) population updating: mixing the previous generation particle population with the newly obtained particles to obtain a temporary population, and selecting particles with the number equal to the set population scale as a new population, wherein the selection rule is as follows: firstly, selecting non-dominant particles from a temporary population according to non-dominant sorting, stopping selection if the number of the selected particles reaches the population scale, and selecting the non-dominant particles from the rest particle populations according to the non-dominant sorting again if the number of the selected particles does not reach the population scale until the number of the selected particles reaches the population scale;
8) updating individual optimum: calculating the fitness of each particle in the new population according to the fitness function, comparing the fitness with the current individual optimum, and selecting a more optimal particle as a new individual optimum;
9) updating the storage set: selecting non-dominant particles from the new population through non-dominant sorting, mixing the non-dominant particles with particles in a storage set, and selecting the non-dominant particles as a new storage set according to the non-dominant sorting;
10) updating global optimum: selecting the optimal particles from the storage set as a new global optimum according to the target function weight and the sequence of the particles;
11) updating parameters: parameters omega, c for velocity update formula and position update formula1，c2And performing dynamic updating, wherein the calculation formula is as follows:
Figure BDA0002792546830000081
Figure BDA0002792546830000082
Figure BDA0002792546830000083
12) adding 1 to the iteration number, and judging whether the iteration number reaches the maximum iteration number tmaxIf not, circulating to the step 6), and if so, entering the next step;
13) and outputting the storage set to obtain a series of combination schemes which meet the requirements of users.
Further, the updating of the global optimum in step 10) is performed by using an objective function weight and a sorting method of the particles, and the method includes: calculating the value of the objective function, i.e. f, for the particles in the memory setPs(CS) and fTi(CS) setting the weight omega according to the degree of the user to pay attention to the twoPsAnd ωTiAnd carrying out weighted summation through two objective functions, and finally selecting the particles with the optimal weighted summation value as the global optimization.
Compared with the traditional multi-target particle swarm algorithm, the improved multi-target particle swarm algorithm is specifically improved as follows:
1) population initialization design
The invention adopts the theory of the best point set to initialize the population, so that the initial population can be uniformly distributed in a decision space.
2) Design of inertia weight and acceleration factor
The invention adopts a nonlinear strategy based on trigonometric function to control inertia weight and acceleration factor, as shown in formula (18), formula (19) and formula (20). The inertial weight ω decreases with increasing number of iterations, the individual acceleration factor c1The global acceleration factor c increases as the number of iterations increases2With the increase and decrease of the iteration times, the strategy can enable the algorithm to have stronger global search capability at the initial optimization stage and stronger local search capability at the later optimization stage, and finally can obtain better solving effect.
3) Design of population update
For population updating, mixing the previous generation particle population with the new generation particle population to obtain a temporary population, selecting particles with the size of the population scale as the new population, wherein the selection rule is as follows: first, non-dominant particles are selected from a temporary population according to a non-dominant sorting, and if the number of selected particles reaches the population size, the selection is stopped, and if the number of selected particles does not reach the population size, non-dominant particles are selected from the remaining particle population according to the non-dominant sorting until the number of selected particles reaches the population size. Compared with a mode that the multi-target particle swarm algorithm directly updates the population through a formula, the method provided by the invention updates the population in a mode of multiple times of non-dominated sorting, and can obtain a new population with higher quality.
4) Selection of global optimum
And for the selection of the global optimum, updating by adopting the objective function weight and the sorting method of the particles. And calculating the objective function value of the particles in the storage set, setting the weight of the QoS attribute according to the degree of the user to pay attention to the two, carrying out weighted summation through the objective function, and finally selecting the optimal weighted summation value as the global optimum. Compared with a grid method adopted by a multi-target particle swarm algorithm, the method disclosed by the invention better meets the requirements of users on Web service combination.
Finally, it should be noted that: although the present invention has been described in detail with reference to the specific embodiments, it should be understood by those skilled in the art that the present invention is not limited to the description of the embodiments, and various modifications and substitutions can be made within the spirit of the present invention.

Claims (3)

1. A Web service combination method based on credibility screening is characterized by comprising the following steps:
(1) aiming at a certain QoS attribute of Web service, calculating the credibility value of the attribute by adopting a Bayesian learning theory and a time factor method;
(2) calculating the credibility value of each QoS attribute of each candidate service in the candidate service set, and deleting the dominated candidate service by adopting a skyline theory according to the credibility value to obtain a credible candidate service set as a solving space for subsequent optimization; the method comprises the following specific steps:
1) according to the reliability value calculation method in the step (1), the reliability value of each QoS attribute of each candidate service in the subtask candidate service set is calculated to obtain the QoS attribute reliability set of the candidate service
Figure FDA0002792546820000011
Wherein the content of the first and second substances,
Figure FDA0002792546820000012
representing candidate services wsiThe set of QoS trustworthiness of (a),
Figure FDA0002792546820000013
representing candidate services wsiQ (a) to (b)nA confidence value for each QoS;
2) according to the credibility set, sequentially comparing the candidate service with other candidate services by using a skyline theory, and deleting the dominated candidate service from the candidate service set, wherein the dominated rule is as follows: suppose candidate services wsmAnd wsnWhen s ismThe confidence values of all QoS attributes are not better than wsnConfidence values corresponding to QoS attributes, and at least one QoS attribute exists, the confidence value being completely inferior to wsnConfidence values corresponding to QoS attributes, wsmIs dominated by;
3) after all candidate services are compared, a credible candidate service set is obtained;
(3) constructing a multi-target model according to each QoS attribute, optimizing the multi-target model by adopting an improved multi-target particle swarm optimization, and solving a Web service combination scheme; the method specifically comprises the following steps:
1) for the condition that the QoS attribute dimensions are not consistent, a normalization method is adopted to convert the QoS attribute value into a [0,1] range;
2) calculating the aggregation value of each QoS of the combined service as a target function according to the calculation formula of the QoS aggregation function under different workflow models, and setting QoS constraint;
3) and taking the credible candidate service set as a solving space to construct a multi-objective optimization model.
2. The Web service composition method based on credibility screening as claimed in claim 1, wherein the step (1) of calculating the credibility value of a certain QoS attribute of the Web service comprises the following steps:
1) and calculating the trust R according to the release value q and the historical operation value record q' of a certain QoS attribute of the Web service, wherein the calculation formula is as follows:
Figure FDA0002792546820000014
2) selecting the historical running record of the attribute in a certain period of time, and obtaining the trust level set of the attribute through the formula (1)
Figure FDA0002792546820000021
Wherein
Figure FDA0002792546820000022
Representing the degree of trust of the i-th run, diRepresenting the number of days from the current time at the i-th run;
3) assuming degree of distrust
Figure FDA0002792546820000023
And calculating the total trust degree and the total distrust degree of the attribute in the period of time according to the following calculation formula:
Figure FDA0002792546820000024
Figure FDA0002792546820000025
wherein, lambda is a time factor, lambda is more than 0 and less than or equal to 1, and represents the importance degree of time to the trust; rλ,qRepresenting the overall trust of the attribute q; i denotes the ith run, i ∈ [1,2, …, n]；
Figure FDA0002792546820000026
Indicating the overall degree of distrust for the attribute;
Figure FDA0002792546820000027
representing the degree of distrust of the i-th run, diRepresenting the number of days from the current time at the i-th run;
4) according to Bayes learning theory, calculating reliability value rep of the attributeqThe calculation formula is as follows:
Figure FDA0002792546820000028
3. the Web service combination method based on credibility screening of claim 1, wherein the improvement of multi-objective particle swarm algorithm in the step (3) comprises the following steps:
1) setting algorithm parameters: the method comprises the steps of (1) including population scale, maximum iteration times, formula parameters and the like;
2) initializing a population: initializing a population by adopting a good point set theory;
3) initializing a memory set: calculating the fitness of each particle according to a fitness function, and selecting non-dominated particles to be added into a storage set through non-dominated sorting;
4) initializing individual optimal and global optimal: taking the fitness value of each particle as the initial individual optimum of each particle, and selecting the optimal particle from the storage set according to the weight sorting of the target function as the global optimum;
5) initializing the iteration number to be 1;
6) update the velocity and position of the particle: calculating to obtain new particles through the following speed updating formula and position updating formula;
vj,t＝ωvj,t+c1r1(pj,best-xj,t-1)+c2r2(gbest-xj,t-1) (5)
xj,t＝xj,t-1+vj,t (6)
wherein j is the sequence of particles; omega is the inertial weight; v. ofj,tThe particle velocity of the particle j when the iteration number is t; x is the number ofj,tThe position of the particle j when the iteration number is t; p is a radical ofj,bestIs optimal for the individual of particle j; gbestThe global optimization is achieved; c. C1Is an individual accelerating factor, c2Is a global acceleration factor; r is1And r2A random number between 0 and 1;
7) and (3) population updating: mixing the previous generation particle population with the newly obtained particles to obtain a temporary population, selecting the particles with population scale quantity as a new population, wherein the selection rule is as follows: firstly, selecting non-dominant particles from a temporary population according to non-dominant sorting, stopping selection if the number of the selected particles reaches the population scale, and selecting the non-dominant particles from the rest particle populations according to the non-dominant sorting again if the number of the selected particles does not reach the population scale until the number of the selected particles reaches the population scale;
8) updating individual optimum: calculating the fitness of each particle in the new population according to the fitness function, comparing the fitness with the current individual optimum, and selecting a more optimal particle as a new individual optimum;
9) updating the storage set: selecting non-dominant particles from the new population by non-dominant sorting, mixing the non-dominant particles with particles in the storage set, and selecting the non-dominant particles as a new storage set according to the non-dominant sorting;
10) updating global optimum: sorting according to the weight sum of the objective functions of the particles, and selecting the optimal particles as a new global optimum;
11) updating parameters: dynamically updating parameters in a speed updating formula and a position updating formula;
12) adding 1 to the iteration number, judging whether the iteration number reaches the maximum iteration number, and if not, circulating to the step 6); if the result is reached, the next step is carried out;
13) and outputting the storage set to obtain a series of combination schemes which meet the requirements of users.
CN202011319892.3A 2020-11-23 2020-11-23 Web service combination method based on credibility screening Pending CN112511346A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011319892.3A CN112511346A (en) 2020-11-23 2020-11-23 Web service combination method based on credibility screening

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011319892.3A CN112511346A (en) 2020-11-23 2020-11-23 Web service combination method based on credibility screening

Publications (1)

Publication Number Publication Date
CN112511346A true CN112511346A (en) 2021-03-16

Family

ID=74959455

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011319892.3A Pending CN112511346A (en) 2020-11-23 2020-11-23 Web service combination method based on credibility screening

Country Status (1)

Country Link
CN (1) CN112511346A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113687780A (en) * 2021-07-30 2021-11-23 济南浪潮数据技术有限公司 Distributed storage server QoS optimization method, system, terminal and storage medium
CN114172963A (en) * 2021-10-15 2022-03-11 合肥工业大学 Multi-granularity task and service matching method and system based on resource utilization rate
CN114625497A (en) * 2021-12-28 2022-06-14 杭州电子科技大学 Credible service combination method based on cooperative sensing

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20130078037A (en) * 2011-12-30 2013-07-10 건국대학교 산학협력단 Qos-aware web service composition method using on-the-fly learning-based search
CN105635255A (en) * 2015-12-23 2016-06-01 河海大学 Skyline-based Web service selection method under data origin credibility constraint
CN106201706A (en) * 2015-04-30 2016-12-07 北大方正集团有限公司 A kind of population method for service selection and system
CN107016077A (en) * 2017-03-28 2017-08-04 南京邮电大学 A kind of optimization method of web oriented Services Composition
CN107070704A (en) * 2017-03-22 2017-08-18 东南大学 A kind of Trusted Web services combined optimization method based on QoS

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20130078037A (en) * 2011-12-30 2013-07-10 건국대학교 산학협력단 Qos-aware web service composition method using on-the-fly learning-based search
CN106201706A (en) * 2015-04-30 2016-12-07 北大方正集团有限公司 A kind of population method for service selection and system
CN105635255A (en) * 2015-12-23 2016-06-01 河海大学 Skyline-based Web service selection method under data origin credibility constraint
CN107070704A (en) * 2017-03-22 2017-08-18 东南大学 A kind of Trusted Web services combined optimization method based on QoS
CN107016077A (en) * 2017-03-28 2017-08-04 南京邮电大学 A kind of optimization method of web oriented Services Composition

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
何干志，刘茜萍: ""基于QoS综合评价的服务选择方法"", 《计算机技术与发展》 *
杨丹榕: ""基于可信度计算与QoS相结合的 Web服务组合优化问题研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113687780A (en) * 2021-07-30 2021-11-23 济南浪潮数据技术有限公司 Distributed storage server QoS optimization method, system, terminal and storage medium
CN113687780B (en) * 2021-07-30 2024-02-23 济南浪潮数据技术有限公司 QoS optimization method, system, terminal and storage medium for distributed storage server
CN114172963A (en) * 2021-10-15 2022-03-11 合肥工业大学 Multi-granularity task and service matching method and system based on resource utilization rate
CN114172963B (en) * 2021-10-15 2023-07-28 合肥工业大学 Multi-granularity task and service matching method and system based on resource utilization rate
CN114625497A (en) * 2021-12-28 2022-06-14 杭州电子科技大学 Credible service combination method based on cooperative sensing
CN114625497B (en) * 2021-12-28 2023-04-07 杭州电子科技大学 Credible service combination method based on cooperative sensing

Similar Documents

Publication Publication Date Title
Chen et al. Deadline constrained cloud computing resources scheduling for cost optimization based on dynamic objective genetic algorithm
CN112511346A (en) Web service combination method based on credibility screening
Zou et al. A novel global harmony search algorithm for task assignment problem
CN109818775B (en) Short-term network flow prediction method
CN107656799B (en) Workflow scheduling method considering communication and computing cost under multi-cloud environment
Xie et al. Trust-driven and PSO-SFLA based job scheduling algorithm on cloud
Zhang et al. A Novel Ant Colony Optimization Algorithm for Large Scale QoS‐Based Service Selection Problem
Zhang et al. A search space reduction-based progressive evolutionary algorithm for influence maximization in social networks
CN112306665B (en) Service integration method driven by sequence QoS
CN114172963A (en) Multi-granularity task and service matching method and system based on resource utilization rate
Teng et al. An Improved Whale Optimization Algorithm Based on Aggregation Potential Energy for QoS‐Driven Web Service Composition
Zhang et al. Optimising web service composition based on differential fruit fly optimisation algorithm
Zhang et al. Information guiding and sharing enhanced simultaneous heat transfer search and its application to k-means optimization
CN116887285A (en) Edge computing server placement method based on dung beetle optimization algorithm
Haytamy et al. Enhanced qos-based service composition approach in multi-cloud environment
Chen et al. GA for QoS satisfaction degree optimal web service composition selection model
CN114980216A (en) Dependent task unloading system and method based on mobile edge calculation
Sun et al. Fast-FFA: a fast online scheduling approach for big data stream computing with future features-aware
Zhou et al. Multi-objective service compositon optimization using differential evolution
Yao et al. Hybrid-tc: a hybrid teaching-learning-based optimization algorithm for service composition in cloud manufacturing
Kumar et al. Enhanced crow search algorithm for task scheduling in cloud computing
Alkhanak et al. A hyper-heuristic approach using a prioritized selection strategy for workflow scheduling in cloud computing
Farid et al. Weighted-adaptive Inertia Strategy for Multi-objective Scheduling in Multi-clouds.
Liu et al. A new task scheduling strategy based on improved ant colony algorithm in IaaS layer
CN115718648A (en) Augmented reality task unloading method based on heuristic algorithm under edge-side collaborative environment

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20210316