Optimization Method Of Service Solution Construction Based On Requirement-service Matching

Nowadays,with the rapid development of information science and technology,a large number of intelligent business services supported by advanced IT technologies have emerged in various fields.Thus these services formed a complex service ecosystem which has evolved from traditional domain separation into multi-domain integration.How to design an optimized service solution for the specific requirement of customer in a complex environment,and how to construct and implement the corresponding service system through means of service resource reuse,has become a difficult problem.Given the problems existing in the process of constructing domain service solutions in the complex service ecosystem,this thesis focuses on the domain service optimization problems(recommendation,selection,and composition).In the context of the tremendous changes in the service ecosystem,this thesis proposes a service solution construction and optimization method based on requirement-service matching.On the one hand,this method uses service history experience to summarize the requirement patterns of customers.On the other hand,this method mines the long-term service patterns of providers to meet customer requirements.Then,the pairwise matching mechanism of requirement patterns and service patterns is established to support the quick mapping from customer requirement to service solution.The main research work includes the following aspects:(1)Based on “Requirement-Engineering” Two-Phase of Service Engineering Paradigm(RE2SEP),this thesis analyzes the significant characteristics of three different types of requirements in the service system to identify the main problems,and introduces how to optimize the system under the guidance of the RE2 SEP.Specifically,the problems in the service system are divided into the service recommendation method based on pattern matching and time interval prediction,the service selection method based on requirementservice probability matrix and the service composition method based on pattern-pairwise matrix.According to the characteristics of different problems,this thesis designs the corresponding solutions,and proposes the optimization method of service solution based on requirement-service matching.(2)This thesis proposes a service recommendation method based on pattern matching and time interval prediction.In view of the sharp increase in the number of services,this thesis effectively discovers the potential requirements of users by mining the knowledgein history records,and recommends the appropriate services to the target users at the right time.Specifically,this thesis proposes a candidate services discovery method based on pattern matching.By analyzing potential requirement patterns,a similar user group of target users is found based on user clustering and preference metrics.The potential candidate services of the target users are discovered by mining the high frequency patterns in the service usage sequence of similar user groups.Secondly,this thesis proposes a service recommendation method based on candidate services and time interval prediction.By generating a probability model,this method can effectively estimate the time interval between different services,and optimize the partial order relationship among services based on Bayesian personalized sorting.The quality of recommendation and the user experience are both improved by recommending the right service at the right time to the target user.(3)This thesis proposes a service selection method based on requirement-service probability matrix.Due to the huge search space,it is difficult for existing methods to select the best quality service from a large number of candidate services within a limited time,and to construct a feasible service solution that meets the user's quality requirements.In order to solve this problem,this thesis proposes an efficient service selection strategy based on priori knowledge.This method uses the prior knowledge obtained from historical data as a guide to construct a probability matching matrix,effectively reduces the search space of service selection problems,so as to improve efficiency.Secondly,Qo S-aware service selection is an NP-hard problem.Researchers apply various evolutionary algorithms to reduce time complexity by finding near-optimal solutions.In order to solve the problem of algorithm parameter tuning,the series of dependencies are studied,which exists among the characteristics of the service selection problem,the parameter values of the evolutionary algorithm and the optimization results.Under the premise of avoiding many cumbersome attempts and seeking an optimal parameter setting,the service selection problem is solved efficiently.(4)This thesis proposes a service composition method based on pattern-pairwise matrix.In order to meet the more complex functional service requirements proposed by users,rather than being limited to the standardized processes provided by the system,service composition technology is widely used to build complex,value-added composite services.In order to improve the efficiency of service composition,the main researches include the ontology construction and fusion method proposed to solve the data hetero-geneity,the functional service pattern mining method that improves the discovery efficiency of related services by discovering the functional service patterns in the historical service record,the formal concept analysis method to discover the functional requirement pattern,the matrix construction method to describe the even relationship between the functional requirement pattern and the functional service pattern,and the service composition method based on the pattern-pairwise matrix.Compared with the method of constructing a service solution from scratch,this method effectively reduces the search space of related services by considering the functional service patterns,and improves the search efficiency,so it has achieved good experimental results.