【作者】 陈阳；

【导师】 张宏国；

【作者基本信息】 哈尔滨理工大学 ， 软件工程， 2020， 硕士

【摘要】 随着大数据和大服务概念的兴起,服务产业迎来了新的变革。伴随着需求方对软件服务性能标准上的提高以及越来越复杂的软件业务流程,传统的软件开发模式已经变得不再适用。如何快速、高质量的开发服务产品成为了如今服务领域的一个热门研究方向。当对现有的软件服务进行组合时,用户除了关心功能性方面的可满足性外,软件服务的非功能性的可满足度也变得愈加重要。传统的服务质量(Quality of Service,Qo S)在对软件服务非功能属性描述时,并未考虑法律或者服务使用权限等方面的约束。针对上述问题,本文提出了SCOM(Service Contract oriented Model,SCOM)驱动的软件服务系统构建方法。在该方法中,本文首先利用Horn子句构建了满足用户功能性需求的抽象服务流程模型,接着引入了服务契约概念,扩充了Petri网的非功能性需求描述能力,构建了满足用户个性化的功能和非功能性需求的SCOM模型。接着,基于SCOM模型,提出了一种混合增强人工蜂群(Hybrid Enhancement Artificial Bee Colony,HEABC)算法,支持SCOM驱动的软件服务系统构建。最后根据上述研究内容,设计了一款小型的Web服务组合系统。具体研究主要包括以下几个方面:(1)提出了一种考虑用户输入和输出请求的Web服务组合方法。在使用Petri网对Web服务的抽象流程进行描述时,首先将注册后的服务转换为一组Horn子句的形式,将用户提供的输入和输出参数转换为Horn子句中的一组事实和目标,通过逻辑推理的方式构建完整的业务流程。(2)提出了基于服务契约(Service Contract,SC)的Petri网模型。原始的Petri网在对Web服务进行服务流程刻画时只包含了功能属性,并不能对非功能属性进行描述。本文将服务契约扩展至Petri网的流程构建过程中,通过基于服务器契约的属性选择匹配原子服务,完成服务组合的链式操作。(3)为了高效地实现服务选择,并利用服务聚合方法构建能够最大化满足用户需求的复杂软件服务系统,本文提出了一种混合增强人工蜂群(Hybrid Enhancement Artificial Bee Colony,HEABC)算法。该算法将K-means算法、KNN算法与ABC算法融合,保证ABC算法在离散解空间更新解时,始终保持解的连续性。进一步地,该算法通过增加蜜蜂群体之间信息共享的能力,增强了蜜蜂群体的探索和开发能力。更多还原

【Abstract】 With the rise of big data and big service concepts,new changes have taken place in the service industry.With the improvement of software service performance standards and the increasingly complex software business processes on the demand side,traditional software development models are no longer applicable.How to develop service products quickly and with high quality has become a hot research direction in the service field today.When combining existing services,in addition to focusing on functional consistency,the pros and cons of non-functionality of services have become increasingly important.Traditional quality of service(Qo S)does not consider legal or service restrictions when describing services.In view of the above problems,this paper proposes a SCOM(Service Contract oriented Model,SCOM)driven software service system construction method.In this method,this article first uses the Horn clause to construct an abstract service process model that meets the functional needs of users,and then introduces the concept of service contracts to expand the non-functional requirements description capabilities of Petri nets,and builds a personalized user satisfaction SCOM model for functional and non-functional requirements.Then,based on the SCOM model,a Hybrid Enhancement Artificial Bee Colony(HEABC)algorithm was proposed to support the construction of SCOM-driven software service systems.Finally,based on the above research content,a small Web service composition system is designed.The specific research mainly includes the following aspects:(1)A web service composition method considering user input and output requests is proposed.When describing the abstract flow of Web services using Petri nets,the registered service is first converted into a set of Horn clauses,and the input and output parameters provided by the user are converted into a set offacts and goals in the Horn clause.To build a complete business process through logical reasoning.(2)A Petri net model based on Service Contract(SC)is proposed.When the original Petri net described the service process of the Web service,it only contained its functional attributes and could not describe the non-functional attributes.In this paper,the service contract is extended to the Petri net process construction process,and the atomic service is selected and matched based on the attributes of the server contract to complete the chain operation of service composition.(3)In order to efficiently implement service selection and use service aggregation methods to build complex software service systems that can maximally meet user needs,a Hybrid Enhancement Artificial Bee Colony(HEABC)algorithm is proposed.This algorithm combines K-means algorithm,KNN algorithm and ABC algorithm to ensure that the ABC algorithm always maintains the continuity of the solution when updating the solution in the discrete solution space.Further,the algorithm enhances the ability to explore and develop bee colonies by increasing the ability to share information among bee colonies.更多还原