A Component Selection Approach In Internetware Based On GA

Internetware as a new software paradigm is coming with the rapid development of information technology; Along with the increasing maturity of the software development methods and technologies based on component, research on Internetware from the angle of component, which is a key point of studying on Internetware. How to make full use of the existing components resource to choose the right components to construct Internetware to help users complete more complex tasks from huge quantities of components is Components Choosing Problem in Internetware. Generally, the choosing of components includes two steps which are component retrieval and component selection. Component selection is choosing the component from the set of the results of the component retrieval.To address the above issues, this paper analyzes the characteristics of the Internetware and the research findings of existing technologies and methods of component choosing, uses the common retrieval method based on function facet to retrieve components which will get retrieval set, then introduces the genetic algorithm to search the retrieval set, and chooses set of components that meet the user needs to construct Internetware to complete the more complex tasks in real word. According to the train of thought above, the main work of the paper are organized as following:Analyzing the relevant research achievement at home and abroad, combining with the characteristics about Internetware, in view of the problem that it is difficult to find the right components meeting non-functional requirements by using existing component retrieval technology and method, this paper describes CSPI with the right method and maps the CSPI into a MMKP combined with the characteristics of the MMKP.Analyzing the MMKP, combining with the characteristics of CSPI, this thesis makes full use of the advantages of GA—strong adaptability,excellent expansibility,high fault tolerance and global exploration and uses GA to solve MMKP to select suitable components to construct Internetware.For the "premature" phenomenon of the experimental process sometimes, the idea of positive feedback in the ACO is introduced into the GA to mutate on purpose, and takes the appropriate selection strategy to choose the more suitable component. Simulation results show that it can better solve the component selection in Internetware by genetic algorithm. The improved genetic algorithm can avoid "premature" phenomenon to a certain extent, and be able to choose right components users need.