Un modelo de Razonamiento Basado en Casos para la captación de requisitos en el desarrollo de proyectos de software

Abstract

Studies show that one of the causes of delay in software projects is in the Collection Requirements. We also know that the cost to repair a requirements error is 5 to 10 times less than in the coding and 200 times less than in maintenance. Therefore, there are many efforts to mitigate this problem. For example: tools and techniques that make working in the administration and management re- quirements, such as RUP, REM, Interviews, surveys, Case uses, Win Win, and so on. However according to the latest reports are the problems. That is why in this paper we propose a model case- based reasoning to capture requirements in the software development project. With this artificial intelligence technique, is used the functional requirements of software projects previously deve- loped to solve or identify functional requirements of a new project. For this purpose the functional, requirements are represented as “cases”. The design has provided CBR tasks: represent the case, model Case Library (LC), define the indexing technique and fill the requirements of LC with previous projects. The model also considers the processes of recovery, reuse, revise and retain a require- ment. The 20 case studies of two projects requirements show that the proposed model reduces to 79.84% the time required to capture obtained by the RUP.