Técnicas automáticas para la elaboración, validación y verificación de requisitos de Software

• Autores: Renzo Gastón Degiovanni
• Directores de la Tesis: Nazareno Matías Aguirre (dir. tes.)
• Lectura: En la Universidad Nacional de Córdoba (UNC) ( Argentina ) en 2015
• Idioma: español
• Enlaces
  • Tesis en acceso abierto en: hdl.handle.net
• Resumen
  • español

    Es ampliamente aceptado que los errores son más fáciles (y menos costosos) de corregir si se capturan lo más temprano posible en el proceso de desarrollo. Luego, obtener una especificación de requisitos de buena calidad, es de fundamental importancia práctica y económica en la mayoría de las metodologías de desarrollo de software modernas. Esto motivó a que las notaciones formales para la especificación de requisitos adquirieran mayor atención en las últimas décadas. En este trabajo, se presentan dos técnicas automáticas que brindan soporte al proceso de elaboración y validación de requisitos de software. En breve, éstas se basan en manipulaciones de formulas lógicas, explotando la eficiencia de varios mecanismos de análisis provenientes de los métodos formales, como SAT solving, model checking e interpolación.

  • English

    Software Engineering aims at providing mechanisms for developing, designing and maintaining large-scale quality software, in an effective and systematic way.

    Several studies have shown that most software errors are caused by inconsistencies, misunderstandings and imprecisions that arise in early phases of software development, e.g., during requirements engineering, and in particular during the elicitation, elaboration and specification of software requirements. In addition, requirements related errors are more expensive and difficult to solve as the development process progresses. Therefore, the quality of software requirements specifications is very important, for practical and economical reasons, in most modern software development methodologies.

    For this reason, in the last decades some formal requirements engineering methods (e.g., statecharts, SCR and other tabular notations, the KAOS method and Clafer) have gained more attention. Typically, these formal requirements engineering methodologies have associated powerful analysis mechanisms to assess and improve the quality of requirements specifications, such as the analysis of consistency (no contradicting requirements) and completeness (no missing cases).

    On the other hand, there currently exist various algorithms that efficiently solve different kinds of analysis over logical formalisms, such as boolean satisfiability (SAT solving) or the computation of interpolants. Because of the efficiency and relative scalability that characterise these algorithms, they have been successfully used in many practical applications, such as software verification, testing and AI planning.

    In this thesis, we present automated techniques that concern software requirements elaboration, validation and verification. Briefly, our techniques are based on the manipulation of logical formulas, taking advantage of the efficiency provided by the mechanisms associated with formal analysis techniques such as SAT solving, model checking and interpolation computation. We will assess our approaches on a variety of case studies, to show they can reach better levels of scalability when compared with related state-of-the-art techniques.