Towards Autonomously Developed Software: A genetic Approach in Critical and Embedded Systems

Abstract

Nowadays, we are still considering changes undergone by a software system as a sporadic phenomenon and we do not sufficiently anticipate future changes during the development phase. Consequently, many problems arise in the maintenance phase. In this study, we present an approach where all changes undergone by a software system are considered as its ontogenetic dimension. We represent this dimension by specific concepts as a continuous and well delimited process that is embedded in the software model of a system. Inspired by genetics, our approach proposes a model where anticipated and unanticipated changes are modeled by a collection of fine grained instructions called genes.