Medidas para avaliação da confiabilidade de sistemas autoadaptativos

Abstract

Self-adaptive systems (SAS) are capable of changing their behavior at runtime and are associated with domains such as intelligent spaces, aviation systems, and healthcare. These systems present dynamic and complex operations, and to provide adequate services, they require high levels of quality. One way to help ensure the quality of these systems is to perform quality assessments that are mostly based on software measures that, in turn, seek to quantify system attributes from various perspectives. However, in order to evaluate the quality of self-adaptive systems, the measures must contain elements appropriate to the specificities of SAS, such as dynamic reconfiguration and communication with sensors, actuators, and other systems. Aiming to identify characteristics, attributes, and quality measures important for SAS, a systematic mapping was first performed. In this work, the reliability characteristic was then identified as one of the most recurrent characteristics in the selected studies and according to the studies selected, reliability has a significant impact on the functioning of self-adaptive systems. The systematic mapping also showed that reliability characteristic had few measures, and its sub characteristics are poorly explored. Then, this work proposes a set of software measures aimed at the reliability of SAS and evaluate them related to their applicability and benefits. At the end, this work aims to provide support to the evaluation of SAS reliability and, consequently, contribute to the improvement of the quality assurance of these systems.