Deep Learning for Self-Adaptive System Analyzer Improvement

The efficiency of the analysis phase in the Autonomic Control Loop (ACL) plays a critical role in ensuring the Quality of Service (QoS) of Self-Adaptive Systems (SASs) with large adaptation spaces. This means that accurately identifying current issues within the allocated time guarantees the feasibility of self-adaptations. However, the analysis phase faces many challenges, due to the inherent complexity and noisy features within SASs. To address these challenges, we introduced an efficient Deep Learning (DL) model for multi-classifying QoS. Our findings demonstrate that our model enhances performance in a short time, ultimately preserving QoS.