Self-configurable Manufacturing Industrial Agents (SMIA) performance evaluation dataset

This dataset presents a comprehensive empirical evaluation of the Self-configurable Manufacturing Industrial Agents (SMIA) software, focusing on its computational efficiency and scalability, in terms of deployment aspects and standardized model analysis and interpretation (Asset Administration Shell (AAS) models enriched by CSS model semantics). The archive contains high-resolution metrics and aggregated data derived from four different scenarios. The first three experiments were structured and designed to obtain meaningful results, considering that this paper focuses on SMIA's architectural innovation and its ability to parse and interpret semantically enriched AAS models, while preparing to become part of a MAS. Two aspects related to the SMIA's key processes were analyzed: the self-configuration process of DTs and their operational readiness in the MAS. Additionally, an exhaustive comparison with alternative solutions (NOVAAS and FA3ST) was also performed. Organized according to FAIR principles, the dataset includes raw time-series metrics, pre-processed analytical files, analytical R commands, reproduction scripts, and the AASX files used as input for the software, facilitating full reproducibility of the reported benchmarks.