FRANCISCO 3.0: A HYBRID DEA-MCDM ARCHITECTURE FOR DECISION SUPPORT IN COMPUTER NETWORKS

The FRANCISCO software is a tool for computer network analysis and optimization based on fractal theory, self-similarity, and Data Envelopment Analysis (DEA). This work presents FRANCISCO 3.0, a hybrid DEA-MCDM architecture that integrates the multicriteria methods VIKOR, ELECTRE III, and PROMETHEE II into the original software environment, incorporating subjective decision-maker preferences into the analysis process. The implementation was carried out in R using the MCDA package while preserving compatibility with the software architecture and interface. The methods were evaluated using FRANCISCO’s default public dataset, composed of virtual network DMUs generated through Apache Bench experiments and compared with the DEA SCCR and multiplicative SCCR super-efficiency models. The results demonstrated consistent rankings and complementarity between objective and subjective decision-support approaches, expanding the software’s analytical capabilities for complex computer network optimization scenarios.