Abstract for: Automation of Loop Eigenvalue Elasticity Analysis

In this paper, the automation of the loop eigenvalue elasticity analysis (LEEA) is proposed. LEEA is a methodology for generating a formal analysis of system dynamics models and relies on the analysis of the structure of a linearized model to infer from its eigenvalues and eigenvectors the model portions most related to a given behavior of interest. In order to apply the LEEA so far has been necessary to get involved in a manual and lengthy process, that includes among its several steps the decoding of complex functions into simplest versions and deleting some parts of the model and adding some others that are used as auxiliary variables in order to have the model generator simulation suite (such as VENSIMŪ) interact with numerical computation software (such as MATLABŪ) back and forth to provide the LEEA analysis. This research provides an efficient tool which integrates the LEEA methodology without modifying the created models. Three models were tested and computational experiments indicate a good behavior of the tool compared with the results obtained in the existing literature.