Quantitative validation of agent-based simulation models: a protocol proposal for sensitivity analysis and model parameter validation with empirical data

• Grant number: 15/09760-3
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• Start date: September 01, 2015
• End date: January 31, 2017
• Field of knowledge: Applied Social Sciences - Economics - Industrial Economics
• Principal Investigator: David Dequech Filho
• Grantee: Marcelo de Carvalho Pereira
• Host Institution: Instituto de Economia (IE). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil

Abstract

The project deals with the development of a methodological protocol and software tools for the validation of agent-based simulation models and their application in three specific cases. It is expected to contribute to the methodological consolidation of such an analytical tool, increasing its robustness and facilitating its dissemination. The expected results are (i) a protocol proposal and the development of user-friendly software based on appropriate econometric methods (meta-modeling and Bayesian inference) for the quantitative validation of agent-based models, in particular to evaluate the parametric configuration space and its relation to the empirical data, and (ii) through the application of these tools, a robust analysis of three simulation models regarding industrial organization and the micro-macro dynamics of the labor market. The economic literature on the subject is still quite limited, demanding methodological research in other scientific fields (computer sciences, mathematics, and engineering). The existing software requires advanced knowledge of IT/programming, currently a major barrier to researchers. The project aims to contribute to mitigating these problems and, in comparison to the current best practices, to a substantially more robust validation of the three proposed models, which are representative of the variety of topics enabled whose research by the simulation methodology.

Scientific publications

(6)

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

DOSI, GIOVANNI; PEREIRA, MARCELO C.; VIRGILLITO, MARIA ENRICA. The footprint of evolutionary processes of learning and selection upon the statistical properties of industrial dynamics. INDUSTRIAL AND CORPORATE CHANGE, v. 26, n. 2, p. 187-210, APR 2017. (15/09760-3)

DOSI, G.; PEREIRA, M. C.; ROVENTINI, A.; VIRGILLITO, M. E.. When more flexibility yields more fragility: The microfoundations of Keynesian aggregate unemployment. JOURNAL OF ECONOMIC DYNAMICS & CONTROL, v. 81, p. 25-pg., 2017-08-01. (15/09760-3)

DOSI, GIOVANNI; PEREIRA, MARCELO C.; ROVENTINI, ANDREA; VIRGILLITO, MARIA ENRICA. The effects of labour market reforms upon unemployment and income inequalities: an agent-based model. SOCIO-ECONOMIC REVIEW, v. 16, n. 4, p. 687-720, OCT 2018. (15/09760-3)

DOSI, G.; PEREIRA, M. C.; ROVENTINI, A.; VIRGILLITO, M. E.. When more flexibility yields more fragility: The microfoundations of Keynesian aggregate unemployment. JOURNAL OF ECONOMIC DYNAMICS & CONTROL, v. 81, n. SI, p. 162-186, AUG 2017. (15/09760-3)

DOSI, G.; PEREIRA, M. C.; VIRGILLITO, M. E.. On the robustness of the fat-tailed distribution of firm growth rates: a global sensitivity analysis. JOURNAL OF ECONOMIC INTERACTION AND COORDINATION, v. 13, n. 1, SI, p. 173-193, APR 2018. (15/09760-3)

DOSI, G.; PEREIRA, M. C.; VIRGILLITO, M. E.. On the robustness of the fat-tailed distribution of firm growth rates: a global sensitivity analysis. JOURNAL OF ECONOMIC INTERACTION AND COORDINATION, v. 13, n. 1, p. 21-pg., 2018-04-01. (15/09760-3)