There is increasing interest towards techniques for calibrating and validating Agent-Based Models (ABMs). Several researchers at EMbeDS and at the Institute of Economics have extensively worked on these topics in recent years. We will organize some brainstorming sessions on calibration and validation of ABMs on February 22-23, 2022, with the goal of discussing recent developments and exploring future directions.
On these days, Jiri Kucacka, an assistant professor at Charles University in Prague and at the Czech Academy of Sciences, will visit us. He will present a paper on “estimation of heuristic switching in behavioral macroeconomic models”, part of his recent research on the use of simulated maximum likelihood methods to calibrate ABM parameters. For more details, see here.
We will also hear two presentations by Mario Martinoli and Marco Pangallo, two postdocs at EMbeDS and at the Institute of Economics. Mario will present his research on “A General Protocol for the Calibration and Validation of Large-scale ABMs”, while Marco will present ongoing work on “Quantifying features of simulation models directly from model code”. These talks are part of the “Young Researchers Informal Seminars” co-organized by Mario and Marco that involve PhD students, postdocs and assistant professors of the Institute of Economics.
We will finally organize brainstorming sessions in small groups. These sessions will cover updates on current research directions in calibration/validation, the creation of an R package for calibration and validation of ABMs, and new research directions to bring ABMs closer to real-world data, including initialization and data assimilation.