Work detail

Agent-based modeling of dual currency economy

Author: Bc. Bořivoj Vlk
Year: 2015 - summer
Leaders: PhDr. Ing. Jiří Skuhrovec PhD.
Consultants:
Work type: Bachelors
Language: English
Pages: 69
Awards and prizes:
Link: https://is.cuni.cz/webapps/zzp/detail/166560/
Abstract: Thesis describes close relationship between Dynamic programming and reinforcement
learning algorithms on the example of a model of a dual currency
economy. Dynamic programming is methodology used for deriving equilibria
of Search-Theoretic equilibrium monetary models, which provide evidence for
emergence of fiat currency or emergence of internationally circulating currencies
without any human institutions. Particular previously published SearchTheoretic
framework of dual currency economy is used as a background for development
of Agent-based Computational model. Both models are compared
based on their ability to reach specified equilibria and their assumptions, with
the conclusion, that models are closely related and with the same assumptions
would have the same results. Agent-based model also provides possibility of
relaxing assumptions on perfect information distribution and static environment.
In this setting, the model will reach different equilibria, that correspond
better to the real human behavior, observed in previously published laboratory
experiments.

Partners

Deloitte

Sponsors

CRIF
McKinsey
Patria Finance