Reinforcement Learning Applied to a Cryptocurrency Portfolio in a Complexity Environment

Resumen

Recently, cryptocurrencies have been used as financial assets and have presented positive returns, albeit their volatility is high. This paper aims to elaborate a hypothetical cryptocurrency portfolio and to do so, employs machine learning and an optimization algorithm to define the ideal amount to be allocated in each asset. The results show the hypothetical portfolio presents superior returns and lesser volatility compared to other allocation strategies.