Publié le 25 avril 2024
–
Mis à jour le 17 mai 2024
Date(s)
le 16 mai 2024
GREDEG
We are delighted to welcome Herbert Dawid, innovation economist specializing in Agent-based Computational Economics, Industry Dynamics and Policy Design in Complex Environments, to Bielefeld University on May, 16, 2024.
Title: Agent-based analysis of the effect of algorithmic decision making in markets
(with Philipp Harting and Michael Neugart)
Abstract
In this paper we explore how the use of machine-learning based price setting algorithms by firms affects the emerging level of prices in oligopoly markets. In particular, we consider Deep-Q Networks algorithms, as an instance of state-of-the-art reinforcement learning, and show based on extensive agent-based market simulations that different aspects of the design of these algorithms have crucial impact on the emerging prices. In this respect our analysis contributes to the active literature on the danger of algorithmic collusion on markets.
Title: Agent-based analysis of the effect of algorithmic decision making in markets
(with Philipp Harting and Michael Neugart)
Abstract
In this paper we explore how the use of machine-learning based price setting algorithms by firms affects the emerging level of prices in oligopoly markets. In particular, we consider Deep-Q Networks algorithms, as an instance of state-of-the-art reinforcement learning, and show based on extensive agent-based market simulations that different aspects of the design of these algorithms have crucial impact on the emerging prices. In this respect our analysis contributes to the active literature on the danger of algorithmic collusion on markets.
