Convergence of trading strategies in continuous double auction markets with boundedly-rational networked traders

Authored by Junhuan Zhang, Peter McBurney, Katarzyna Musial

Date Published: 2018

DOI: 10.1007/s11156-017-0631-3

Sponsors: No sponsors listed

Platforms: No platforms listed

Model Documentation: Other Narrative Flow charts Mathematical description

Model Code URLs: Model code not found

Abstract

This paper considers the convergence of trading strategies among artificial traders connected to one another in a social network and trading in a continuous double auction financial marketplace. Convergence is studied by means of an agent-based simulation model called the Social Network Artificial stoCk marKet model. Six different canonical network topologies (including no-network) are used to represent the possible connections between artificial traders. Traders learn from the trading experiences of their connected neighbours by means of reinforcement learning. The results show that the proportions of traders using particular trading strategies are eventually stable. Which strategies dominate in these stable states depends to some extent on the particular network topology of trader connections and the types of traders.
Tags
Social networks Agent-based modeling behavior Dynamics market microstructure Model Prediction Zero-intelligence Stock-market Reinforcement learning Investment decisions Automated trading Continuous double auctions Neural-network