About this project#
Platform for Artificial Market Simulations is OSS project for artificial market simulations. This is implemented and distributed as a python package. This project inherits the concept of plham . However, corresponding to the recent usage of deep learning on python, this project decided to employ a python-based system.
Goals#
Artificial market simulation is a multi-agent simulation and run virtual markets on your computer. In artificial markets, there exists virtual markets and virtual agents imitating the real traders. By piling up the agents' behaviors, markets works well and the market prices are decided. Through artificial market simulations, we can analyze the mechanisms of what happens in the actual markets.
Recently, deep leaning is introduced to many researches, and it is also important for financial markets. Owing to the development of deep learning, the combination of artificial market simulations and deep learning also become important. For addressing this motivation, PAMS are developed.
Features#
The features of PAMS are the following:
time scale is tick-time
can replicate flash crashes
long-term simulations, such as simulations over the years, are also possible
multiple stocks can be simulated
index-ETF or options can be handled
Moreover, the followings are the feature of this platform:
It can be run on stand-alone computer, clusters, google colab, or many other computers
Small setting modification can be available only by changing config file
New classes (agents, markets) are easy to be introduced