Application of a Switched PIDD Type Control Strategy to the Model-Free Algorithmic Trading

Abstract

Our paper is devoted to an applications of the feedback control methodology to the modern Algorithmic Trading (AT). The controller we propose performs a specific four term control design consisting of proportional, integral, derivative, and second order derivative terms (PIDD). The designed PIDD based trading strategy also incorporates a switched (dynamic) structure. We develop a specific switching mechanism that involves a backtesting driven profit maximization procedure on the historical data. We also study here the resulting model-free PIDD trading strategy in the frequency domain. Additionally, we give a rigorous formal estimation of the win rate for the trading algorithm under consideration. We finally apply the obtained AT strategy to a real-world stock market, namely, to the Binance Bitcoin/USD price dynamics.