Optimal Technical Indicator-based Trading Strategies Using NSGA-II

TL;DR

This paper introduces an optimized method for technical indicator-based stock trading strategies using NSGA-II, focusing on maximizing returns and minimizing risks through a bi-objective evolutionary algorithm.

Contribution

It presents a novel application of NSGA-II to optimize technical trading strategies, incorporating transaction costs and domain expertise for improved stability.

Findings

Strategies outperform baseline methods during stable periods

Inclusion of transaction costs improves real-world applicability

Bi-objective optimization balances return and risk effectively

Abstract

This paper proposes non-dominated sorting genetic algorithm-II (NSGA-II ) in the context of technical indicator-based stock trading, by finding optimal combinations of technical indicators to generate buy and sell strategies such that the objectives, namely, Sharpe ratio and Maximum Drawdown are maximized and minimized respectively. NSGA-II is chosen because it is a very popular and powerful bi-objective evolutionary algorithm. The training and testing used a rolling-based approach (two years training and a year for testing) and thus the results of the approach seem to be considerably better in stable periods without major economic fluctuations. Further, another important contribution of this study is to incorporate the transaction cost and domain expertise in the whole modeling approach.