Enhancing trading strategy performance by using memetic algorithm with a diverse technical indicator pool

In investment decision-making, traders and investors often rely on technical analysis, which focuses on price behavior. Technical analysis involves using indicators such as moving averages, momentum, and volume to determine optimal buy or sell points. However, the challenge lies in selecting the right combination and parameters for these indicators. Existing research proposed optimization methods using memetic algorithms that combine global search techniques like genetic algorithms with local search methods like simulated annealing to optimize trading strategies. This paper proposes the Memetic-based Algorithm with a Diverse Technical Indicator Pool (MADTIP) framework for trading strategy optimization, which integrates indicators from four key categories—trend, momentum, volume, and volatility. By ensuring that a trading strategy is composed of diverse indicators, the MADTIP framework enhances profitability and provides more robust solutions that can effectively navigate the dynamic nature of financial markets. Experimental results on real datasets indicate a significant improvement in performance from the proposed approach compared to the previous approach.