A hybrid (μ+λ) evolutionary algorithm for evolving simple trading rules: Case study on stock Exchange of Thailand (SET50)

This paper proposes a new hybrid method for generating trading rules. The proposed technique employs genetic algorithm to evolve the rules´ structure and uses evolution strategy to determine constant parameters, simultaneously. The evolved trading rules are tested with historical data of SET50. The results are compared with a commonly used indicator (MACD). The testing period starts from June to October 2013, which the SET50 index dropped 3.77%. Generally, it is difficult for traders to earn profit in a downtrend market. However, the proposed method yields the 3.94% average profit return while MACD incurs -0.52%. It shows that the proposed technique can find profit generating rules although the overall market is bearish.