Using Genetic Algorithms to Construct a Low-risk Fund Portfolio Based on the Taiwan 50 Index

This research is mainly based on the index funds constructing methods of Oh, et al (2005) and Chang and Lai´s (2006) Adjusted-GA model. Then, a new model is constructed based on the Taiwan market environment and its characteristics. The fixed-rate deposit offered at the Taiwan was added to the modified model to achieve optimal portfolio performance. The results show that the modified stock selection model is capable of reducing the tracking error and is more suitable for investment portfolio selects stocks in Taiwan. The modified model shows that the optimal portfolio is comprised of 20 stocks in order to diversify risk. Furthermore, adding the fixed-rate deposit into the modified model will result in better portfolio performance.