Applying Constraint Satisfaction Mode on Genetic Algorithm for Formulating Stock Trading Strategies

• Main Authors: Kai-Ling Yao, 姚凱齡
• Other Authors: Cheng-Liang Yang
• Format: Others
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/86233238092254620067

碩士 === 大同大學 === 資訊經營學系(所) === 97 === This research applies genetic algorithms (GAs) on investment decision making for selection, timing and capital allocation. Combine technical analysis with GAs in the timing decision can find out the optimal investment decision and help investors earning profits. The research proposes new encoding and decoding methods that chromosomes can conditionally fit in when initialing population and combining selection, capital allocation and timing decisions in one single chromosome. The experimental consequences point out that the new encoding method simplifies the procedure of evolution and accelerate the speed for searching solution. Successfully combined technical analysis with genetic algorithms and helped investors can therefore be benefited to trade stocks at appropriate opportunity in simulate transactions. The new model indeed earned profits and exceeded the interest rate of deposit certificate, as well as the average return rate of eight major types stock price index gained by using buy-and-hold strategy.