Developing a two level option trading strategy based on option pair optimization of spread strategies with evolutionary algorithms

Artificial intelligence methods are being used for a long time for portfolio optimization and asset pricing. In this study, a two level option trading strategy has been modeled and optimized by Genetic Algorithm and Particle Swarm Optimization. It is known that performance of financial strategies go up when prices are grouped by trend. Therefore, trending strategy is optimized in the first level. Short-term and long-term average prices are calculated for trending strategy. When short-term average is greater than long-term average, it is accepted as a sign of upward trend; and when short-term average is less than long-term average, it is accepted as a sign of downward trend. Length of short-term and long-term ranges are optimized in the first level. In the second level, trend is identified by results of the first level. Then, The Bull Call Spread strategy is used in upward trend days and The Bear Call Spread is used in downward trend days. Strike prices and expiration dates of the options to be traded are optimized. This model is tested on 5 different ETF's, results are compared with 3 different strategies in the literature and it is observed that this model makes highest profit among these strategies.