Genetic algorithms on technical trading rules
S.M. Hofman · 2013
In this paper, well know technical trading rules are used in different genetic algorithms to increase the performance on the S&P 500 exchange traded fund traded on the NASDAQ. Moving average rules, support and resistance rules, filter rules, momentum in price rules and channel breakout rules are combined to create one optimal rule to use on this data. The genetic algorithm creates an initial population of random rules and updates his population untill a population with this best rule is reached. The updating is of the same principle as Darwin’s survival of the fittest. In this way, less performing will not survive through the next generation and better performing rules are. With Genetic Programming, an extension of the GA developed by Koza (1992), another type of optimal rules will be created. We will check whether this method creates better trading rules than our previous work and will compare several different ways to create these rules, when we end with a best rule for our data.