Selection of parameters and stages of genetic algorithm for investment purposes

Abstract

The stages and structure of the genetic algorithm, which can be used to optimize the parameters of investment strategies, are described, such as evaluation functions, technical indicators, strategy models (chromosomes), and the stages of the algorithm itself - selection, crossover, mutation. Possible parts of the chromosome - strategies in investing can be the parameters of technical indicators, for example, for a moving average, these are periods of a moving average, for a stochastic oscillator - time periods, conditions for entry and exit from a position, a method of capital management, risk strategies, i.e., conditions for behavior in risky situations operations such as stop-loss and take-profit to minimize losses and maximize profits. Various strategies can be used for the selection stage, including: proportional selection, tournament selection, ranked selection, elite selection, selection by stochastic universal sampling.