Autómata Evolutivo (AE) para el mercado accionario usando martingalas y un algoritmo genético

Abstract

Evolutionary Automaton (AE) for the Stock Market Using Martingales and a Genetic Algorithm

The aim of this paper is to develop an Evolutionary Automata (EA) that operates considering a martingale model, which defines investment strategies using immediate historical information, profit or lost limits and stopping time. The EA offers signals to buy, sell or hold a determinate stock, these signals are based on the optimal combination of simple moving average through a genetic algorithm. The EA was tested with two stock indices, before, during and after the subprime crisis. Finding that when the stock markets were in bull market, the Buy-Hold (BH) strategy presented superior performance than the EA; in contrast, during the crisis period, the performance of the EA was better than BH, finally, for all of the test period, the performance of the EA was superior. The EA have the restriction that can be used by one stock/index for period, though this can be solved by cycling the number of instruments of the portfolio. The authenticity of this research is the combination of the above models that match between them in order to generate a system that helps investor decisions.