Parallel evolutionary algorithms for stock market trading rule selection on many-core graphics processors

Abstract

This chapter concerns stock market decision support systems that build trading expertise on the basis of a set of specific trading rules, analysing financial time series of recent stock price quotations, and focusses on the process of rule selection. It proposes an improvement of two popular evolutionary algorithms for rule selection by reinforcing them with two local search operators. The algorithms are also adapted for parallel processing on many-core graphics processors. Using many-core graphics processors enables not only a reduction in the computing time, but also an exhaustive local search, which significantly improves solution quality, without increasing computing time. Experiments carried out on data from the Paris Stock Exchange confirmed that the approach proposed outperforms the classic approach, in terms of the financial relevance of the investment strategies discovered as well as in terms of the computing time. © 2011 Springer-Verlag Berlin Heidelberg.