Genetic algorithms running into portable devices: A first approach

Abstract

Nowadays, smartphones and tablets are essential parts of our daily life. In research, lines of work in advanced algorithms are always wanted to explore the advantages and difficulties of new computing platforms. As an obvious combination of these two facts, analyzing the performance of intelligent algorithms (such as metaheuristics) on these portable devices is both interesting for science and for building new high-impact apps. Thus, we here design and evaluate a genetic algorithm executed over two kinds of portable devices (smartphone and tablet), as well as we compare its results versus a traditional desktop platform. Among several contributions, we mathematically model the running time to analyze the numerical performance of the three devices. Also, we identify weak and strong issues when running an intelligent algorithm on portable devices, showing that efficiency and accuracy can also come out of such computing limited systems.