| Abstract |
In the future smart city, citizens are interconnected and easily share information anywhere, anytime. Through a sensor network integrated with real time monitoring systems, data are collected, processed and analyzed. Of course, this is already happening, in part. Nowdays, the number of portable devices that are available to all people is huge and we can find them everywhere, they are not only smartphones but also tablets, IoT, and other. This is a perfect scenario to start new lines of research on the actual suitability of portable devices to solve real optimization and machine learning problems. On the one hand, the fact that they are everywhere encourages research aimed at their collaboration in a distributed way. On the other hand, genetic algorithms are metaheuristics where parallelization takes on great importance. In this paper, we analyze the numerical behavior of distributed genetic algorithms on portable devices. We focus on the behavior of the distributed algorithm when we modify the number of interconnected devices, as well as the behavior of the algorithm when the devices with different performances collaborate together. As a conclusion, the numerical results support the future research in the concept of distributed intelligence everywhere, since algorithms worked out accurate and efficient results. |
