Individualization of control devices in ambient intelligence based on evolutionary algorithms

Abstract

Individualization of control devices in ambient intelligence based on evolutionary algorithms

Manucharjan L.H.., Pachev A.N.

Incoming article date: 07.12.2017

Since users can have very different preferences, the personalization of surrounding devices is of paramount importance. Several approaches have been proposed for establishing such personalization in the form of machine learning or more specialized approaches to learning based on scientific knowledge and innovations. Despite great advances in optimization, evolutionary algorithms in this context have been little studied, mainly because they are known as elements that are slow to learn. Anyway, at present there are quite fast optimizers based on evolutionary algorithms. In this article, an analysis is made of the suitability of evolutionary algorithms for "ambient intelligence".

Keywords: ambient intelligence, evolutionary algorithms, personalization, optimizer, CMA-ES, user, controller, machine Learning, sensor