Energy consumption is the result of interactions between hardware, software, users, and the application environment. Optimization of energy consumption has become crucial, energy metric is considered a critical metric, so it is important knowing how to measure and understand how energy is consumed on mobile devices. Accurate knowledge will allow us to propose different solutions to reduce energy consumption in order to improve the user experience. The main objective of this thesis is to model the energy consumption of a particular application running on a mobile device. In this manuscript, we propose a model of energy behavior monitoring, we also describe a methodology to identify model parameters. To this end, we analyzed a collection of experimental data collected during a “Tour de France” in a wheelchair. Then we applied statistical tools to obtain the parameters of the model. We finally validate the model by comparing results against other experimental data. In order to create models based on concrete cases, three case studies have been developed and presented in this thesis. — The first case study compares the evolution of the energy cost in mobile environments as well as the energy behavior of the different components of smart phones based on several energy models. — The second case study deals with the evaluation and measurement of the energy cost consumed, as well as the problems encountered in the methods used for the evaluation of energy consumption. For a better evaluation, we have introduced the study of energy behavior using virtual machines. — The third case study is based on the treatment of the results of the measurements obtained during my tour of France in a connected electric wheelchair. The objective is to anticipate resources by making real measurements and then monitor energy behavior in a real and diversified environment. The model can be used to define an optimal frequency in terms of energy consumption for precise situations without degrading too much the quality of service desired by the user. Collection of scientific data was carried out on the basis of several states on different places by fixing metrics previ-ously. These measures comparisons were carried out in a real environment and confronted with the consumption measures carried out beforehand in a controlled environment.
Directeurs de thèse : M Pierre BOULET / M Mohammed BERRAJAA Rapporteurs : Mme Fouzia. BENABBOU / M Yacine. BELLIK Examinateurs : M Giuseppe LIPARI / M Mostafa AZIZI
Thesis of the team Émeraude defended on 21/06/2018