This project intends to characterize, design, and evaluate metaheuristic techniques able to solve real world problems, and then, particular applications of a communication network of vehicles (VANET). Our hypothesis is that standard metaheuristics are not able by themselves to address the tight requirements of many complex problems like this one, since we are dealing with execution times of a few seconds, specific user constraints, scalable to very big dimension problems, and robust to work in different scenarios.
The goals of the project are to design and exploit a communication network between cars by developing efficient distributed, dynamic, self-adaptive, and hybrid metaheuristics. To this end, we will develop theory (convergence and elementary landscapes), define a methodological approach for designing metaheuristics (statistically sound, software and hardware centric, explicitly demanding tight levels of performance, comparing/beating the best state of the art in academic benchmarks), and put all them to work in several problems, in special that of the hot topic of VANETs.
This project has foreseeable transfers to industry (EPO attractive) and focuses on multidisciplinary advances from a quantified and verifiable point of view, in applications of interest for the Society and for Computer Science (TIN). The proposal is expected to endorse the visibility of Spain in improving the existing state of the art in this field of intelligent systems, as well as to extend the application field of metaheuristics. We will tackle therefore a mission for internationalization, with frequent visits to/from foreign teams, numerous high-impact publications, research staff training, and also directly involving world-class international researchers in our team like Prof. Darrell Whitley and Prof. Xin Yao.