Abstract
In wireless sensor networks (WSNs), prototyping systems facilitate the realization of real node deployment, enabling to test new algorithms, protocols, and networking solutions. This paper investigates the 3D indoor redeployment problem in WSNs by finding the positions where nodes should be added in order to improve an initial deployment while optimizing different objectives. For this purpose, an approach based on a recent evolutionary optimization algorithm (NSGA-III) is used. The latter algorithm is hybridized with a strategy of incorporating of the user preferences (PI-EMO-VF). The major contributions of this work are as follows: testing the NSGA-III efficiency in the case of real world problems, comparing it with another recent many-objective algorithm (MOEA/DD), and incorporating the concept of preferences of users into NSGA-III. The real experiments performed on our testbeds indicate that the results given by the proposed algorithm are better than those given by other recent optimization algorithms such as MOEA/DD.