Optimizing a Sensor Deployment with Network Constraints Computable by Costly Requests
Résumé
This paper deals with the planning of networked sensors. In such sensors deployment, two main aspects have to be considered: cover most strategic areas, and ensure the connectivity between sensors and network nodes. The connectivity is computed by expensive simulations, and these computation requests should be optimized as an experiment design in order to enhance the constraint models, while the sensor deployment itself is solved by usual optimizer. We propose a simulated approach, inspired from the known algorithm, EGO, in order to address this two-level problem.