Discrete Planning of Sensor Networks with Mobile Nodes via Mixed Integer Programming and Simulation Evaluation

This work investigates a discrete problem of planning wireless sensor networks with mobile nodes, motivated by Internet of Things applications in agricultural environments. The goal is to select, from candidate positions, a subset of fixed nodes that preserve multi-hop connectivity between mobile devices and a sinkhole along a discretized horizon. To this end, a formulation by Mixed Linear Integer Programming is proposed, which integrates installation decisions, link activation and time-dependent flow routing. The model is used to generate structurally viable candidate topologies, which are later evaluated with metrics from a simulation campaign in Cooja/Contiki-NG. In the analyzed instance, MILP generated 27 different topologies, allowing us to observe trade-offs between structural density, latency, energy consumption, and volume of data delivered.