Integer Programming Applied to Wireless Sensor Networks Topology Optimization

Wireless Sensor Networks (WSNs) are systems with great potential for applications in the most diverse areas such as industry, security, public health, and agriculture. In general, for a WSN to achieve high performance, multiple criteria must be considered, such as coverage area, connectivity, and energy consumption. In this work an Integer Programming (IP) model to solve a Sensor Allocation Problem (SAP) is presented. The IP model considers a heterogeneous WSN and deterministic locations to positioning of sensors. The proposed model was validated using the IBM ILOG CPLEX solver. A several computational experiments were performed, and an analysis through small and mediumsized instances of the problem under study are presented and discussed. The proposed model presents good results given the problem premises, constraints and considered objectives, achieving 0.0099% optimality gap for the best scenarios where networks are fully connected and are feasible to implement. Other suboptimal evaluated scenarios with denser distribution of sensor nodes depict about 0.04% of isolated node positioning, spite maintaining overall balance between energy consumption and coverage. Therefore, the proposed model shows promise for achieving practical solutions, i.e., those with implementation feasibility in most considered heterogeneous network scenarios.