Optimal deployment of heterogeneous sensor networks for a hybrid point and barrier coverage application

Abstract

Heterogeneous sensor networks (HSNs) comprise of different types of sensors with varying range and sensing quality. Coverage is one of the most important performance metrics for sensor networks (SNs), and it reflects how well an area of interest is monitored. Compared with homogeneous SNs, design and control of HSNs for coverage purposes are more complicated. This study considers a hybrid point and barrier coverage application for HSNs in which a decision-maker seeks to locate sensors along a two-dimensional belt-shaped barrier region for two purposes: (1) protecting critical facilities located inside the barrier region, and (2) preventing illegal border crossings. Using this multi-objective setting, this study incorporates the following factors: multiple types of sensors, intruders and critical facilities, diffuse sensor coverage model, cooperative cover allowance, and technical constraints.To solve the defined problem, we first develop a multi-objective integer non-linear program (INLP) formulation. Since the INLP is non-convex it may not produce globally optimal solutions. For this purpose, using a special mapping technique, we reformulate the problem as a multi-objective integer linear program (ILP). Then, we employ a genetic algorithm (GA) metaheuristic to solve the problem. We then perform extensive simulation runs to measure and compare the performances of the proposed INLP, ILP and GA solution approaches. Our results indicate although the ILP is efficient for small size problems, it requires longer computing times to deliver globally optimal solutions. The INLP and GA, on the other hand, provide a balance between the solution quality and computation time for larger problem instances. The performance of the proposed hybrid modelling approach is demonstrated through sensitivity analysis runs.