Solving a multi-objective heterogeneous sensor network location problem with genetic algorithm

In this paper, we consider a multi-purpose two-level location problem introduced by Karatas (2020) to improve the coverage performance of heterogeneous sensor networks. The problem basically seeks to determine the best location scheme of sensors of different types and characteristics in a belt-shaped boundary area with the purpose of providing a sufficient level of field, point and barrier coverage against different types of intruders. To solve the problem, first, the Mixed Integer Linear Programming (MILP) model developed by Karatas (2020) is used together with a commercial solver and a Non-Dominated Sorting Genetic Algorithm-II (NSGA-II), adapted to solve especially large-sized instances of the problem, is applied. Next, we compare the NSGA-II heuristic with the MILP solved via a commercial exact solver on a number of test instances. The experiment results suggest that the heuristic algorithm can produce a large number of diverse and high-quality solutions in very short computation times in comparison to the exact solver.

Automated summary

This paper considers a multi-purpose two-level location problem aiming to enhance the coverage performance of heterogeneous sensor networks by determining the best location scheme of sensors in a belt-shaped boundary area. The study finds that a heuristic algorithm can generate diverse and high-quality solutions in a short computational time compared to an exact solver.