Optimization of sensor selection problem in IoT systems using opposition-based learning in many-objective evolutionary algorithms

In the Internet of Things (IoT) systems, physical objects are connected to each other through sensor devices that serve multiple functionalities. The sensor selection is known to be an NP-hard problem. Thus, Evolutionary Algorithms (EAs) can be incorporated to solve the sensor selection problem in IoT systems. Previously, researchers have been working on sensor selection problems with two or three objectives. Recently, this problem is formulated as a many-objective optimization problem and solved using a Decomposition-based Many-Objective Evolutionary Algorithm (MOEA/D). In this paper, we consider the sensor selection problem as a many-objective problem with 5 objectives. To accelerate the convergence, we incorporate Opposition Based Learning (OBL) in the general framework of MOEA/D. Furthermore, we use a well-known many-objective algorithm known as the Non-dominated Sorting based Genetic Algorithm incorporated with OBL (NSGA-III/OBL) to enhance its convergence and diversity. The experimental results show that NSGA-III/OBL outperforms all other compared algorithms.

Automated summary

This paper studies the sensor selection problem in IoT systems, treating it as a multi-objective optimization problem with 5 objectives. A Decomposition-based Many-Objective Evolutionary Algorithm and Opposition Based Learning are incorporated to solve the problem. Experimental results show that the proposed algorithm outperforms other compared algorithms.