Hybrid multi-objective evolutionary algorithms based on decomposition for wireless sensor network coverage optimization

Ying Xu,Ou Ding,Rong Qu,Keqin Li

doi:10.1016/j.asoc.2018.03.053

Abstract

In Wireless Sensor Networks (WSN), maintaining a high coverage and extending the network lifetime are two conflicting crucial issues considered by real world service providers. In this paper, we consider the coverage optimization problem in WSN with three objectives to strike the balance between network lifetime and coverage. These include minimizing the energy consumption, maximizing the coverage rate and maximizing the equilibrium of energy consumption. Two improved hybrid multi-objective evolutionary algorithms, namely Hybrid-MOEA/D-I and Hybrid-MOEA/D-II, have been proposed. Based on the well-known multi-objective evolutionary algorithm based on decomposition (MOEA/D), Hybrid-MOEA/D-I hybrids a genetic algorithm and a differential evolutionary algorithm to effectively optimize sub-problems of the multi-objective optimization problem in WSN. By integrating a discrete particle swarm algorithm, we further enhance solutions generated by Hybrid-MOEA/D-I in a new Hybrid-MOEA/D-II algorithm. Simulation results show that the proposed Hybrid-MOEA/D-I and Hybrid-MOEA/D-II algorithms have a significant better performance compared with existing algorithms in the literature in terms of all the objectives concerned.

Highlights

Wireless Sensor Networks (WSNs) are self-organized networks consisting of sensor nodes capable of sensing, processing and wireless communication
We proposed a hybrid MOEA/D algorithm based on the work in [9], namely Hybrid-MOEA/D-I, by combining Genetic Algorithm (GA) and Differential Evolution (DE) as the mixed reproduction operator to optimize each sub-problem
Better non-dominated solution set (NDS) sets have been obtained by Hybrid-MOEA/D-II compared with Hybrid-MOEA/D-I for four different sizes of WSN, which demonstrate the effectiveness of the discrete binary particle swarm optimization (DPSO) enhancement strategy

Summary

INTRODUCTION

Wireless Sensor Networks (WSNs) are self-organized networks consisting of sensor nodes capable of sensing, processing and wireless communication. Experimental results demonstrated that both the simulated annealing based strategies and the genetic local search can efficiently identify high quality non-dominated solution sets for the problems and outperform other conventional multi-objective evolutionary algorithms. To further enhance the search ability of Hybrid-MOEA/D-I and preserve high quality individuals in each generation, a new Hybrid-MOEA/D-II algorithm is devised to integrate an improved discrete binary particle swarm optimization algorithm in [15] as the enhancement strategy to obtain a better Pareto solution set. The coverage field of si at position (xi, yi) is indicated by the green circle area with a radius of ri

The Network Model of WSN

RELATED WORK FOR THE MULTI-OBJECTIVE COVERAGE OPTIMIZATION PROBLEM

I I TABLE II THE RELATIONSHIP OF

An Illustrative Example of the Proposed Hybrid Algorithms

The Time Complexity Analysis

SIMULATION RESULTS

Performance Evaluation of Different Algorithms

The Comparison of Total Number Targets Detected

The Comparison of the Remaining Energy

The Comparison of Non-dominated Solution Sets

The Comparison of Set Coverage Metric

CONCLUSION