TY  - EJOU
AU  - Houssein, Essam H. 
AU  - Ibrahim,  E. 
AU  - Wazery, Yaser M. 
AU  - Samee, Nagwan Abdel 
AU  - Emam, Marwa M. 

TI  - Energy-Efficient Optimization in Wireless Sensor Networks: A Systematic Review of Metaheuristic and AI-Based Approaches, Taxonomy, and Research Gaps
T2  - Computers, Materials \& Continua

PY  - 
VL  - 
IS  - 
SN  - 1546-2226

AB  - Wireless Sensor Networks (WSNs) remain constrained by limited battery capacity, which restricts their operational lifetime in Internet of Things (IoT) applications. We present a PRISMA 2020-guided systematic review of 89 primary studies published during 2016–2026, focusing on metaheuristic, hybrid, and Artificial Intelligence (AI)-based methods for energy-efficient clustering and routing. A five-database search strategy yielded 876 records, followed by duplicate removal, title/abstract screening, full-text eligibility assessment, and structured quality appraisal. The synthesis indicates that Ant Colony Optimization (ACO) and Particle Swarm Optimization (PSO) remain the most frequently studied individual algorithms, each appearing in 23.6% of the 89 reviewed studies (SI as a family accounts for 33.7% of all studies), whereas recent work increasingly emphasizes hybrid optimization and learning-based adaptive routing. Among the reviewed studies, Deep Reinforcement Learning with Graph Neural Networks (DRL-GNN) reports up to 4100 operational rounds compared with 1200 rounds for the Low-Energy Adaptive Clustering Hierarchy (LEACH) under reported simulation settings. However, these comparisons should be interpreted cautiously because the benchmark relies on cross-study normalization across heterogeneous simulation assumptions, hardware settings, and reporting practices, and most primary studies do not provide variance estimates, confidence intervals, or effect-size analyses. The evidence base also remains predominantly simulation-driven, which limits direct inference about real-world deployment performance. Within these constraints, the review contributes an integrated taxonomy spanning classical, hybrid, and AI-driven approaches, a harmonized comparative benchmark across widely reported WSN performance metrics, a domain-oriented analysis of application trends, and a structured discussion of research gaps related to security, scalability, hardware feasibility, and deployment realism.
KW  - Wireless sensor networks; energy-efficient routing; clustering; metaheuristic optimization; artificial intelligence; deep reinforcement learning; Internet of Things; PRISMA 2020

DO  - 10.32604/cmc.2026.086392