• Formulates MSNS as an NP-complete Minimum Dominating Set problem. • Integrates residual-energy-aware sleep/awake scheduling to balance energy consumption among sensor nodes. • Proposes a hybrid memetic algorithm for optimal minimal sensor node selection. • Enhances network lifetime using energy-aware sleep scheduling and achieves higher coverage efficiency with fewer active nodes. Wireless Sensor Networks (WSNs) are critical for Internet of Things (IoT) applications, enabling data collection through battery-powered sensor nodes.The Minimal Sensor Node Selection (MSNS) problem aims to select the smallest subset of active nodes. MSNS also ensure target coverage while maximizing network lifetime through energy efficient sleep scheduling. This paper formulates MSNS as a Minimum Dominating Set (MDS) problem, proven NP-complete, and proposes an improved hybrid memetic search algorithm for clustered WSNs. To enhance performance over the basic memetic approach, we introduce a refined mathematical model. This strengthens the fitness function and local search, together with a new local search dominance rule that forces replacement of any low energy node with a higher energy neighbor that covers the same or more uncovered targets. These additions make the memetic algorithm much more aggressive at protecting higher energy nodes and produce noticeably better (smaller + long lasting) sets. By spatial correlation and sleep-awake mechanisms, the approach minimizes redundant data transmission and energy consumption. Network simulations on clustered WSNs suggests that the improved protocol promising results over the basic greedy MDS, modern algorithms and standard genetic algorithm. This helps achieving a substantial reduction in the number of active nodes, better approximation ratios (1.12–1.18), higher coverage efficiency (up to 8.8 targets per active node), and significantly extended network lifetime under the simulated conditions. These theoretical improvements stem from the memetic algorithm’s hybrid design. The integration of greedy initialization for strong starting solutions, energy aware local search with dominance rules that preferentially retain high residual energy nodes. The refined fitness function penalizing energy depletion, and sleep awake scheduling exploiting spatial correlation to suppress redundant activations and transmissions. In contrast, the standard genetic algorithm lacks these local refinement and energy prioritization mechanisms, leading to larger active sets, poorer approximations, lower efficiency, and shorter lifetimes. Whereas, the basic greedy MDS provides reasonable but suboptimal heuristics without evolutionary exploration. The results and the findings confirm the effectiveness of the proposed algorithm, showcasing its capacity to provide significant energy savings. And also ensuring scalability and dependable target coverage in real world IoT applications using wireless sensor networks.
Shalu et al. (Sun,) studied this question.