AI-assisted load balanced and dominating set–based framework for energy-efficient IoT-WSNs in sustainable smart city applications

Abstract IoT-based Heterogeneous Wireless Sensor Networks (HWSNs) are a vital enabling technology for Sustainable Smart Cities (SSCs), but their heterogeneous energy resources make lifetime-aware management a critical challenge. A practical solution is to partition the network into disjoint Dominating Sets (DSs) and activate only one DS at a time, allowing the remaining nodes to sleep and conserve energy. This paper proposes the AI-Assisted Load-Balanced Dominating-Set-Based Framework (AIALBDSF) for IoT-based HWSN-enabled SSCs. AIALBDSF integrates three main components: ESA-Skyline for energy- and size-attentive DS construction, SLADS-4C for adaptive size- and lifetime-aware DS scheduling, and a chain-based intra-DS aggregation and routing layer. Unlike conventional DS-based approaches that mainly focus on DS construction or static activation policies, AIALBDSF jointly integrates ESA-Skyline-based disjoint DS construction, fuzzy four-state DS scheduling, and communication-aware intra-DS routing within a unified framework. ESA-Skyline generates multiple DS candidates at each iteration and applies the BNL_Skyline filtering procedure to retain only non-dominated solutions with respect to DS lifetime and size, while SLADS-4C employs fuzzy c-means-based four-state characterization to schedule DSs that effectively balance workload among DSs adaptively. After activation, the selected DS performs chain-based aggregation using an energy- and distance-aware leader selection strategy to reduce redundant long-range transmissions, and forwards the aggregated data to the sink. Simulation results demonstrate that AIALBDSF improves workload balance among DSs, reduces premature depletion of fragile sets, and significantly extends HWSN lifetime compared with representative baseline methods, while reducing redundant long-range transmissions and maintaining effective DS-based operation.