SINK RELOCATION IN MULTI-HOP WIRELESS SENSOR NETWORKS: A STATISTICALLY VALIDATED BENCHMARK AND ENERGY–LONGEVITY TRADE-OFF ANALYSIS

Wireless Sensor Networks (WSNs) are known to exhibit an unfair distribution of energy consumption around the sink, where nodes close to the sink carry the load of the rest of the network, and thus have their batteries exhausted earlier than others — the “hotspot” effect. The relocation of the sink is a broadly suggested solution; however, previous studies have tended to analyse their proposed relocation schemes either independently, with different simulation settings, and from the results of a single simulation run, making it hard to differentiate between the actual relocation scheme and the simulation setting to gain insight into the effectiveness of the relocation logic itself. We contribute to that gap directly in this paper, as we run four different strategies of sink relocation on one identical MATLAB simulation environment that has the same multi-hop relay routing model and radio energy cost, and validate all results for 30 independent random-seed simulations, reporting their mean ± standard deviation, and comparing the four strategies by two-sample significance testing. We also give a formal computational complexity analysis, demonstrating that all four strategies have the same asymptotic complexity class, O(R•N). Unlike common belief, simple periodic recentering (Center of Active Nodes) and even a static sink outperform any set of tested configurations of Energy Weighted Relocation (EWR) (over a systematic search of the EWR weighting exponents, smoothing factor, and relocation frequency) on total network residual energy. We do, however, find that there is a real but previously unquantifiable trade-off: configurations that emphasize proximity to nodes, increase the number of surviving nodes by up to 13%, while simultaneously decreasing the total residual energy in the network, which means that the weighted-centroid relocation redistributes, but does not save, network energy. We argue that residual-energy- and distance-based weighting do not address the key factor behind the hotspot effect under multi-hop relay routing, which is per-node relay load, and propose a load-aware relocation rule as a concrete step for future research, backed by evidence.