للتجميع في شبكات الاستشعار اللاسلكية تحت نماذج قنوات راديوية واقعية (fss) دراسة متانة خوارزمية البحث بالمجموعة الثابتة

Abstract

This thesis evaluates the robustness of the Fixed-Set Search (FSS-WSN) cluster-head selec tion algorithm in Wireless Sensor Networks under realistic radio channel models. Most prior comparisons rely on Heinzelman's deterministic energy model, which assumes perfect packet delivery — an assumption far from experimental reality. We conduct a systematic compar ison of six algorithms (LEACH, HEED, SEP, PSO, GWO, FSS-WSN) under three channel models: ideal, log-normal shadowing, and log-normal plus Rayleigh fading. The factorial plan covers 720 simulations, analyzed with paired Wilcoxon signed-rank tests, Holm correc tion, and effect-size estimation. Two hypotheses are tested: stability of FSS-WSN's advantage across channels (H0), or its dependence on the channel model (H1). The work contributes methodologically by extending the validation framework of FSS-WSN, empirically by deliv ering the first documented benchmark under realistic channels, and practically by providing a robustness guide for algorithm selection based on the expected deployment environment.