Non-Uniform Transmission Scheduling Under Jamming in a Mobile Ad Hoc Network Model: An ns-3 Simulation Study

Paper 17 in The Geometry of the Critical Line programme. We test whether non-uniform spatial phase offsets applied to packet transmission scheduling across nodes of a mobile ad hoc network (MANET) reduce vulnerability to radio-frequency jamming. The hypothesis is motivated by the spectral-smearing mechanism identified in forced Kuramoto networks (Paper 11), where aperiodic spatial phase offsets suppress synchronisation lock-in by 80–100 percentage points. Using the ns-3 network simulator, we model a 50-node MANET under three threat models (barrage, sweep, and follower jamming) with six scheduling patterns (uniform, sinusoidal, prime-indexed, golden-ratio LDS, van der Corput, and random-fixed) across 7 jammer power levels and 100 independent seeds per condition (12,600 total runs). The network operates in a severe congestion regime: baseline packet error rates exceed 99% due to MAC-layer broadcast contention. Among equal-load static patterns, all non-uniform schedules outperform the coherent uniform baseline in every condition tested (Welch t > 10, p ≪ 0.001), delivering 1.37–1.57× more packets despite an absolute PER improvement of only 0.22 percentage points. Fixed-random scheduling is numerically strongest; deterministic low-discrepancy sequences offer no advantage over random. Exploratory follower-jammer diagnostics show increased tracking difficulty under non-uniform schedules. The results partially support the spectral-smearing hypothesis: the directional effect is confirmed but the practical impact is severely bounded by protocol-layer saturation. Together with Paper 14's negative Kuramoto–Sivashinsky result, this draws the second weight limit of the programme. Archive includes LaTeX source and complete simulation data (12,600 rows).