A NUMERICAL STUDY OF LINK AND PATH DURATIONS IN MOBILE AD HOC NETWORKS

A theoretical analysis has shown that under a set of assumptions, the distri-bution of path duration can be well approximated by an exponential distribution when the path hop count is sufficiently large. The goal of this thesis is two folds: Using NS-2 simulations to (i) Investigate how fast the path distributional conver-gence takes place, and how quickly the inverse of the expected duration of a path converges to the sum of the inverses of the expected durations of the links along the path, and (ii) Validate the conditions under which the distributional convergence is established. Simulations are run under four different mobility models and two different on-demand routing protocols. Simulation results show that the path duration distri-bution can be accurately approximated by an exponential distribution for path hop count larger than 5 or 6 with fitting error less than 0.05 using Kolmogorov-Smirnov test (K-S test) for all considered scenarios. However, the ratio of the inverse of the expected path duration to the sum of the inverses of the expected link durations along the path does not get close to one for path hop count less than 12 in the case