Binary search algorithm to predict the time of next link change in wireless mobile networks

We propose a binary search algorithm of logarithmic time complexity to predict the time for next link change in a wireless mobile network (WMN). The search space for the algorithm ranges from the current time instant (initial left index) to the last time instant of the simulation (initial right index). The algorithm always maintains an invariant that the WMN has not undergone any link change at the time instant corresponding to the left index and has undergone at least one link change at the time instant corresponding to the right index. The algorithm goes through a sequence of iterations (the search space reduces by half in each iteration) until the difference between the right index and left index is within a termination threshold (e). We also show that the predicted times of next link changes follow an exponential distribution, especially for lower values of the termination threshold.