Analyzing non-equilibrium statistical behaviors within the OLSR MANET routing protocol

Optimized Link State Routing (OLSR) is a well known proactive MANET routing protocol. Much of the existing OLSR research has been based on packet-level simulation studies in which the results are reported as the averages computed across a conducted set of Monte Carlo runs. This validity of such averaging rests on the presumption that the OLSR behaves, in terms of its network-level quality of service (QoS) measures, as an equilibrium stochastic process, (i.e., a process that converges towards a unique statistical steady state). This work empirically explores the degree to which this presumption holds for OLSR. The work shows that the degree to which OLSR exhibits non-equilibrium behaviors can be clearly influenced by adjusting its run-time tuneable parameters, (e.g., hello interval and topology control interval).