Routing under uncertainty: A comparative study

Multihop mobile wireless networks, characterized by volatile state and limited resources, present a challenging environment for effective and efficient route selection. In such networks, nodes may be restricted in their use of resources for obtaining state information and making routing decisions and often may have to base their decisions on incomplete or out-of-date information. In this paper, we compare six variants of well-known routing algorithms with respect to their communication, computation, and memory demands and the characteristics of the routes they produce. We show that algorithms that probabilistically select routes based upon locally-available state information derived through dynamic programming techniques result in efficient use of transmission capacity and hence hold promise for use in highly-dynamic networks with limited resources