A self-optimization path-finding for geographic forwarding to avoid dead-end in Mobile Ad Hoc Networks

The term Organic Computing is that the system you used having the Self-X properties and autonomously. Self-organization and self-configuration will be automatically performing some mechanisms after system constructed. Self-healing when the mal-function operates or damaged the system. Self-optimization when system found a way that could be better than it is now. And context-aware is that when the working environment has changed, system would adjust itself. We simply used the conception of self-optimization for geographic forwarding in Mobile Ad-hoc Networks (MANETs). The geographic forwarding is a position-based routing that use positioning device like GPS, and try forwarding data packets without knowing the network topology or maintain routing tables. There is an instinctive problem called dead end that occurs when the current relay node cannot forwarding data to destination directly. Once you encountered this situation, an addition overhead must be paid for route to other path. Here we proposed a method that we first divided the MANETs into several squares, and each node in the square will elect their agent to be a path finder. After that we use a special self-optimization algorithm called ldquoA* algorithmrdquo to find a cursory path before using geographic forwarding. This algorithm simply combined the advantages of Dijkstra´s algorithm and best-first search. Source node could perform geographic forwarding and no longer encounter the dead end situation according to use the heuristic function to estimate a cursory path.