Genetic Algorithm Based Congestion Aware Routing Protocol (GA-CARP) for Mobile Ad Hoc Networks

This paper represents a genetic algorithmic approach to the congestion aware routing problem in Mobile Ad hoc Networks. Variable-length chromosomes (strings) and their genes (parameters) are the sources for encoding the problem. The crossover operation exchanges partial chromosomes (partial routes) and the mutation operation maintains the genetic diversity of the population. The proposed congestion aware routing fitness function algorithm is capable of curing all the infeasible chromosomes with an adaptive repair function. The congestion aware fitness function gives an improved quality of solution and enhanced rate of convergence. The performance metrics throughput, packet delivery ratio and delay are taken into account for computer simulations which shows the proposed algorithm exhibits a much better quality of solution (congestion aware routing) and a much higher rate of convergence than other conventional algorithms in mobile ad hoc networks.