Knowledge sharing agents using genetic algorithms in mobile ad hoc networks

We present a novel approach for knowledge sharing mobile agents to achieve a uniform spread of autonomous mobile nodes over an unknown geographical terrain. In this research, knowledge sharing agents adjust their speed and directions within a mobile ad hoc network (manet) using a decentralized topology control mechanism based on genetic algorithms (GAs). The genetic information that each mobile agent exchanges with other neighboring agents within its communication range includes the node's location, speed, and movement direction.;We show that our GA-based topology control system can be modeled as a dynamical system in order to provide formalism to study its convergence trajectory in the space of possible populations. We apply this discrete time dynamical system model for calculating the cumulative effects of genetic operators such as selection, mutation, and crossover as a population of possible solutions evolves through generations using our GA-based approach.;To demonstrate applicability of our topology control algorithms to real-life problems and evaluate their effectiveness, we implemented a simulation software system and several different testbed platforms. The simulation and testbed experiment results indicate that, for important performance metrics such as normalized area coverage (NAC) and convergence rate, our algorithms can be effective in deploying nodes under restrained communication conditions in manets operating without prior geographical terrain knowledge. Since our topology control algorithms adapt to any local environment rapidly and do not require global network knowledge, they can be used as real-time topology controllers for realistic military and civilian applications.