Markovian agent modeling swarm intelligence algorithms in wireless sensor networks

Wireless Sensor Networks (WSN) are large networks of tiny sensor nodes that are usually randomly distributed over a geographical region. The network topology may vary in time in an unpredictable manner due to many different causes. For example, in order to reduce power consumption, battery operated sensors undergo cycles of sleeping–active periods; additionally, sensors may be located in hostile environments increasing their likelihood of failure; furthermore, data might also be collected from a range of sources at different times. For this reason multi-hop routing algorithms used to route messages from a sensor node to a sink should be rapidly adaptable to the changing topology. Swarm intelligence has been proposed for this purpose, since it allows the emergence of a single global behavior from the interaction of many simple local agents. Swarm intelligent routing has been traditionally studied by resorting to simulation. The present paper aims to show that the recently proposed modeling technique, known as Markovian Agent Model (MAM), is suited for implementing swarm intelligent algorithms for large networks of interacting sensors. Various experimental results and quantitative performance indices are evaluated to support this claim. The validity of this approach is given a further proof by comparing the results with those obtained by using a WSN discrete event simulator.