Autonomic and Coevolutionary Sensor Networking with BiSNET/e

Wireless sensor networks (WSNs) applications are often required to balance the tradeoffs among conflicting operational objectives (e.g., latency and power consumption) and operate at an optimal tradeoff. This chapter proposes and evaluates a biologically-inspired architecture, called BiSNET/e, which allows WSN applications to overcome this issue. BiSNET/e is designed to support three major types of WSN applications: data collection, event detection and hybrid applications. Each application is implemented as a decentralized group of software agents, which is analogous to a bee colony (application) consisting of bees (agents). Agents collect sensor data or detect an event (a significant change in sensor reading) on individual nodes, and carry sensor data to base stations. They perform these data collection and event detection functionalities by sensing their surrounding network conditions and adaptively invoking behaviors such as pheromone emission, reproduction, migration, swarming and death. Each agent has its own behavior policy, as a set of genes, which defines how to invoke its behaviors. BiSNET/e allows agents to evolve their behavior policies (genes) across generations and autonomously adapt their performance to given objectives. Simulation results demonstrate that, in all three types of applications, agents evolve to find optimal tradeoffs among conflicting objectives and adapt to dynamic network conditions such as traffic fluctuations and node failures/additions. Simulation results also illustrate that, in hybrid applications, data collection agents and event detection agents coevolve to augment their adaptability and performance . Department of Computer Science University of Massachusetts, Boston e-mail: {pruet,jxs}@cs.umb.edu 1 2 Boonma and Suzuki 1.