Dynamic Resource Allocation for a Sensor Network

Abstract : We describe a project that analyzes the performance of a sensor network trying to detect enemy agents in an urban environment. We simulate a network of units that can detect different enemy types using a variety of sensors. Our work aims to address key issues about networked systems, including the effectiveness of cooperative strategies, the balance of multiple constraints, and the tradeoffs between resource allocation and performance. We make use of evolutionary computing algorithms to search the space of possible configurations, carrying out parameter sweeps to test how various dimensions of the design space impact performance over a range of scenarios. Our results provide some quantitative insights into the performance of networked systems.