An Assignment Based Approach to Distributed Resource Management

Numerica proposes the development of a distributed communications and sensor resource management system. Sensor resource management (SRM) is performed at one level and then communications resource management (CRM) at a finer level with feedback to the SRM. This architecture accommodates the much faster time frame over which CRM operates. Our approach consists of: a) dynamic myopic and non-myopic sensor tasking based on a resource constrained assignment of a group of sensors to each task; b) adaptive myopic and non-myopic communications resource allocation across a network with a feedback to the sensor tasking algorithms; c) a stochastic dynamic optimization framework called partially observable Markov decision processes (POMDP); d) distributed algorithms for a) - c) consistent with the needs of the peer-to-peer network and MANETS (Mobile Ad-hoc NETworks). The philosophy is that communications and sensor resource management should not be considered independently of one another in that a sensor should not be asked to send information that cannot be transmitted because of network constraints, nor should a sensor be asked to perform task that it cannot achieve just because of network bandwidth availability.