Agent-based Sensor-Mission Assignment for Tasks Sharing Assets

A sensor network may be required to support multiple mission to be accomplished simultaneously. Furthermore, the environment may change at any time; i.e. a new mission may arrive at any time. In solving this many-mission, many-sensor problem in dynamic environments, conflicts between missions may occur for the use of sensor resources. A mechanism to match sensor resources to mission demands thus becomes necessary. In this paper, motivated by the conservation of resources, we consider the problem of sensor-mission assignment, in which sensors may be shared and reassigned between tasks. To achieve this, sensors are represented by agents, which coordinate to establish virtual organizations to meet mission requirements. The agent coordinating the
achievement of a mission utilises a novel multi-round, Knapsack based algorithm, GAP-E, to allocate sensor agents to tasks based on bids received. Through simulations, we empirically demonstrate that this model provides a significant improvement in the number of completed missions as well as execution time.