Maintaining Coherence among Entities’ States in a Distributed Multi-Agent System ii

For multi-agent systems to interact meaningfully in a distributed environment, the coherence among the entities’ states must be maintained. Because continuous state updates normally require large amounts of network bandwidth, Newtonian-based equations (i.e., dead-reckoning models) are frequently used to reduce the number of communications updates. However, even with the use of such dead-reckoning models, networking and communications limitations still exist in currently fielded systems. A more effective approach to reducing the communications requirements is achieved by replacing the dead-reckoning models with neural networks. This paper presents the background and motivation for the research, the architecture and training algorithms of the networks, and the integration of the networks into a large-scale simulation environment. Quantitative measures from the experiments reveal that the use of neural networks can significantly reduce the number of communication updates required to maintain entity-state coherence. However, the neural networks are also more difficult to scale than the currently used dead-reckoning algorithms.