Simulating asynchronous, decentralized military command and control

The pace and scope of modern warfare have made some aspects of traditional, centralized decision making obsolete. A decentralized scheme grants greater autonomy to fighting units, which process data locally for efficient decision making. In experimental simulations using a decentralized algorithm, units react to information faster in both offensive and defensive scenarios. We used asynchronous, distributed, discrete event simulation techniques to model the command and control problem. These algorithms are appropriate for problems involving discrete data transfer, geographically dispersed decision making entities, asynchronously generated stimuli, and data feedback. To compare our algorithm with traditional algorithms and identify scenarios for which it is effective, we used a loosely coupled parallel processor as a simulation testbed. To the best of our knowledge, this research is the first attempt to scientifically model decentralized C/sup 3/ and assess its performance through extensive parallel simulation. State-of-the-art battlefield simulators such as Simnet and CATT provide training environments in which human operators make the decisions. In contrast, our testbed provides an automated environment for fast, efficient, accurate warfare modeling. By "accurate" we mean spatial and timing resolution; we do not claim to represent all battlefield details, such as terrain.