Dynamic Visualization of Battle Simulations
Paul R. Cohen, James Davis, John L. Warwick · 2000
We present a case study of visualization in understanding encounters between multiple agents in an adversarial environment. The information visualized consists of time series of attributes and relations such as mass, velocity and distance, which we preprocess with a Bayesian clustering algorithm. We differentiate between the encounters based on their outcomes, and generate two and three-dimensional maps that can be used to determine good courses of action from different points in the agents' environments. Keywords: Bayesian Clustering, Dynamic Maps 1. INTRODUCTION This is a case study of visualizations of encounters between adversarial agents in simulated war games. We visualize time series data which captures the dynamics of the encounters. As a preprocessing step, we use a Bayesian clustering algorithm to partition large numbers of encounters of data into clusters which share important dynamical similarities. Our visualizations allow an observer to analyze past events leading to t...