Researching Adaptive Systems with RRobots

This research effort proposes a framework for conducting simulation-based research experiments using a pre-existing system. RRobots is a simulation of robots competing within a two-dimensional battlefield. This RRobots simulation has been enhanced to support adaptive-systems research efforts probability based behavior selection and self-healing systems. It supports experiment automation. Researchers have used this as a framework to explore the speed of adaptation. New reporting data elements were added to the output system. The team added tools to support the execution of large numbers of simulation runs, distributing these runs within federation of RRobots Ruby on Rails servers. The initial generation of experiment results using a large data set is analyzed. Next-generation job processing components within the RRobots code are also discussed. This paper discusses these improvements from the general RRobots baseline and uses these to compare a faster more reliable RRobots code base with the original simulation.