Parallel simulation uses multiple processors to reduce the execution time of a
single simulation model. This technology enables off-line decision aide
simulations to be transformed into real-time decision aides for time critical
situations without loss of model accuracy.
As a case study of applying parallel simulation technology to critical battle
management applications, we have undertaken the
parallelization of an existing large-scale Ada simulator called TISES, and
documented our experiences. TISES
stands for THAAD Integrated System Effectiveness Simulation and models all
activities performed by a collection of THAAD missile batteries during an
engagement scenario. TISES is approximately 300,000 lines of source code.
Because of its enormous size and complexity, parallelization of the entire
system was beyond the scope of our proof-of-concept effort. Instead, we
selected a sub-component of the simulation that performs threat evaluation and
Our approach for the parallelization of this sub-component makes use of an
optimistic synchronization protocol, called Time Warp. Using our highly
optimized Time Warp simulation executive, called Georgia Tech Time Warp (GTW)
we obtain speedups on the order of 12 on 16 processors
Using the knowledge gained from this effort and our parallelization approach,
we believe the entire TISES system could be parallelized using Time Warp, as
well as other missile defense applications.
Also see our reference
related to this research.