Aas 10-225 Efficient Parallelization of Nonlinear Perturbation Algorithms for Orbit Prediction with Applications to Asteroid Deflection and Fragmentation

Orbit determination and control problems are sometimes solved using linearized methods, even though nonlinear effects can sometimes dominate the system. This paper addresses ways to parallelize computational algorithms for numerical integrators and nonlinear relative motion equations. An application of the proposed methods to mutual gravitation of a fragmented asteroid is presented, with analysis of the relative efficiency of parallel architechtures. Hydrodynamic simulation of a subsurface nuclear explosion is used to generate the initial conditions, after which the fragmented system is propagated in parallel including mutual gravitational terms. Efficient algorithms for this problem emphasize the achievement of verifiable results for asteroid deflection/fragmentation research using limited or budget-allocated computer resources.