Abstract:

Dynamic Bayesian networks are an important tool for the nonlinear dynamical systems with uncertainty inference. A fast forwards-backwards algorithom is proposed by introducing a new computation method into the improved forwards-backwards (IFB) algorithom. The fast forwards algorithm and backwards algorithm are deduced in theory, and the two algorithms are combined to deduce the fast forwards-backwards algorithm. According to the complexity analysis, it’s easy to see that the complexity of the proposed algorithm is lower. It is proved by the simulation experiments that the algorithm is correct and efficient.