Structure-variable Hybrid Dynamic Bayesian Networks and its inference algorithm

Abstract

Structure-variable Discrete Dynamic Bayesian Networks can model under the situation n of the process of mutation and the change of discrete network structure and parameters, but can't model and reason the system containing both continuous variables and discrete variables. Focusing on this question the concept of Structure-variable Hybrid Dynamic Bayesian Networks is proposed, introduce the algorithm of the discrete forward-backward into Discrete Dynamic Bayesian Network inference to get inference algorithm of more hidden nodes in Structure-variable Discrete Dynamic Bayesian Networks, and combine with the forward-backward reasoning algorithm of continuous variables, propose inference algorithm for Structure-variable Hybrid Dynamic Bayesian Networks. The simulation background is the assessment of UAV under dynamic environment situation and validates the algorithm .The results of simulation demonstrates the validity of the proposed concepts and reasoning algorithm for Variable Structure Hybrid Dynamic Bayesian Networks.