Asymptotic Observability of Boolean Networks With Markovian Data Loss
研究了实际网络系统中因随机干扰导致的数据丢失对布尔网络可观测性的影响，提出了一种基于马尔可夫链的模型，并给出了渐近可观测性的充要条件，降低了计算复杂度。
Interactions in practical networked systems are subject to temporally correlated data loss caused by stochastic disturbances. In contrast to conventional models that assume independent Bernoulli data loss, this article investigates the asymptotic observability of Boolean networks (BNs) under Markovian loss dynamics. Within a newly developed algebraic framework, the evolution of the data loss switching signal is modeled as a Markov chain, and a column similarity matrix is constructed to characterize the set of states that are distinguishable with probability one in finite time. By establishing an equivalence between the asymptotic reachability of the distinguishable state set and the asymptotic observability of the system, necessary and sufficient conditions for asymptotic observability are derived with reduced computational complexity. Practical numerical examples are provided to illustrate the proposed results.