Observability Of Boolean Networks Research Articles

Observability of Boolean networks via matrix equations

From the new perspective of logical matrix equations, observability of Boolean networks (BNs) is investigated in this paper. First, it is shown that one BN is locally observable on the set of reachable states if and only if the constructed matrix equations have a unique canonical solution. Then, combining with an equivalence relation, a novel condition is established to verify global observability. Finally, an example is worked out to illustrate the obtained results.

Observability of Boolean Networks With Redundant Channels

In this brief, observability of Boolean networks (BNs) with redundant channels is considered. Each output node with multi-channels is designed to receive the system state information under certain probabilities. By resorting to the semi-tensor product of matrices, the algebraic form of the proposed model is provided. An algorithm to verify the observability of the proposed model is then given by fully considering the randomness properties of the proposed model. In addition, a new observable BN with a set of redundant channels is proposed to improve the reliability of the BNs. Finally, two numerical examples are presented to illustrate the main results.

Observability of Boolean networks via set controllability approach

Controllability and observability of Boolean control networks (BCNs) are two fundamental properties. But verification of the latter is much harder than the former. This paper considers the observability of BCNs via controllability. First, set controllability is proposed, and a necessary and sufficient condition is obtained. Then a technique is developed to convert observability into an equivalent set controllability problem. Using the result for set controllability, a necessary and sufficient condition is also obtained for the observability of BCNs.

Linearization of nonlinear filter generators and its application to cryptanalysis of stream ciphers

Nonlinear filter generators are commonly used as keystream generators in stream ciphers. A nonlinear filter generator utilizes a nonlinear filtering function to combine the outputs of a linear feedback shift register (LFSR) to improve the linear complexity of keystream sequences. However, the LFSR-based stream ciphers are still potentially vulnerable to algebraic attacks that recover the key from some keystream bits. Although the known algebraic attacks only require polynomial time complexity of computations, all have their own constraints. This paper uses the linearization of nonlinear filter generators to cryptanalyze LFSR-based stream ciphers. Such a method works for any nonlinear filter generators. Viewing a nonlinear filter generator as a Boolean network that evolves as an automaton through Boolean functions, we first give its linearization representation. Compared to the linearization representation in Limniotis et al. (2008), this representation requires lower spatial complexity of computations in most cases. Based on the representation, the key recoverability is analyzed via the observability of Boolean networks. An algorithm for key recovery is given as well. Compared to the exhaustive search to recover the key, using this linearization representation requires lower time complexity of computations, though it leads to exponential time complexity.

Observability of Boolean networks: A graph-theoretic approach

Boolean networks (BNs) are discrete-time dynamical systems with Boolean state-variables and outputs. BNs are recently attracting considerable interest as computational models for genetic and cellular networks. We consider the observability of BNs, that is, the possibility of uniquely determining the initial state given a time sequence of outputs. Our main result is that determining whether a BN is observable is NP-hard. This holds for both synchronous and asynchronous BNs. Thus, unless P=NP, there does not exist an algorithm with polynomial time complexity that solves the observability problem. We also give two simple algorithms, with exponential complexity, that solve this problem. Our results are based on combining the algebraic representation of BNs derived by D. Cheng with a graph-theoretic approach. Some of the theoretical results are applied to study the observability of a BN model of the mammalian cell cycle.