Cactus-Expandable Graphs

Cactus graphs, which are connected graphs with no edge contained in two or more different simple cycles, are known to provide useful properties for network systems. Meanwhile, it has been found that a noncactus graph can be equivalently transformed into a cactus graph by using a conventional technique for simplifying block diagrams. Based on such graph transformation, we can transform a network system on a general graph (where nodes have deterministic dynamics with a multidimensional state variable) into a network system on a cactus graph while preserving the original dynamics, which allows us to apply existing results for network systems on a cactus graph. However, the condition under which such transformation is possible has never been clarified so far. This paper addresses the problem of finding graphs which can be transformed into a cactus graph. We first introduce the notion of cactus expandability, which corresponds to the existence of a cactus graph equivalent to the original graph. A sufficient condition is presented based on two graph characteristics, called the nucleus graphs and doubly bidirectionally connected pairs. Furthermore, we derive a sufficient condition for the interconnection of graphs. Based on these results, we also present a method to obtain a cactus graph equivalent to the original graph and discuss its computational complexity. We finally apply the results to the robustness analysis of biological networks to demonstrate the potential for systems biology.

[1]  S. Davis,et al.  Mathematical modeling of an oscillating gene circuit to unravel the circadian clock network of Arabidopsis thaliana , 2013, Front. Plant Sci..

[2]  Ching-tai Lin Structural controllability , 1974 .

[3]  S. Kauffman Homeostasis and Differentiation in Random Genetic Control Networks , 1969, Nature.

[4]  Murat Arcak,et al.  Diagonal Stability on Cactus Graphs and Application to Network Stability Analysis , 2011, IEEE Transactions on Automatic Control.

[5]  Shun-ichi Azuma,et al.  A sufficient condition for cactus-expandability of graphs , 2017, 2017 IEEE 56th Annual Conference on Decision and Control (CDC).

[6]  Michael K. Ng,et al.  Algorithms for Finding Small Attractors in Boolean Networks , 2007, EURASIP J. Bioinform. Syst. Biol..

[7]  A. Goldberger,et al.  Emergence of complex dynamics in a simple model of signaling networks. , 2004, Proceedings of the National Academy of Sciences of the United States of America.

[8]  L. Chua,et al.  Topological criteria for nonlinear resistive circuits containing controlled sources to have a unique solution , 1984 .

[9]  Shun-ichi Azuma,et al.  Structural bistability analysis of Boolean networks , 2016, 2016 IEEE 55th Conference on Decision and Control (CDC).

[10]  Paul E. Brown,et al.  Extension of a genetic network model by iterative experimentation and mathematical analysis , 2005, Molecular systems biology.

[11]  Mohsen Zamani,et al.  Structural controllability of multi-agent systems , 2009, 2009 American Control Conference.

[12]  Shun-ichi Azuma,et al.  Structural oscillatority analysis of Boolean networks , 2015, 2015 54th IEEE Conference on Decision and Control (CDC).

[13]  Shun-ichi Azuma,et al.  Structural Monostability of Activation-Inhibition Boolean Networks , 2017, IEEE Transactions on Control of Network Systems.

[14]  Gilbert Labelle,et al.  Enumeration of m-Ary Cacti , 1998, Adv. Appl. Math..