Identification of Boolean Control Networks

Abstract

Identification is essential for the application of theoretical results on Boolean control networks to practical problems because knowing the structure of a network is the first priority when solving any related problems. Particularly, when a large or, potentially, huge network (e.g., a genetic regulatory network) is considered, we need to identify it or approximate it by using a small network.

In this chapter necessary and sufficient conditions are first obtained for the identifiability of state equations from input-state data. An alternative necessary and sufficient condition for a controllable Boolean network to be observable is then presented. Based on these two results, necessary and sufficient conditions for identifiability from input–output data are obtained, and some particular cases are considered. To identify practical models, a numerical algorithm is proposed. This chapter is based on Cheng and Zhao (Automatica, 2010).