Fuzzy Logic in Evolving in silico Oscillatory Dynamics for Gene Regulatory Networks

Abstract

This chapter investigates empirically the influence of control logic on regulatory dynamics in computational models of genetic regulatory networks. The gene regulatory network motif considered in this work consists of three genes with both positive and negative feedback loops. Two fuzzy logic formulations are studied in this work, one is known as the Zadeh operator, and other is the probabilistic operator. The evolved dynamics of the network motifs is then verified with perturbed initial system states. Our empirical results show that with the probabilistic ‘AND’ operator and the probabilistic ‘OR’ operator, the system is able to evolve sustained oscillation with a low probability. However, sustained oscillation is not evolvable when the Zadeh operator is employed. In addition, we also show that regulatory motifs with the probabilistic operators possess much richer dynamics than that with the Zadeh operators.