An optimization framework of biological dynamical systems |
| |
Authors: | Horie Ryota |
| |
Affiliation: | Laboratory for Language Development, RIKEN Brain Science Institute, 2-1 Hirosawa, Wako, Saitama 351-0198 Japan |
| |
Abstract: | Different biological dynamics are often described by different mathematical equations. On the other hand, some mathematical models describe many biological dynamics universally. Here, we focus on three biological dynamics: the Lotka-Volterra equation, the Hopfield neural networks, and the replicator equation. We describe these three dynamical models using a single optimization framework, which is constructed with employing the Riemannian geometry. Then, we show that the optimization structures of these dynamics are identical, and the differences among the three dynamics are only in the constraints of the optimization. From this perspective, we discuss the unified view for biological dynamics. We also discuss the plausible categorizations, the fundamental nature, and the efficient modeling of the biological dynamics, which arise from the optimization perspective of the dynamical systems. |
| |
Keywords: | Constrained optimization Lotka-Volterra equation Hopfield neural networks Replicator equation Riemannian geometry |
本文献已被 ScienceDirect PubMed 等数据库收录! |
|