Abstract
Evolutionary dynamics provides us the mathematical framework to explore the dynamics of evolution in diverse research topics, such as theoretical ecology, population genetics, epidemiology, immunology and etc. The basic building blocks of evolutionary dynamics are reproduction, selection and mutation. To discuss the basic properties and research topics on selection and mutation, we only address two important questions about the origin of life and the biodiversity in ecological systems in the thesis. Start from the quasispecies equations which describes the dynamics of genome frequency under mutation-selection processes, an important relationship among mutation rate, genome length and relative fitness called error threshold is discovered to maintain the evolutionary advantage of the fittest genome in the single-peak fitness landscape. Apply the idea and the techniques to the peak-mesa-background fitness landscape, the universal phase diagram is found to explain the error threshold under different conditions. Next, we switch the focus to the evolutionary dynamics in population level by considering the frequency-dependent selection without mutation in replicator equations. The Reference-Gamble-Birth (RGB) algorithm, which has the same dynamics as replicator equations in infinite population limit, is introduced to handle the stochastic interactions in agent-based simulation. The numerical results of RGB algorithm shows that the population dynamics shares the same universal extinction patterns in generalized rock-paper-scissors communities under different types of spatial networks. The universal extinction patterns show that the population dynamics can be divided into two phases with a critical point at the phase transition, and therefore the extinction patterns can be viewed as the indicator of ecological stability (biodiversity) for ecological systems.