Abstract:We propose a class of evolutionary models that involves an arbitrary exchangeable process as the breeding process and different selection schemes. In those models, a new genome is born according to the breeding process, and then a genome is removed according to the selection scheme that involves fitness. Thus the population size remains constant. The process evolves according to a Markov chain, and, unlike in many other existing models, the stationary distribution -- so called mutation-selection equilibrium -- can be easily found and studied. The behaviour of the stationary distribution when the population size increases is our main object of interest. Several phase-transition theorems are proved.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to propose a class of evolutionary models. These models can use an arbitrary commutative process as the reproduction process and adopt different selection schemes. In these models, new genomes are generated according to the reproduction process, and then a genome is removed according to the selection scheme involving fitness, so that the population size remains constant. This process evolves according to a Markov chain. Different from many other existing models, the steady - state distribution (i.e., mutation - selection balance) of this model can be easily found and studied. The paper pays special attention to the behavior of the steady - state distribution when the population size increases and proves several phase - transition theorems.
Specifically, the paper proposes the following key points:
1. **Model Introduction**: A class of probabilistic evolutionary models is introduced. This model has three purposes: as an abstract model of biological evolution; as an efficient genetic algorithm that is easy to analyze theoretically; and as a bridge between genetic algorithms and Bayesian non - parametric MCMC methods.
2. **Steady - State Distribution**: For any fitness function, the steady - state distribution of the model can be expressed in a closed form, which makes it possible to study the model behavior under different population sizes, mutation rates, and fitness scalings.
3. **Phase - Transition Phenomenon**: Two phase - transitions are discovered, and these phase - transitions exist for all fitness functions. These phase - transitions mainly occur at the transition points between the influence of fitness and the influence of the prior distribution.
4. **Mathematical Formulation**: The paper describes in detail the mathematical formulation of the model, including Markov chains, commutative processes, Deligne distributions, de Finetti's theorem, weak convergence of probability measures, etc.
5. **Limit Results**: The paper explores the limit behavior of the steady - state distribution when the population size tends to infinity. Specifically, when the prior distribution is independent of the population size \(n\), and the fitness function \(w_n\) depends on \(n\), the paper proves that the limit behavior of the steady - state distribution will be different under different values of \(\lambda\):
- When \(\lambda\in[0, 1)\) and \(\phi(1)<\phi(2)\), only the fittest genotype survives in the limit.
- When \(\lambda = 1\), the steady - state distribution converges to a non - degenerate distribution that depends on the prior distribution and the function \(\phi\).
- When \(\lambda> 1\), the influence of fitness disappears, and only the prior distribution plays a role.
6. **Dirichlet Prior**: The case of Dirichlet prior is particularly considered. In this case, the reproduction process is a generalized Pólya urn process. The paper proves that in this case, phase - transitions still occur and gives the specific limit distribution.
In conclusion, by proposing a new evolutionary model, this paper not only solves the problem of difficulty in finding the steady - state distribution in traditional models but also reveals the profound influence of fitness and prior distribution on population dynamics.