Fixation probabilities in evolutionary dynamics under weak selection (Q2659034): Difference between revisions

The aim of this paper is to generalize the weak selection method for calculating the fixation probability in a wide class of evolutionary models. The main result is the expansion of weak selection of the first order of the probability of fixing a mutant for any initial condition. This result is applicable to arbitrary shapes of spatial structure and frequency-dependent selection, and the expansion can be calculated for any particular model and initial configuration by solving a system of linear equations. Under conditions that are applicable to most of the models of interest, the size of this system -- and therefore the difficulty of computing this extension -- exhibits polynomial growth in population size. The main result of the authors is to prove Theorem 2 that there is a first-order expansion of the fixation probability in the intensity of choice for any initial configuration. This expansion has three main components: (1) reproductive value, which quantifies the expected contribution to future generations; (2) neutral residence time, which can be interpreted as the average number of steps at which all individuals have a type before absorption, taking into account the initial state of the population; and (3) Fourier coefficients of the first-order effects of the probability that a single replace occurs in one update step. From Theorem 2 it follows that the complexity of computing this first-order decomposition is O (N3 (D + 1)), where D is the degree of the process. This boundary can be further improved in some cases, taking into account the structural properties of the population, for example, in the case of updating ``Death-birth'' on a regular graph. In any case, for a fixed degree D, the size of the system demonstrates polynomial growth in N, while the number of states in the evolutionary process grows exponentially in N. From these results, many well-known results can be inferred on the critical benefit-to-cost ratios for making collaboration preferable in addressing social dilemmas. The results provide more information than simply when weak selection favors a particular trait; the results also allow us to investigate this issue for a fairly large class of evolutionary update rules, population structures, and initial configurations. The results apply to a wide class of stochastic evolutionary models in which size and spatial structure are arbitrary (but fixed).