Independent scholarship · Open accessThe monthly digest

Indexed text attachment

evolutionary-game-theory-in-higher-categories.tex

Source for Evolutionary Game Theory in Higher Categories

View paper · Download original file

Lines 909–1008 of 3949. Source text is displayed literally.

909states transform by the same permutations. Thus the910payoff family and population bundle can carry911nontrivial label transport. For example, a loop can912act by a permutation fixing the payoff class and913moving some points of the simplex. A solution family914that is required to be a section must respect that915transport. A numerical solution set at a single916parameter point does not specify how to continue the917individual solutions around such a loop.918 919A payoff-preserving surjection of type sets is a920different operation. It can remove duplicate labels by921summing their population weights. Its fibers over922interior mass assignments can have positive dimension.923It therefore fails to identify the underlying state924spaces even when the expected payoff descends. In925particular, if a state puts positive mass above a926target type that has at least two preimages, a nonzero927transfer of mass within that fiber leaves its image928unchanged. Chapter~\ref{ch:ess} shows that such a transfer929prevents the specified lift from being an ESS. The930groupoid equivalence for normalization uses actual931comparisons with inverses. A coarse aggregation needs932a separate argument for each invariant transported933along it.934 935\section{Games with selected residents}936 937The total solution space is more informative for families938than a list of stable residents for each matrix. For the939ESS predicate, form the subspace940\[941 \mathcal S_I=942 \{(q,p)\in\mathcal Q_I\times\Delta(I):943                   p\text{ is evolutionarily stable at }q\}.944\]945The definition is independent of the representative of \(q\).946Permuting the game and the resident together preserves this947subspace, by the payoff computation in948Chapter~\ref{ch:ess}. Projection therefore gives a functor949\(\mathcal S_I//\mathfrak S_I\to950\mathcal Q_I//\mathfrak S_I\). In a parameter family, a951selected resident is a lift along this projection.952In local coordinates it consists of a choice on each chart953and compatibility of the choices on every overlap.954 955Over a chosen game class with a specified identification of956its labels, the stable residents form the ordinary set957\(\operatorname{ESS}(q)\). If one instead restricts to the958full subgroupoid of games isomorphic to \(q\), then the959category of marked residents is equivalent to the action960groupoid of \(\operatorname{Stab}(q)\) on961\(\operatorname{ESS}(q)\). Indeed, bringing a game in that962orbit back to \(q\) takes its resident to one at \(q\),963and different choices of a permutation differ by a964stabilizer element. The groupoid records how residents can965be identified by changes of labels. This construction only966uses the stated stability predicate and its invariance967under permutations. It applies equally to the Nash and968neutral predicates.969\par970 971\bookpart{II}{Dynamics on population states}972\chapter{Replicator flows and mutation}973\label{ch:replicator}974 975\section{From abundance to proportion}976 977Let \(S\) be a nonempty finite strategy set and write \(\Delta(S)\)978for its probability simplex. A fitness assignment is a map979\(F:\Delta(S)\to\mathbb R^S\). Its \(i\)-th entry is the rate980associated with strategy \(i\), evaluated in the population \(x\).981For a matrix \(A\), the assignment \(F(x)=Ax\) evaluates the982expected payoff of each strategy against the population. The983derivation below applies to nonaffine assignments as well.984Taylor and Jonker give an early dynamical treatment, and985Alexander explains the abundance and revision interpretations986of the equation \cite{taylor-jonker,alexander}.987 988Suppose \(n_i\) is the abundance of strategy \(i\) and that989\(\dot n_i=n_iF_i(x)\), where \(x_i=n_i/\sum_j n_j\).990Writing \(M=\sum_j n_j\) and \(\phi(x)=\sum_jx_j F_j(x)\),991one obtains \(\dot M=M\phi(x)\). The quotient rule gives992\begin{equation}993  \dot x_i = x_i\bigl(F_i(x)-\phi(x)\bigr).994  \label{eq:replicator}995\end{equation}996The subtraction of mean fitness is thus the correction due to997normalization. Conversely, if \(x\) solves \eqref{eq:replicator},998then \(M(t)=M(0)\exp(\int_0^t\phi(x(s))\,ds)\) recovers a999positive total abundance and \(n_i=Mx_i\) solves the abundance1000equation. The abundance formulation uses independence of the1001fitness assignment from the total. If the rates depend on total1002abundance, the proportions generally retain that dependence1003and need an additional state variable.1004 1005There are two simple symmetries of \eqref{eq:replicator}.1006Adding \(c(x)\) to every \(F_i(x)\) adds \(c(x)\) to \(\phi(x)\)1007and leaves the field fixed. Relabeling \(S\) transports \(x\)1008and \(F\) together. For an affine assignment, adding the same