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 829–928 of 3949. Source text is displayed literally.

829 830If a solution curve for \(X_A\) is parameterized by831\(t\), changing its parameter to \(\lambda t\) gives a832solution for \(X_B\) on the corresponding time interval.833Positive scaling therefore gives a comparison of time834units. Keeping time fixed gives the sharper quotient835\(\mathcal Q_I\). Neither equivalence identifies the836unadjusted absolute payoffs. Those retain a role if837another process, such as a baseline contribution to838fitness, is sensitive to their absolute levels.839 840\section{Relabeling as a groupoid action}841 842Let \(\mathfrak S_I\) be the permutation group of the843type set. A permutation \(\sigma\) acts by844\(A\mapsto P_\sigma A P_\sigma^{-1}\), where845\(P_\sigma\delta_i=\delta_{\sigma(i)}\).846It preserves \(W_I\), since it fixes \(\mathbf 1\).847It also commutes with \(\Pi_I\), since it preserves848the sum of the coordinates. The action therefore849passes to \(\mathcal Q_I\). Its action groupoid has850objects \(q\in\mathcal Q_I\) and arrows851\(q\to q'\) given by permutations \(\sigma\) for852which \(q'=\sigma q\). Two matrices in the same orbit853determine isomorphic objects. The permutations fixing854a class are the automorphisms of that object. Passing855to the set of orbits forgets this group. Accordingly,856the orbit set classifies a single game up to relabeling,857while the groupoid can also record changes of858presentation along a parameter space.859 860One can equivalently keep the raw matrices and allow861translation by \(W_I\) in the arrows. The group862\(W_I\rtimes\mathfrak S_I\) acts on \(\mathcal A_I\)863by \((w,\sigma)A=\sigma A+w\). Its multiplication864uses the evident permutation action on \(W_I\).865Consider the functor of action groupoids866\[867 \mathcal A_I//(W_I\rtimes\mathfrak S_I)868       \longrightarrow869 \mathcal Q_I//\mathfrak S_I870\]871sending \(A\) to its class and an arrow \((w,\sigma)\)872to \(\sigma\). It is fully faithful: if the classes of873\(A,B\) satisfy \([B]=\sigma[A]\), then874\(w=B-\sigma A\) belongs to \(W_I\) and is the unique875translation giving an arrow above \(\sigma\). It is876essentially surjective by the canonical normalized877representative \(\Pi_I A\) in each class. A quasi-inverse878uses this representative, sending each permutation to an879arrow with translation zero. The comparison back to a880raw matrix is translation by its difference from its881normalized form. The formulas are continuous and even882smooth as functions of the matrices. Consequently they883also give equivalent presentations for families with884continuous or smooth transition data. Here the removal885of redundant payoff coordinates gives an equivalence of886groupoids because the arrows over each permutation are887recovered uniquely.888 889\begin{figure}[htbp]890\centering891\[892\begin{tikzcd}[column sep=3em,row sep=2.4em]893  \mathcal A_I\arrow[r,"\sigma"]\arrow[d,"\Pi_I(-)"']894    &\mathcal A_I\arrow[d,"\Pi_I(-)"]\\895  \mathcal Q_I\arrow[r,"\sigma"]&\mathcal Q_I896\end{tikzcd}897\]898\caption{Canonical normalization respects every permutation of labels.}899\label{fig:payoff-normalization}900\end{figure}901 902\section{Aggregation and local labels}903 904For a parameter base \(S\), a family with local labels905can be given on an open cover by maps to \(\mathcal Q_I\)906along with permutations on overlaps satisfying the907cocycle law. The comparison on overlaps is the one908in Figure~\ref{fig:payoff-normalization}, and population909states 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 transfer