Indexed text attachment
evolutionary-game-theory-in-higher-categories.tex
Source for Evolutionary Game Theory in Higher Categories
View paper · Download original file
Lines 3801–3900 of 3949. Source text is displayed literally.
3801Over a commutative ring the kernel checks3802\[3803 \sum_i F(x,u)_i=(1-m(x))\mu_x(u).3804\]3805It follows that the drift is tangent to the mass-one hyperplane.3806The zero-coordinate property, invariance under adding a common payoff3807on that hyperplane, and linearity under uniform payoff scaling are3808also checked. These are identities of algebraic expressions used in3809Chapter~\ref{ch:replicator}.3810 3811Over a field the project defines the reproductive update3812\[3813 S(x,u)_i=\frac{x_i u_i}{\mu_x(u)}.3814\]3815It verifies unit mass when the denominator is nonzero. Over a linearly3816ordered field with $x$ in the simplex and strictly positive fitness3817$u$, Lean derives positivity of the denominator and proves that the3818output lies in the simplex. For affine reproductive weights3819$w_i=1+\eta u_i$, under $m(x)=1$ and $1+\eta\mu_x(u)\ne0$, it also checks3820the exact discrete-to-drift identity3821\[3822 S(x,w)_i-x_i3823 =\frac{\eta}{1+\eta\mu_x(u)}F(x,u)_i .3824\]3825No time limit is needed to obtain this equality.3826 3827\section{Deterministic aggregation}3828 3829For a map $q:I\to J$, the Lean definition of aggregation is3830$(q_*x)_j=\sum_{q(i)=j}x_i$. The project proves preservation of mass3831and of simplex membership, functoriality under composition, and the3832pullback-pushforward pairing3833\[3834 \mu_x(v\circ q)=\mu_{q_*x}(v).3835\]3836For $v:J\to R$ it proves the two commutation identities3837\[3838 q_*F(x,v\circ q)=F(q_*x,v),\qquad3839 q_*S(x,v\circ q)=S(q_*x,v).3840\]3841A state-dependent form and a specialization to pullbacks of payoff3842matrices make the first equation applicable to the deterministic3843semiconjugacy of Chapter~\ref{ch:semiconjugacy}. Each result uses3844finite sums and function extensionality. The second identity is3845algebraic as an equality of field terms. Its population interpretation3846uses the nonzero normalizer as in the preceding section.3847 3848\section{Reproducing the check}3849 3850From the extracted Lean source, run \texttt{lake exe cache get},3851\texttt{lake build}, and \texttt{lake env lean Audit.lean}.3852The manifest pins the mathlib revision. The last command prints3853the transitive axioms of the audited declarations. The supplied audit3854records standard Lean axioms and no admitted proofs. The appendix3855covers the algebraic vector field and reproductive update. The main3856text gives the analytical and higher-categorical arguments, including3857existence of flows, stochastic limits, and the channel covariance3858formula in Chapter~\ref{ch:stochastic-compression}.3859\chapter{Subject index}3860\label{app:index}3861\small3862\setlength{\parskip}{0pt}3863\noindent References give the chapters or appendices where each term is developed.3864\smallskip3865\begin{multicols}{2}\raggedright3866\begin{description}[leftmargin=0pt,labelindent=0pt,itemsep=.10\baselineskip]3867\item[Aggregation] \ref{ch:aggregation}, \ref{ch:semiconjugacy}, \ref{ch:stochastic-compression}.3868\item[Assortative interaction] \ref{ch:structured}.3869\item[Best response] \ref{ch:revision}.3870\item[Coarse correlated equilibrium] \ref{ch:agent-learning}.3871\item[Context] \ref{ch:contextual-games}, \ref{ch:higher-equivariance}.3872\item[Continuation] \ref{ch:open-games}, \ref{ch:compositional-games}.3873\item[Coplay] \ref{ch:open-games}.3874\item[Covariance defect] \ref{ch:stochastic-compression}.3875\item[Descent] \ref{ch:contextual-games}, \ref{ch:higher-equivariance}.3876\item[Equilibria] \ref{ch:ess}, \ref{ch:higher-equivariance}.3877\item[Evolutionarily stable strategy] \ref{ch:ess}, \ref{ch:superiority}, \ref{ch:lyapunov}.3878\item[Exponential weights] \ref{ch:agent-learning}.3879\item[Finite populations] \ref{ch:markov}, \ref{ch:mean-field}.3880\item[Fitness] \ref{ch:population}, \ref{ch:payoff-equivalence}.3881\item[Formalization] \ref{app:formalization}.3882\item[Homotopy quotient] \ref{ch:higher-equivariance}.3883\item[Invasion] \ref{ch:ess}, \ref{ch:superiority}.3884\item[Lens] \ref{ch:open-games}.3885\item[Logit dynamics] \ref{ch:revision}.3886\item[Lumpability] \ref{ch:aggregation}.3887\item[Lyapunov functions] \ref{ch:lyapunov}.3888\item[Markov kernel] \ref{ch:markov}, \ref{ch:aggregation}.3889\item[Mass action] \ref{ch:open-dynamics}.3890\item[Mean-field limit] \ref{ch:mean-field}.3891\item[Monodromy] \ref{ch:contextual-games}.3892\item[Multipopulation dynamics] \ref{ch:structured}.3893\item[Mutation] \ref{ch:replicator}, \ref{ch:stochastic-stability}.3894\item[Open dynamics] \ref{ch:open-dynamics}.3895\item[Open game] \ref{ch:open-games}, \ref{ch:compositional-games}.3896\item[Payoff shift] \ref{ch:payoff-equivalence}, \ref{ch:agent-learning}.3897\item[Pushout gluing] \ref{ch:open-dynamics}.3898\item[Regret] \ref{ch:agent-learning}.3899\item[Revision protocol] \ref{ch:revision}, \ref{ch:compositional-games}.3900\item[Replicator equation] \ref{ch:replicator}, \ref{ch:lyapunov}.