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 2329–2428 of 3949. Source text is displayed literally.

2329                     \in B_{\mathcal H\circ\mathcal G}(x,k)\\2330 &\quad\Longleftrightarrow\2331   (\sigma,\sigma')\in B_{\mathcal G}(x,k_\tau)2332   \quad\text{and}\quad2333   (\tau,\tau')\in B_{\mathcal H}(y,k).2334\end{aligned}2335\]2336The two tests use the incumbent context. The second component sees2337the output generated by \(\sigma\). The first component evaluates2338alternatives using \(\tau\) as a fixed strategy, including its play2339and coplay at changed inputs. This is a prescription for testing2340responses with the other components' strategies held fixed.2341 2342\begin{proposition}[Associativity of the response construction]2343Sequential composition is associative under the canonical2344reassociation of strategy products. Singleton-strategy identity2345lenses with the sole identity response give identities.2346\end{proposition}2347 2348\begin{proof}2349Write \(\mathcal J:(Z,Q)\to(W,T)\) for the third component and2350\(\upsilon\) for its incumbent. For an outer continuation2351\(\ell:W\to T\), put2352\[2353 k_{\upsilon}(z')=2354 C_{\mathcal J}\bigl(\upsilon,z',\ell(P_{\mathcal J}(\upsilon,z'))\bigr).2355\]2356With \(y=P_{\mathcal G}(\sigma,x)\) and2357\(z=P_{\mathcal H}(\tau,y)\), either parenthesization tests the2358three responses in the contexts2359\[2360\begin{array}{rl}2361 \mathcal G:&2362 \bigl(x,\ y'\mapsto C_{\mathcal H}2363             (\tau,y',k_{\upsilon}(P_{\mathcal H}(\tau,y')))\bigr),\\2364 \mathcal H:& (y,k_{\upsilon}),\\2365 \mathcal J:& (z,\ell).2366\end{array}2367\]2368For the first parenthesization these contexts result from first2369transporting through \(\mathcal J\) and then through \(\mathcal H\).2370For the second, the coplay of the downstream lens composite expands2371to the same substitution. The play and coplay agree by lens2372associativity. Inserting an identity leaves the transported state and2373continuation unchanged and adds the sole relation on a singleton.2374\end{proof}2375 2376Taking strategy sets up to reindexing by these coherent bijections2377gives an ordinary category. Keeping actual strategy sets makes the2378reassociation visible as a comparison between games. The next chapter2379also keeps reindexings that need not be invertible.2380 2381\section{A compositional equilibrium test}2382 2383The diagonal of the response formula gives the equilibrium criterion2384\[2385  \sigma\in E_{\mathcal G}(x,k_\tau),2386  \qquad2387  \tau\in E_{\mathcal H}(P_{\mathcal G}(\sigma,x),k),2388\]2389where \(E_{\mathcal G}\) denotes the contextual equilibrium set.2390This is a system of mutually dependent tests. Retaining only2391\(E_{\mathcal G}\) is enough for this particular diagonal calculation2392when all contexts and the fixed-strategy lenses are retained. Off-diagonal2393responses are needed to study transformations that preserve response2394possibilities or to introduce an update rule using those possibilities.2395For evolutionary models the selected continuation can include2396frequency-dependent payoffs. One must still specify how the current2397population determines that continuation, which strategy information is2398available, and how admissible responses change frequencies. The2399interface supplies a way to keep these choices explicit.2400 2401For instance, an adapter can expose the population state as an input2402and compute a payoff vector as its output. A downstream selection2403component then evaluates each strategy against that vector.2404Compositional optimization tests whether a strategy passes the2405resulting response relation. A dynamical model adds a state transition2406or vector field on populations, whose state change is an additional2407map. These two observations about the system can then be checked for2408agreement under the hypotheses of Chapter~\ref{ch:lyapunov}.2409 2410\section{Deterministic adapters}2411 2412The response formula gives a useful rule for changing the language in2413which an interface is described. Suppose2414\(L:(X',S')\to(X,S)\) and \(M:(Y,R)\to(Y',R')\) are lenses.2415Regard each as a singleton-strategy game with its sole response.2416For a context \((x',k')\) of \(M\circ\mathcal G\circ L\), the2417response test of the middle game has context2418\[2419 \left(v_L(x'),\ y\longmapsto u_M(y,k'(v_M(y)))\right).2420\]2421This follows by first applying the lens substitution to the return2422from \(M\), then carrying the input forward through \(L\). The2423singleton comparisons impose no further response condition, and the2424same formula therefore computes equilibria of the adapted game.2425If a downstream adapter translates every return by the same amount2426for the relevant decision context, a real maximizing response test2427is unchanged. More generally a strictly increasing transformation2428of every return in that test preserves the ordering of alternatives.