Scope
This monograph takes the classical organization of J. McKenzie Alexander’s Evolutionary Game Theory (Cambridge University Press, 2023, DOI 10.1017/9781108582063) as a point of departure. It develops evolutionary stability and invasion, replicator and revision dynamics, stochastic selection and mean-field limits, open games and dynamical systems, and families over a smooth site.
Mathematical results
The book compares exact aggregation of drift and Markov kernels with the error incurred by stochastic compression. A covariance formula identifies the pointwise failure of a stochastic channel to intertwine replicator drift and gives a composition law and finite-time error estimate. The higher-categorical chapters formulate coherent actions and descent for parametrized systems and relate an action groupoid to symmetric equilibrium and isotropy. Online selection connects these constructions to learning rules.
Evidence and authorship
The companion Lean 4.19.0 artifact contains 23 audited algebraic theorems, including simplex preservation for the positive-fitness update and exact commutation of drift and normalized selection with deterministic aggregation under compatible payoffs. The analytical and higher-categorical proofs are in the manuscript. The complete standalone LaTeX file is designated for text indexing on publication. This anonymous manuscript was prepared with AI assistance in research, drafting, typesetting, and verification.