Evolutionary dynamics of zero-sum games (Q1060991)

The model under consideration is based in game theory with differential equations arising in dynamical systems. The authors first attempt to give numerical solutions using the simplest approach, namely Euler's polygonal approximation method. One of the general systems they introduce was considered by \textit{P. D. Taylor} and \textit{L. B. Jonker} [Math. Biosci. 40, 145-156 (1978; Zbl 0395.90118)] as a dynamical model. It applies the ideas of \textit{J. M. Smith} [see Evolution and the theory of games. (1982; Zbl 0526.90102)] of game theory to evolution. Independently this system of equations was introduced by Eigen's group [see \textit{M. Eigen} and \textit{P. Schuster}, The hypercycle. Springer (1979)] studying origin of life problems. The discrete time model is due to \textit{W. G. S. Hines} [J. Appl. Probab. 17, 333-340 (1980; Zbl 0439.92021)] and to the authors [J. Math. Biol. 17, 241-251 (1983; Zbl 0519.92014)]. In this paper the authors examine the case antipodal to the population genetic models, which is the case where the payoff matrix is antisymmetric. This condition says that the games considered in the model are zero-sum. This special case of evolutionary game models has a separate pattern of applications all its own. \textit{T. Nagylaki} [Proc. Natl. Acad. Sci. USA 80, 5941-5945 and 6278-6281 (1983; Zbl 0521.92014 and Zbl 0528.92013 resp.)] has introduced an important equation with antisymmetric payoff function as a model for game conversion. The authors classify and study the general behavior. They introduce an equilibrium and give some important result about a log function which might be important in applications. Relationship with equilibria in the boundary is studied in the autosymmetric case. The interior equilibrium case is also considered and the empty case is presented. In the following, they describe some concepts from topological dynamics due to \textit{C. Conley} [Isolated invariant sets and the Morse index. (1978; Zbl 0397.34056)] and \textit{S. Smale} [Global Analysis, Proc. Sympos. pure Math. 14, 289-297 (1970; Zbl 0205.541)] and apply such results obtaining important theorems. Some of them are related to equilibria, and they state a conjecture about the set of equilibria. Finally the authors study Hamiltonians for the continuous system. Foliations, symplectic forms, exterior algebra, vector tangents etc. are presented and studied for the corresponding case. They conclude that because they are usually structurally unstable, Hamiltonian differential systems are only appropriate when the conservation effects are essential in the underlying theory rather than accidental consequences of the choice of model design. In particular the authors think that Nagylaki is correct in his preference for the discrete time model for the biological applications he has in mind.