Conflict solutions with Mathematica. Two-person games (Q3835992)

While the simplex method provides an algorithm for solving matrix (zero-sum) games, the emphasis in this book is on a Mathematica program for an algorithm based on the linear complementarity problem for finding all Nash equilibria in a finite noncooperative (bimatrix) game. Several conflict situations are examined with this tool. The algorithm is limited to games with relatively few pure strategies, but a nice feature is that one can get solutions in symbolic form, often revealing more of the structure of the game than when only numerical results are found. Thus via the computer algebra the emphasis is more on experimenting and less on classical analytical solution techniques. The intent is that when the algorithms for experimenting are understood the underlying theory is also better understood. The first chapter introduces the various concepts in bimatrix games and the linear complementarity problem. Chapter two uses these tools to analyze inspection games. Chapters 3 and 4, on evolution games and perfect equilibrium, introduce refinements which restrict the number of equilibria. The last chapter deals with zero-sum games. Each chapter closes with suggestions for further experimentation on the computer. Two short appendices discuss eigenvectors and eigenvalues and nonlinear dynamics. A third discusses the Lemke-Howson algorithm for the LCP, and a fourth describes the author's Mathematica package GameTheory`Bimatrix'. Prerequisites are some acquaintance with the Mathematica language and some familiarity with vectors and matrices. An accompanying diskette contains Mathematica notebooks for all the chapters as well as the package GameTheory`Bimatrix'.