Evolution, constraint, cooperation, and community structure in simple models

This dissertation studies models for evolutionary change in ecological communities, particularly the Lotka-Volterra equations and the adaptive dynamics tradition of evolutionary modeling. A recurring theme is the role of constraints on evolution, and how they shape the direction evolution takes, particularly with regard to cooperation. In several different modeling settings, an argument is constructed that cooperation can take many forms, and is not always undermined by temptation to behave selfishly.; The first chapter is a nontechnical introduction, which talks in general terms about the role of scientific models in larger social discourses. Chapters 2 and 3 review the published literature on Lotka-Volterra equations and adaptive dynamics. Chapter 3 introduces new techniques for studying the effects of constraints on evolutionary motion.; Chapter 4 studies evolution in Lotka-Volterra models when interactions are derived from simple characteristics of single populations; chapter 5 removes that assumption and allows the interactions to evolve directly. This choice produces a kind of diversification that is not seen in previously known adaptive dynamics models, and illuminates how phenotypes can constrain the evolution of cooperation. Chapter 6 applies the techniques of chapter 5 to a variety of models.; Chapter 7 shifts to a game theory model, exploring ways out of the "prisoner's dilemma", and chapters 8 and 9 look at cooperation in the form of large-scale ecological regulation. In these chapters as well, the tension between constraints and cooperation is central. Also in chapter 9, I describe a process with some similarity to animal learning that may contribute to the regulation of planetary processes and the Gaia hypothesis.; The concluding chapter explores the themes of constraint and cooperation that recur throughout, and discusses the learning hypothesis of chapter 9 and problems of degeneracy that occur when models are oversimplified.