Game-Theoretic Models of Animal Behavior Observed in Some Recent Experiments

In this dissertation, we create three theoretical models to answer questions raised by recent experiments that lie beyond the scope of current theory. In the landmark-effect model, we determine size, shape and location for a territory that is optimal in the sense of minimizing defense costs, when a given proportion of the boundary is landmarked and its primary benefit in terms of fitness is greater ease of detecting intruders across it. In the subjective-resource-value model, we develop a game-theoretic model based on the War-of-Attrition game. Our results confirm that allowing players to adapt their subjective resource value based on their experiences can generate strong winner effects with weak or even no loser effects, which is not predicted by other theoretical models. In the rearguard-action model, we develop two versions of a game-theoretic model with different hypotheses on the function of volatile chemical emissions in animal contests, and we compare their results with observations in experiments. The two hypotheses are whether volatile chemicals are released to prevent the winner of the current round of contest from translating its victory into permanent possession of a contested resource, or are used to prevent a winner from inflicting costs on a fleeing loser.