Comparative choice and its fitness consequences in a model environment

Much of behavioral ecology is related to the decisions animals make and their resulting fitness consequences. Many of the models employed to predict behavior make assumptions of economic rationality. These assumptions ensure that rational, optimizing models will be solvable and tractable and allows them to make useful predictions. Unfortunately, these assumptions distance models from the real world where the assumptions of rationality are never entirely sound. Economic rationality supposes that outcomes can be assigned objective values within a stable valuation framework, and that choices are made to maximize a decisionmaker's expected payoff. However, there has been substantial work in psychology showing that people's behavior routinely breaches the assumptions of economic rationality and thus deviates from models that include those assumptions. Studies of animal behavior have found similar breaches of economic rationality. These breaches of economic rationality observed in humans and other animals may be due to comparative assessment of options, where the evaluation of an option is based on a relative comparison with the other options available, rather than by an estimation of inherent value. Comparative assessment seems to describe a range of observed behaviors well, but is not consistent with the assumptions of economic rationality. I investigate the potential for non-rational, comparative assessment to be a fitness maximizing strategy in a range of model environments. The models incorporate several realistic features such as: short-run requirements for resources; uncertainty about the options that will be available in the future; and imperfect perception of the options that are currently available. I identify a range of conditions in the simulation where a strategy based on relative comparisons outperforms an economically rational strategy. This work provides insights into where we might expect animals to use relative comparisons and, correspondingly, where we might expect to see deviations from economically rational models of behavior.