Optimal Foraging Theory: Lessons and Application to Adaptive Engineering Systems

In biological systems, optimal strategy is generally defined as optimizing fitness, measured as reproductive value (RV), the expectation of producing surviving offspring from time t onward, given that an organism is in state S(t). Any action can be associated with an expectation of immediate reproductive success. Maximum RV results from the action that maximizes the sum of immediate and future surviving offspring. Adaptive biological behavior is the product of historical experience, heritability, individual variation, and differential fitness among individuals. Foraging tasks are a standard test bed for robot research because of their applicability to many problems. Optimal foraging theory offers explanations and predictions with direct applicability to engineering problems. Much theory development involves optimal solutions based on complete information about the system, but animals do not always conform to predictions of such models. Adaptive approximations to optimality in biological systems offer models for design of engineered systems.