A rational agent is an agent that selects actions expected to maximise its performance measure given its knowledge, percept history and available actions, a foundational concept in decision theory and classical AI. Rationality is typically formalised through an explicit utility function over outcomes, and mechanism design and multi-agent systems often assume agents behave rationally when reasoning about incentives. The concept contrasts with agents that act on fixed rules or bounded heuristics rather than expected-utility optimisation.