Non-cooperative games are mathematical models of interactive strategic decision situations.In contrast to cooperative models, they build on the assumption that all possibilities for commitment and contract have been incorporated in the rules of the game.This contribution describes the main models (games in normal form, and games in extensive form), as well as the main concepts that have been proposed to solve these games.Solution concepts predict the outcomes that might arise when the game is played by "rational" individuals, or after learning processes have converged.Most of these solution concepts are variations of the equilibrium concept that was proposed by John Nash in the 1950s, a Nash equilibrium being a combination of strategies such that no player can improve his payoff by deviating unilaterally.The paper also discusses the justifications of these concepts and concludes with remarks about the applicability of game theory in contexts where players are less than fully rational.