Not sure what you mean by this. Even in the most basic rational choice analysis, where the players are agents seeking to maximize a utility function, the "rewar

Not sure what you mean by this. Even in the most basic rational choice analysis, where the players are agents seeking to maximize a utility function, the "reward" can be nonlinear. I'm not familiar with any economic model that unintentionally restricts the agent's utility to be linear. Sometimes you assume that agents have a linear "utility of money", but everyone in economics knows that this is a taylor expansion around a small region where the linearity assumption is reasonable, and not an actual fact about human preferences.