That is the simplest model of rationality, but there are others. Wikipedia[1] lists three others in addition to perfect information:
> Perfect information: The simple rational choice model above assumes that the individual has full or perfect information about the alternatives, i.e., the ranking between two alternatives involves no uncertainty.
> Choice under uncertainty: In a richer model that involves uncertainty about the how choices (actions) lead to eventual outcomes, the individual effectively chooses between lotteries, where each lottery induces a different probability distribution over outcomes. The additional assumption of independence of irrelevant alternatives then leads to expected utility theory.
> Inter-temporal choice: when decisions affect choices (such as consumption) at different points in time, the standard method for evaluating alternatives across time involves discounting future payoffs.
> Limited cognitive ability: identifying and weighing each alternative against every other may take time, effort, and mental capacity. Recognising the cost that these impose or cognitive limitations of individuals gives rise to theories of bounded rationality.
I didn't invent rational choice theory, so don't ask me to justify it.
Ok, well, I'll take a stab at it—the whole concept is exquisitely well crafted as a basis for the argument that liberal (in the classical sense) economic, and even political, systems are ideal in terms of experienced human outcome.
> Wouldn't economic rationality assume that the decisions of actors are rational rather than based on perfect information about their consequences?
Economic rationality is defined to include both perfect information and optimality given that information, so there is no “instead of”—your two alternatives are different phrasings of the same thing.