Implications of routine-based learning for decision making

Abstract
Various mathematical models have been proposed in the economic literature for the modelling of boundedly rational learning. Each of these models considers one type of learning, like imitation, satisficing, trial-and-error or reinforcement learning. This paper combines several types of boundedly rational learning and develops a more general learning model. Then, the characteristics of this learning model are examined for a quite general decision situation involving risk. It is studied under which conditions learning leads to an expected utility maximising behaviour. Furthermore, it is analysed what kind of deviations from the utility maximising behaviour are caused by the different aspects of learning.