For both humans and animals, choice is a necessary part of life. This chapter focuses on the processes mediating choice in service of a local goal--particularly when the chooser has repeated exposures to the same choice point. Problem-solving tasks offer many examples where humans must choose in service of a local goal; here the local goal is reaching a solution. For example, when solving an algebra equation, solvers often have multiple strategies available (e.g.,graphing, quadratic formula), and they must choose among these strategies in order to reach a solution. Animals also face choice in service of a local goal. For example, in foraging, the local goal is to obtain some food, and the animal must choose among multiple patches in which food may be sought.
This chapter raises several issues regarding ACT-R's ability to simulate human and animal choice in changing environments. In particular, the focus is on ACT-R predictions when time-based decay is incorporated into the computation of productions' success. A variety of examples demonstrate that the decay-based parameter-learning mechanism allows ACT-R models to account for a variety of learning and choice data at a fine-grained level of detail.
Myers Experiment Model (with decay)
Building Sticks (with decay) Model