Abstract : The concept of production-executing machines has been used to construct a number of simulations of human problem solving. With a few exceptions the simulations have been of problem solving in situations in which there is no real-time pressure to respond. Such situations are typical of the paradigms used to study attention and performance. A model of production selection and execution has been developed that subsumes the previous problem solving models and that can be applied to real-time situations. The model has been used to simulate human performance in choice reaction time, stimulus repetition, dual task, and cue-conflict (Stroop) situations. This extends the use of production system models to encompass both problem solving and attention limited behavior. (Author)