I performed my doctoral research in the Department of Information and Computer Science at the University of California, Irvine. Working with Pat Langley, I developed a system that relies on a spreading-activation model of memory for problem solving and learning. Experiments with the system account for basic learning phenomena exhibited by human problem solvers. From this work, we learned how the constraints of limited memory can impact high-level human problem solving. In addition, the work demonstrates how mechanisms that overcome these constraints may naturally lead to many of the regularities in learning that have been identified in human problem solving.