General Problem Solver (A. Newell & H. Simon)
The General Problem Solver (GPS) was a theory of human problem solving stated in the form of a simulation program (Ernst & Newell, 1969; Newell & Simon, 1972). This program and the associated theoretical framework had a significant impact on the subsequent direction of cognitive psychology. It also introduced the use of productions as a method for specifying cognitive models.
The theoretical framework was information processing and attempted to explain all behavior as a function of memory operations, control processes and rules. The methodology for testing the theory involved developing a computer simulation and then comparing the results of the simulation with human behavior in a given task. Such comparisons also made use of protocol analysis (Ericsson & Simon, 1984) in which the verbal reports of a person solving a task are used as indicators of cognitive processes (see http://www.rci.rutgers.edu/~cfs/472_html/CogArch/Protocol.html)
GPS was intended to provide a core set of processes that could be used to solve a variety of different types of problems. The critical step in solving a problem with GPS is the definition of the problem space in terms of the goal to be achieved and the transformation rules. Using a means-end-analysis approach, GPS would divide the overall goal into subgoals and attempt to solve each of those. Some of the basic solution rules include: (1) transform one object into another, (2) reduce the different between two objects, and (3) apply an operator to an object. One of the key elements need by GPS to solve problems was an operator-difference table that specified what transformations were possible.
While GPS was intended to be a general problem-solver, it could only be applied to "well-defined" problems such as proving theorems in logic or geometry, word puzzles and chess. However, GPS was the basis other theoretical work by Newell et al. such as SOAR and GOMS. Newell (1990) provides a summary of how this work evolved.
Here is a trace of GPS solving the logic problem to transform L1= R*(-P => Q) into L2=(Q \/ P)*R (Newell & Simon, 1972, p420):
Goal 1: Transform L1 into LO
For more on GPS and the work of Newell & Simon, see:
© 2013 Richard Culatta | Multiple domain web hosting provided by InnovativeLearning.com