General Problem Solver

Quick Reference

A computer program designed to simulate human problem solving, introduced in 1958 by the US cognitive scientist Allen Newell (1927–92) and the US economists and decision theorists John Clark Shaw (born 1933) and Herbert A(lexander) Simon (1916–2001), and developed further in 1972 by Newell and Simon. In this program, a problem is represented as a table of connections showing the distances between all pairs of states (initial, intervening, and final), and problem solving is modelled as a search through the problem space using permissible operators (actions), the task being to find a path of operators from an initial state to a goal state. A typical initial state is the starting position in a game of chess, and the corresponding goal state is checkmating the opponent; a more commonplace initial state is being at home with a child and a car that has a flat battery, and the corresponding goal state in this case might be delivering the child to nursery school. In general, the problem space is too large for exhaustive search to be feasible, and therefore the cognitive heuristic of means-end analysis is implemented. The program is able to solve water-jar problems, the Tower of Hanoi problem, and many logic problems and chess problems. It represents one of the first major landmarks in the history of artificial intelligence. GPS abbrev.

Subjects: Psychology.

Reference entries