Atlas record: https://www.aitimeline.dev/en/events/evt-0089/
Research · June 1959
General Problem Solver
Newell, Shaw and Simon separated the solving method from the particular problem: the program reduces the difference between the current and the goal state.
Why it matters
Means-ends analysis was the first attempt to describe problem solving in general rather than for one domain.
GPS worked on a problem given as states, operators and a table of differences. It solved puzzles and symbolic transformations, but required a person to describe the space in advance. Simon and Newell treated the program both as an engineering tool and as a model of human thinking.
Event record
- Event date
- June 1959
- Timeline date
- Event date
- Verification
- Sources gathered automatically · September 17, 2026
- Lines
- ID
- evt-0089
Presented at the IFIP conference in Paris in June 1959; the RAND preprint exists with the date 30 December 1958 (Stanford's Feigenbaum papers) and in a copy dated February 1959.
Related events
Earlier
Later
- Related The A* algorithm 1968links to this record
The algorithm formalises the guided search GPS performed heuristically.
- Contrasts What Computers Can't Do 1972links to this record
The book takes apart the GPS authors' claims about the generality of the method.
- Contrasts The Art of Artificial Intelligence 1977links to this record
It directly rejects the bet on a general problem-solving mechanism.
- Related Herbert Simon's Nobel Prize 1978links to this record
The award recognises the decision-making research that Logic Theorist and GPS grew out of.
- Contrasts Elephants Don't Play Chess 1990links to this record
It directly denies the programme the General Problem Solver set out.