GPS (General Problem Solver)
A landmark AI program that modeled human problem-solving using heuristic search.
The General Problem Solver (GPS) was an early artificial intelligence program designed to simulate the way humans approach and solve problems. Developed by Allen Newell and Herbert A. Simon in 1957, GPS was built on the hypothesis that intelligent behavior could be captured by a general-purpose symbolic reasoning system, rather than requiring specialized knowledge for each domain. This made it one of the first serious attempts to build a domain-agnostic AI system capable of tackling a wide variety of tasks.
GPS operates by representing problems as a formal structure consisting of an initial state, a goal state, and a set of operators that transform one state into another. Its core mechanism, known as means-ends analysis, works by identifying the difference between the current state and the desired goal, then selecting operators that reduce that difference. This process recurses through subgoals until the problem is solved or no further progress can be made. The approach treats problem-solving fundamentally as a search through a space of possible states, a framing that became central to AI research for decades.
The significance of GPS extends well beyond its practical capabilities, which were limited to relatively simple symbolic tasks like logic puzzles and algebraic manipulation. It established problem-solving as a computational process amenable to formal study, and it introduced concepts—such as heuristic search, goal decomposition, and state-space representation—that remain foundational in AI planning, automated reasoning, and cognitive architectures. The program also served as a computational model of human cognition, directly influencing the field of cognitive science and the information-processing theory of mind.
While GPS itself was eventually superseded by more powerful and specialized systems, its intellectual legacy is substantial. It demonstrated that symbolic reasoning could be mechanized and generalized, laying groundwork for later developments in expert systems, planning algorithms, and knowledge representation. Newell and Simon's work on GPS contributed directly to their receiving the Turing Award in 1975, cementing GPS as a foundational milestone in the history of artificial intelligence.
