heuristiques

Heuristics are simple, efficient rules or mental shortcuts that people use to make quick judgments and solve problems. Rather than exhaustively analyzing all information, a heuristic aims for a good-enough solution in a reasonable time. They are not guaranteed to be correct, but in many situations they produce accurate or satisfactory results. The term comes from the Greek heuriskein, meaning to discover. In cognitive psychology, heuristics contrasts with algorithms, which are step-by-step procedures that guarantee a correct outcome given sufficient time and information.

Common heuristics include: the availability heuristic, which estimates the likelihood of events based on how easily examples come to mind; the representativeness heuristic, which judges probability by similarity to a prototype; anchoring and adjustment, where people start from an initial value and insufficiently adjust from it; the recognition heuristic, which chooses the recognized option when only some options are known; and affect and fluency heuristics, which depend on feelings or ease of processing.

Applications of heuristics span daily life, business, and policy, helping speed decision making under uncertainty. In medicine, guidelines and heuristics support diagnosis and treatment when data are incomplete. In computer science and artificial intelligence, heuristic methods guide search and optimization (for example, A* search uses a heuristic function to estimate remaining cost). In human-computer interaction, usability experts use heuristic evaluation to identify design problems.

Limitations: heuristics can bias judgments and lead to systematic errors, especially when information is framed in particular ways or base-rate data are ignored. They may underperform for rare events, base rates, or atypical cases. To reduce errors, analysts combine heuristics with data, domain knowledge, probabilistic reasoning, and, when possible, formal methods.

Related topics include the availability and representativeness heuristics, anchoring, and heuristic evaluation in UX and AI search.