Heuristic

A heuristic is a problem-solving method, strategy, or rule of thumb that is not guaranteed to be optimal, complete, or even correct, but is sufficient for reaching a satisfactory solution within available time and cognitive resources. The term derives from the Greek heuriskein (to find or discover) and was popularized in computer science and psychology by Herbert Simon's concept of bounded rationality and his distinction between satisficing (seeking a good-enough solution) and optimizing (seeking the best possible solution).

In computer science, heuristics are essential when the problem space is too large for exhaustive search. A* search uses a heuristic function to estimate the cost to the goal. Simulated annealing uses a temperature parameter to escape local optima. Genetic algorithms use population-based heuristics to explore design spaces. In each case, the heuristic is not an approximation of the true solution; it is a strategy for navigating a space whose true structure is unknown or intractably large.

The systems-theoretic significance of heuristics is that they are not cognitive failures. They are adaptive responses to complexity. An organism that optimized every decision would never act. A proof assistant that searched exhaustively would never prove. A society that deliberated until consensus was unanimous would never govern. Heuristics are the mechanisms by which systems operate under bounded rationality — the rationality of agents with limited information, limited time, and limited computational capacity.