Situational Control

Origin. Dmitry Pospelov and the Soviet AI school (1970s-1980s); developed at the Institute of Control Sciences (IPU), Moscow; distinct from Western AI's focus on search and logic.

Mechanism. Represents the world as a structured situation rather than as a state in a state space. A situation is a configuration of objects, relations, and context that has meaning as a whole; the same objects in different relations constitute different situations. Control proceeds by recognizing the current situation type, retrieving the appropriate response schema, and instantiating it. The method is knowledge-intensive rather than search-intensive: expertise is encoded as situation-action pairs, not as goal-means reasoning.

Procedure. Enumerate the situation types the system must handle. For each type, define the recognition criteria: what configuration of observables identifies this situation? For each type, define the response schema: what actions are appropriate, in what sequence, with what adaptations? When a new observation arrives, match against situation types. If a match is found, execute the schema. If no match, either default to a safe action or escalate to a higher level that can reason about novel situations.

Applies to. Expert systems, diagnostic systems, robotics, and any domain where expertise consists of pattern recognition followed by scripted response — troubleshooting, medical diagnosis, operational procedures.

Limitations. Brittle at situation boundaries. Real situations are continuous; the method discretizes them into types, and observations near type boundaries produce unstable recognition. Also: the enumeration of situation types is complete only for well-understood domains; novel situations fall through. The method encodes expertise but does not generate it; it cannot handle situations the designer did not anticipate.