What are the characteristics that make complex problem solving complex?

Question

In real-world problem-solving tasks that many people call "complex" (like flying a jet, programming, fixing a car, fighting a fire - the type investigated by the naturalistic decision making community) what are the key characteristics that separate problem solving in these types of tasks and problem solving in "toy" or "experiment" tasks where one or two stimuli are presented to participants?

Are there levels of complexity that can characterize the differences?

Answer 1

The complexity of the environment certainly is an issue. According to AI - A modern approach, the environment may be

• observable / partly observable
• deterministic / stochastic / strategic
• episodic / sequential
• static / dynamic
• discrete / continuous
• single agent / multiple agent

See also WikiDoc for a short overview. In my oppinion, these criteria apply also to human problem solving.

Another issue is the type of problem to be solved. A measure for complexity may be the number of steps that must be anticipated to solve the problem, e.g. solving a chess endgame is more complex than solving tic tac toe.