The Thinking Library

What does problem solving mean?

A problem arises when the route from a current state to a desired state is not immediately known. Some problems have clear goals, rules and correct answers. Others require first deciding whose needs to meet, what can change and what success means. A mathematical proof, a faulty machine and a long queue in an organization do not all call for the same kind of answer.

Before solving a problem, we need a useful representation of it. A complaint expressed in words may be represented with numbers, a diagram or a model of relationships. This selects what matters for a decision rather than capturing all of reality. Even an elegant model can mislead if we fail to notice what it leaves out.

How does the ability develop?

Both methods and subject knowledge matter. Algorithms, estimates, analogies and system models help in different settings. Part of expertise is recognizing which method applies and where its assumptions break down.

Reading explanations of worked examples, comparing different applications of a principle and checking your predictions against actual results help organize knowledge. Collecting answers alone leaves their reasoning unexplained. Nor is there a guarantee that applying the same four or five steps will solve every problem.

Experience changes how a problem is seen

A novice may group problems by visible objects or words. An expert may organize them by underlying principles of the field. Research comparing physics experts and novices provides one basis for understanding this difference. Knowing many facts and recognizing the relevant relationship when needed are different things.

Analogy: shared relationships, not just shared words

Using another problem’s solution requires understanding which relationships match. A hospital problem and an engineering problem may use different words but share a resource-allocation structure. An analogy can fail when important constraints differ. Comparison involves more than recalling a story.

Searching, selecting and reassessing

When the space of possible solutions is large, we cannot test every route. Simple cases, estimates and small models narrow the search. They save time, but we first need to know how accurate a result must be. Where precision matters for safety, a quick estimate cannot replace final verification.