On Well- and Ill-Defined Problems
December 22, 2021
The concepts of well-defined and ill-defined problems are of fundamental importance in design research. This is because most non-trivial design problems belong to the class of ill-defined problems.
Well-Defined Problems
Simon (1977) formulated a list of six requirements that a problem must satisfy in order to be considered a well-structured one:
- Objective solution tests exist.
- A problem space (set of states) exists for the problem.
- Transitions between states can be represented in the problem space.
- New knowledge can be represented.
- If the problem solver acts in a physical world, a problem space for the physical world exists as well.
- All five precedings items are practically computable.
Well-defined problems are also called “tame” because they are structured in advance. Here, “structured” means that all its constituent states and feasible solving strategies are known in advance.
Ill-Defined Problems
“Ill-defined” problems are now defined simply as a residual of well-defined problems. That is, we say that a problem is ill-defined when it is not well-defined. For a readable introduction into this topic, see Simon (1977).
Comparison
The following table (Raymond, 1990) provides an informative side-by-side comparison of well-defined and ill-defined problems.
Well-defined problems | Ill-defined problems | |
---|---|---|
Problem specification: | complete, unambiguous | incomplete, ambiguous |
Solution evaluations: | definite criteria | no definite criteria (i.e., no stopping rule) |
Knowledge utilized: | represented in one or more problem spaces (in advance of problem solving) | many sources (i.e., problem spaces) that cannot be determined in advance and need to be integrated |
Operators: | enumerable and all known in advance | no exhaustive, enumerable set |
Solution path: | exists and is predetermined (but has to be found) | no predetermined solution path exists |
Examples: | chess, theorem-proving | design, planning, document composition |
References
-
Herbert A Simon. The structure of ill-structured problems. In Models of discovery, pages 304–325. Springer (1977).
-
Raymonde Guindon, Knowledge exploited by experts during software system design, International Journal of Man-Machine Studies, 33(3):279–304, 1990.