Questions: Constraint Satisfaction in Problem-Solving

Question 1

A person is stuck on the nine-dot problem, silently assuming that lines must stay inside the 3×3 grid of dots. When they suddenly realize there is no such rule and the solution appears, this 'aha moment' is best explained as:

• A. Completing an exhaustive search through all remaining possibilities
• B. Relaxing an incorrectly assumed constraint that was blocking productive search
• C. Randomly guessing a new approach after failing with deliberate strategies
• D. Increasing working memory capacity to hold more options simultaneously

Answer: (B)

The nine-dot problem is a classic insight problem where solvers impose a constraint — lines must stay within the grid — that is never stated. The 'aha' corresponds to recognizing and relaxing that false constraint, which instantly opens a region of solution space that was previously invisible. This is the constraint satisfaction framework's central account of insight: insight is not random, it is constraint relaxation.

Question 2

Two problems have equally large search spaces. Problem A has many explicit constraints with strong propagation (each constraint eliminates many candidates). Problem B has few constraints. Which is likely easier to solve, and why?

• A. Problem B — fewer constraints means fewer rules to keep track of
• B. Problem A — constraint propagation prunes the search space before costly search begins
• C. They are equivalent — search space size is the only determinant of difficulty
• D. Problem B — constraints create conflicts that slow the solver down

Answer: (B)

A common intuition is that more constraints make a problem harder. The constraint satisfaction framework reveals the opposite: constraints are the solver's friend. Each constraint eliminates candidates; propagating one constraint often triggers further eliminations. A heavily constrained problem like Sudoku is solvable precisely because propagation does most of the work. Difficulty is shaped by constraint structure, not just search space size.

Question 3

Adding explicit constraints to a design problem — formally writing down every requirement and asking what each one rules out — can make the problem easier to solve, not harder.

• T. True
• F. False

Answer: True

This is counterintuitive but central to the framework. Explicit constraints enable propagation: once you know that requirement X rules out options A, C, and F, your search space shrinks immediately. Problems that feel intractable often become tractable when constraints are made explicit and propagated systematically. The act of clarifying requirements is itself a problem-solving move.

Question 4

Human insight in problem-solving occurs because people systematically and exhaustively search the problem space until they stumble upon the correct solution.

• T. True
• F. False

Answer: False

Research shows that human problem-solvers do not exhaustively search problem spaces — they maintain implicit constraint representations that prune options before conscious deliberation. Insight occurs not at the end of exhaustive search but when a faulty constraint is relaxed, suddenly revealing previously blocked solution paths. Insight feels sudden because the pruned region of solution space becomes accessible all at once.

Question 5

Why does making constraints explicit often make a problem easier to solve, even though it might seem like more rules would add complexity?

Model answer: Explicit constraints enable propagation: each constraint eliminates candidates, and satisfying one constraint often cascades into eliminating candidates for other variables. This pruning reduces the search space that must be explored by guessing or backtracking. Implicit or vague constraints cannot propagate — they sit inert while the solver wastes effort on already-ruled-out options.

The key insight is that constraints and search are in tension: more constraints mean less search, not more. Problems with rich constraint structure — where constraints link many variables and propagate widely — can often be solved with little search at all. The work is done by propagation rather than by trial and error.