Problem Solving and Reasoning Flashcards
Dunker (1926) definition of Problem Solving
A problem arises when a living organism has a goal, but does not know how this goal is to be reached.
Solution is not in memory
Goal directed
Mayer (1992) definition of Problem Solving
A sequence of mental processes or operations performed on information in a subject’s memory.
Behaviour due to cognitive processes
Tautological Problems
Always true
2 Research Methods of Studying Problem solving
- Protocol analysis:
Ask people how they solve a problem
Record ‘thinking aloud - Computer simulations
Well-defined Problems
Clear start state, goal state and operators
3x = 6, what is x?
Real world problems messy
Find basic principles from easy tasks
Generalise to complex problems
Ill-defined Problems
Lack the qualities of well defined problems
More like real life
Search strategies not used
Harder to theorise about
State Space Diagram
Well-defined Problems No knowledge necessary Solved quickly Can be systemically changed Tower of Hanoi
Solving Well-defined problems
1 Prune the state space
2 Exploratory strategies
3 Hill climbing
Prune the state space - Apply heuristics Exploratory strategies - Weak heuristics e.g. avoid loops Hill climbing - Pick a move that takes you closer to goal
Means-Ends analysis - Newell & Simon (1972)
ToH can be broken into subgoals Evaluate task: How do start & goal differ? Apply operator to reduce difference If obstacle prevents this – re-evaluate Create subgoal: remove obstacle Solve subgoal Set/solve next subgoal if necessary HOWEVER: Not all problems have clear subgoals & ID's
Problem Solving by Analogy - Mayer (1992)
Analogical reasoning occurs when we abstract a solution strategy from a previous [base] problem and relate that information to a new [target] problem that we are trying to solve. Indeed, we engage in analogical reasoning when we solve a new problem by using what we know about a related problem that we can solve.
Isomorphic Problems - Hayes & Simon, 1977
Why is the ToH problem easier to solve than the Monsters and Globes problem? (3 reasons)
- Harder to learn rules for change problem
Change rules – 3.5 mins
Move rules – 1.5 mins - Illegal moves not clear for change problem so harder to verify
Change rules – 5.7 secs
Move rules – 4.7 secs - Rules more consistent with real world knowledge for move problem
easier to learn
Easier to spot illegal moves
Why does external representation affect the difficulty of problems?
External representation affects difficulty Changes load on working memory If WM load is raised… If options are harder to identify… …then planning ahead is prevented
Duncker’s Radiation Problem can be solved how?
By analogy. Map the features of the fortress story [base] onto this one [target] and infer missing detail.
Problem Solving Schemas - Gick & Holyoak (1983)
Crit problem solving by analogy
Schema = abstract description of common features & general principles
Domain specific
Schema make transfer easier
But:
Multiple stories needed to create schema
Target problem must be similar → activation
Features of Inductive Reasoning
Inferring rules
Making generalisations
Detecting order out of chaos
Cannot be valid, but can be plausible