Problem Solving (1) Flashcards
What is meant by General Problem Solver?
An information processing approach to solving problems - a computer programme
What is meant by Means-end Analysis?
A technique used to solve problems that considers your current state to where you want to end up at/be (the desired state)
What is meant by Well/ill-defined Problems?
Problems that have a clear end people (or less clear end-point)
What is a problem?
- A situation where your current state isn’t the same (or good enough) as your goal state
What are the types of problems?
Lab problems
Real-world problems
Well-defined problems
Ill-defined problems
What is meant by action selection and outcome evaluation?
Action selection = you don’t have to attempt a problem at all
Outcome evaluation = how do you decide if the problem is over?
What is Thorndike’s law of effect?
Thorndike (1898 - 1911):
- keep doing satisfying things, stop doing dissatisfying things
- used trial and error approach
What is imagery?
- Pinker’s ‘mentalese’ - visualise the world
- mental imagery
What is the mental rotation task?
Shepard Melzer task:
- Do the shapes match?
- imagine what the shapes look like at different angles
- (mental rotation)
- looked at eye movement and compared them to the mental rotation only task = have to solve SM with mental rotation
Explain Gestalt Pscyhology
Gestalt Psychology follows ‘the whole is greater than the sum of its parts’ - a holistic viewpoint
- Productive = come up with a new idea to solve the problem
- Reproductive = analyse the problem, apply existing knowledge to solve the problem
What is Problem Solving?
- Governed by recentering something which originates in the desire to get comprehensive insight
- brings transformation, seeing things as parts of a new, clear structure
- 2 direction are involved in problem solving: getting a whole consistent picture and seeing what the structure of the whole requires for the parts - Wertheimer, (1945)
What did Kohler do/study with Chimpanzees?
- Sultan (a chimpanzee)
- Kohler got Sultan to solve novel problems in order to get a food reward
- Sultan couldn’t reach it
- So restructured the problem by using a box to stand on
- Is this evidence for productive thinking?
What evidence is there for Gestalt Theory?
Pask et al., (1967):
- Students given cards to sort, using any method they want to
- some detailed descriptions(/descriptors) were given and others were general statements
- Answer/Findings = hierarchy thinking, like biological classification when solving problems/sorting the cards
Additional findings:
- some start with general statements and fill in the descriptors later
- others work from the detailed descriptors up to the general statements
What some the limitations to Gestalt Theory?
How is a bad structure restructured into a better one?:
- Incomplete mechanism
- “self-organising properties of the cortex”
How does this work?:
- Based on simple lab tasks
But are real-world problems that simple and easy to solve?
What is information processing?
- WWII: technological advances e.g. fight tanks, artillery, training needs code breaking all required the investigation of human cognition
- J.J. Gibson: training pilots in WWII led to his theories about direct perception and ecological visual processes
- during WWII - German U-boats communicated via radio signals using coding device called the enigma machine
- British and American mathematicians, mainly Alan Turing (1912- 1954) developed and broke then enigma code to understand it (and win the war)
What is the Turing Machine?
How does a machine solve problems?
- Turing = famous mathematician
- Problem = crack morse codes (of the German’s) during WWII
- Solution = built a machine to do it (Turing machine)
- A finite state machine
What is the problem space theory?
Newell and Simon (1972):
- The more the steps = the harder the problem
- navigate through the steps of the problem
- e.g. Donald + Gerald = Robert
What is the maze analogy?
- Problem = get to the middle
- Initial state = entrance
- Goal state = middle
- Operators = the possible moves needed to get to the middle/solve the problem
- (Means-end analysis?)
What is the information processing approach?
- Problem = the current state isn’t the goal state
If we assume: - Novel problem (no prior experience)
- clear rules
- well-defined start and end point
- non-adversarial - when solving the/a problem, it isn’t against an opponent
- How does a machine solve problems? - (computational approach)
What is the general problem solver (GPS)?
Newell & Simon (1972):
Aim: asked people to think out loud to work out the strategy they were using to solve problems
Procedure:
- Researchers wrote a computer program to solve problems, based on the strategies used by humans
Findings:
- The result = the ‘General Problem Solver (GPS)’
- they define a goal - the solution to the problem
- they define the current state - where you are now
- they define the mental operators - possible moves to get to the goal state
What is the Tower of Hanoi (study)?
Operator:
- Only 1 disk can be moved at a time
- you can only move the top disc in a stack of discs
- you can’t put a larger disc on top of a smaller disc
What is ‘Means-end Analysis’?
- Look at where you currently are, where you want to be and how to get there
- problem solving is a production of knowledge states by the application of mental operators
- the operators encode legal moves and restrictions disallowing illegal moves
- there may be alternative/multiple paths from an initial state to a goal state
What is meant by depth-first search?
- Don’t just scroll through all the potential move available (even though this algorithm is guaranteed to reveal the answer)
- Instead our working memory limitations might focus us to try a limited series of moves
- you may do a depth-first search as for example - enigma machine has 158+ million settings and would take too long/a long time to check them all so something else is needed
Depth VS Breadth approaches:
Depth: Breadth:
- +ive: faster - +ve: guarantees success
- +ve: often successful - -ive: requires more
effort/cognitive resources
- +ve: requires less effort - -ive: usually slower
- -ve: doesn’t guarantee success
What is meant by navigating problem space?
- People use a combination of their own knowledge and various heuristics to create new knowledge states and thereby solve problems
- all of these processes are subject to the limited space that they are situated in (i.e. human problem solvers have working memory constraints)
