problemsolving

1
problem-solving

• when you are trying to a achieve a goal, but the
  solution is not immediately obvious
• e.g., 2 X 5 ?
• not a problem, solution is immediately obvious
• e.g., 23 X 57 ? (1,311)
• problem, because answer not immediately obvious

2
Algorithms vs. heuristics

• some problems have guaranteed solutions (i.e.,
  algorithms)
• typically, math problems, physics problems,
  chemistry problems
• without algorithms, use heuristics (rules of
  thumb)
• how to spend your money wisely the most
  efficient driving route money to raise a happy,
  health baby finding a mate

3
parts of a problem

• clearly specified goal (goal-state) (aka,
  end-state)
• starting point (our current situation)
  (start-state) (aka, initial state)
• solution path the way to get from the
  start-state to the goal-state
• actions that we can take operators

4
unhappy unmarried
PROBLEM SPACE
start-state
operators
intermediate state
solution path
Be happy Be married Financially stable In love
goal-state
5
Problem space

• problem space all of the possible states,
  operators, and solutions for a problem (aka,
  search space)
• heuristics help us find a solution path tell us
  which actions to try but, not guaranteed to work

6
systematic random search

• trial and error try one solution
• if it works, youre done!
• if not, then you try another solution
• keep going until you find the solution
• not efficient could spend a lot of time trying
  to solve your problem
• random search try operators randomly (could
  possibly use the same attempt twice)
• systematic keep track of solutions that dont
  work

7
backward search

• problem-solving heuristic
• start (imagining) at the goal-state
• seek a path from the goal-state to the
  start-state
• imagine what (intermediate) state you were in
  just prior to (immediately before) you got to the
  goal-state

8
Other issues

• rules determine what are acceptable actions
  (operators)
• e.g., student code legal system
• must be extremely specific with your goal

9
Means-end analysis

• You are trying to change your current state to
  look as much as possible like the goal state
• 1. make a list of possible operators that could
  change our current state to look like goal state
• Evaluate each operator how far will it go
  towards goal?

10
More on m-e analysis

• 2. apply (use) operator that takes you the
  farthest toward your goal
• 3. go back to step 1 and do it over, starting at
  your new state
• 4. keep going until you have solved your goal