problem solving what is ai way of solving problem?
Post on 19-Dec-2015
228 views
Embed Size (px)
TRANSCRIPT

Problem Solving
What is AI way of solving problem?

CS 331 Dr M M Awais 2
Problem Solving
Example
CAR BREAK DOWN

CS 331 Dr M M Awais 3
Problem Solving
Steps
• Analyze the problem
• Represent it
• Observe the present state
• Infer actions
• Solve to achieve the goal
Development Phase
Implementation Phase

CS 331 Dr M M Awais 4
Problem Solving
Analyze• Identify possible states• Identify possible actions• Identify possible goals• States, Actions, and Goals relate
through rules• Generate Rules that govern the
problem/or are helpful in searching the goal

CS 331 Dr M M Awais 5
Problem Solving
Analyze: Car Break Down
• Possible States (that you can observe or measure): –Air fuel ratio–Temperature–Engine sound–Fuel Consumption

CS 331 Dr M M Awais 6
Problem Solving
Analyze: Goals
• Reduction in Fuel Consumption
• High Pickup
• Low Coolant Temperature
• High Air-Conditioning Efficiency
• etc…

CS 331 Dr M M Awais 7
Problem Solving
Analyze: Actions
• Change condenser
• Change carburetor
• Get wheel alignment
• Etc…

CS 331 Dr M M Awais 8
Problem Solving
Analyze: RulesIf the air fuel (A/F) ratio is not correct (state observed)
THENTHENCarburetor needs adjustment (action identified)
WHICHWHICH Will achieve correct A/F ratio (Goal to be achieved)

CS 331 Dr M M Awais 9
Problem Solving
Represent Problem
• Convert the problem into a format that is understood by the computer system for automatic processing
• Numerous Techniques are available• Propositional /Predicate Calculus (formal)• Semantic / Conceptual Graphs (Graphical)• Search Trees (Graph Theory)• Rule Based/Expert Systems (Informal)• Learning Methods (Statistical/Symbolic)

CS 331 Dr M M Awais 10
Problem Solving
Represent: Example
IF the A/F ratio > 1.5
THEN change carburetor
IF coolant temperature > 250 c
THEN change condenser
MATLAB,
Prolog,
C++,
Lisp
As the course progresses we will study more of these representational techniques

CS 331 Dr M M Awais 11
Problem Solving
Implementation Phase
• Observe:– Use sensors/stored information/user input to get the
information about the present state within the problem domain.
• Infer:– Use above information to select appropriate actions that
will lead to the goal.
• Achieve Goal:– Apply actions in sequence/parallel to achieve goal

CS 331 Dr M M Awais 12
Problem Solving
Problem types
• Deterministic, fully observable
• Partial Deterministic and observable– Partial knowledge of states and actions

CS 331 Dr M M Awais 13
Problem Solving
Problem types
• Deterministic, fully observable
single state problem– know exactly present state;– solution is a sequence of known actions.

CS 331 Dr M M Awais 14
Problem Solving
Problem types
Partial knowledge of states and actions:• Non-observable sensorless or conformant problem
• May have no idea where one is; • solution (if any) is a sequence.• As if you are blind folded
• Nondeterministic and/or partially observable contingency problem
• Percepts provide new information about current state • solution is a tree or policy; often interleave search and execution.• States, actions and Rules of the environment are known, but current
state assessment is only possible after observation
• Unknown state space exploration problem (“online”)– When states and actions of the environment are unknown.

CS 331 Dr M M Awais 15
Problem Solving
More Examples
• Path Finding
• 8 Puzzle
• Robot Assembly
Primary FOCUS of discussion:
How to Analyze a given problem?

CS 331 Dr M M Awais 16
Problem Solving
Example: Romania

CS 331 Dr M M Awais 17
Problem Solving
Example: Romania
• On holiday in Romania; currently in Arad– Flight leaves tomorrow from Bucharest
• Formulate goal– Be in Bucharest
• Formulate problem– States: various cities– Actions: drive between cities
• Find solution– Sequence of cities; e.g. Arad, Sibiu, Fagaras, Bucharest, …
Identify Current StateIdentify GoalAnalyzeSolution

CS 331 Dr M M Awais 18
Problem Solving
Example: vacuum world
• Goal: Clean all the rooms
• Current State: 5• Single state, start in
#5. Solution??• Possible Actions

CS 331 Dr M M Awais 19
Problem Solving
Example: vacuum world
• Single state, start in #5. Solution??
[Movement, clean action]
[Right, Vacuum]

CS 331 Dr M M Awais 20
Problem Solving
Example: vacuum world• Single state
start in #5. Solution??[Right, Vacuum]
• Sensorless: start in {1,2,3,4,5,6,7,8} Solution??
• Contingency: start in {1,3}. Solution??

CS 331 Dr M M Awais 21
Problem Solving
Example: vacuum world
• States??• Initial state??• Actions??• Goal test??• Path cost??

CS 331 Dr M M Awais 22
Problem Solving
Example: vacuum world
• States?? two locations with or without dirt: 2 x 22=8 states.• Initial state?? Any state can be initial• Actions?? {Left, Right, Vacuum}• Goal test?? Check whether squares are clean.• Cost?? Number of actions to reach goal.

CS 331 Dr M M Awais 23
Problem Solving
Example: 8-puzzle
• States??• Initial state??• Actions??• Goal test??• Path cost??

CS 331 Dr M M Awais 24
Problem Solving
Example: 8-puzzle
• States?? Integer location of each tile • Initial state?? Any state can be initial• Actions?? {Left, Right, Up, Down}• Goal test?? Check whether goal configuration is reached• Cost?? Number of actions to reach goal

CS 331 Dr M M Awais 25
Problem Solving
Example: 8-queens problem
• States??• Initial state??• Actions??• Goal test??• Cost??

CS 331 Dr M M Awais 26
Problem Solving
Example: 8-queens problem
Incremental formulation vs. complete-state formulation• States?? • Initial state??• Actions??• Goal test??• Cost??

CS 331 Dr M M Awais 27
Problem Solving
Example: robot assembly
• States?? • Initial state??• Actions??• Goal test??• Cost??

CS 331 Dr M M Awais 28
Problem Solving
Example: robot assembly
• States?? Real-valued coordinates of robot joint angles; parts of the object to be assembled.
• Initial state?? Any arm position and object configuration.• Actions?? Continuous motion of robot joints• Goal test?? Complete assembly (without robot)• Cost?? Time to execute

CS 331 Dr M M Awais 29
Problem Solving
Summary
• Analyze:– States, Actions, Goals, Rules
• Represent: – Using any appropriate technique – Estimate the Cost factors (Time and Space Complexity)
• Implement:– Do actual coding/real world implementation– Validate and Test