Docstoc

slides - UBC - Computer Science

Document Sample
slides - UBC - Computer Science Powered By Docstoc
					     Computer Science CPSC 322


           Lecture 12
Constraint Satisfaction Problems
             (4.1-4.4)
            Oct 4, 2010
            Announcements

• Assignment due on Friday
  • Remember to USE HANDIN!

• Check the Discussion Board: some of your
  questions may have already been answered
  there



                                       Slide 2
             Lecture Overview


• Recap of Lecture 11
• Intro to CSP
  - Variables/Features
  - Constraints
• Algorithms


                                Slide 3
             Recap: Search Wrap Up
          Selection      Complete   Optimal   Time    Space
DFS         LIFO            N         N       O(bm)   O(mb)

BFS         FIFO            Y         Y       O(bm)   O(bm)
IDS         LIFO            Y         Y       O(bm)   O(mb)
LCFS       min cost        Y **      Y **     O(bm)   O(bm)

Best        min h           N         N       O(bm)   O(bm)
First
 A*         min f          Y**       Y**      O(bm)   O(bm)
B&B     LIFO + pruning      N         Y       O(bm)   O(mb)
IDA*        LIFO            Y         Y       O(bm)   O(mb)

MBA*        min f          Y**       Y**      O(bm)   O(bm)


** Needs conditions
                  Standard Search VS. CSP
• To learn about search we have used it as the reasoning
  strategy for a simple goal-driven agent.
• Standard search problem: An agent can solve a
  problem by searching in a space of states
• State is a “black box” : any arbitrary data structure that
  supports three problem-specific routines
   – goal test: goal(state),
   – heuristic evaluation function: h(state)
   – finding successor nodes: neighbors(state)
• In the next few weeks, we’ll see more specialized
  versions of search for three problem categories in
  deterministic settings
                    Course Overview
                                                       Representation
                               Environment
                                                       Reasoning
                        Deterministic  Stochastic       Technique
Problem Type
                             Arc
                        Consistency           We’ll start
       Constraint Vars +
       Satisfaction ConstraintsSearch         from CPS
Static
                                       Belief Nets
        Query       Logics             Variable
                            Search
                                       Elimination
                                       Decision Nets
Sequential           STRIPS
                                        Variable
      Planning                          Elimination
                           Search
                                       Markov Processes
                                        Value
                                        Iteration
Standard Search vs. Specific R&R systems
Constraint Satisfaction (Problems):
    • State
    • Successor function
    • Goal test
    • Solution
Planning :
    • State
    • Successor function
    • Goal test
    • Solution
Inference
    • State
    • Successor function
    • Goal test
    • Solution
  CPSC 322, Lecture 11                Slide 7
      Constraint Satisfaction Problems (CSP)
• In a CSP
  – state is defined by variables/features Vi with values from
    domain Di
  – goal test is a set of constraints specifying
    (1) allowable combinations of values for subsets of variables
    (hard constraints)
    (2) preferences over values of variables (soft constraints)
      Constraint Satisfaction Problems (CSP)
• Allow useful general-purpose algorithms with more
  power than standard search algorithms

• They exploit the multi-dimensional nature of the
  problem and the structure provided by the goal




  set of constraints, *not* black box.
             Lecture Overview


• Recap of Lecture 11
• Intro to CSP
  - Variables/Features
  - Constraints
• Algorithms


                                Slide 10
    Dimensions of Representational Complexity
                 (from lecture 2)
• Reasoning tasks (Constraint Satisfaction / Logic&Probabilistic
  Inference / Planning)
• Deterministic versus stochastic domains
Some other important dimensions of complexity:
• Explicit state or features or relations
  Explicit state or features or relations
•   Flat or hierarchical representation
•   Knowledge given versus knowledge learned from experience
•   Goals versus complex preferences
•   Single-agent vs. multi-agent
       Explicit State vs Features (lect. 2)
How do we model the environment?
• You can enumerate the states of the world.
   • As we have done in the standard search so far, or
• A state can be described in terms of features/variable
  • Often it is more natural and concise to describe states in
       terms of assignments of values to features (variables).
   •   30 binary features (also called propositions) can represent




                230=1,073,741,824 states
    Variables/Features and Possible Worlds
• Variables / features
   • we denote variables using capital letters
   • each variable V has a domain dom(V) of possible values

 • Variables can be of several main kinds:
    • Boolean: |dom(V)| = 2
    • Finite: the domain contains a finite number of values
    • Infinite but Discrete: the domain is countably infinite
    • Continuous: e.g., real numbers between 0 and 1

   • Possible world: a complete assignment of values to a set of
      variables
        This is equivalent to a state for the time being, but the term “possible
         world” is introduced for consistency with later topics, when we will
         need the distinction between these two terms
              Example (lecture 2)
Mars Explorer Example
Weather               {S, C}
Temperature         [-40, 40]
Longitude            [0, 359]
Latitude             [0, 179]

One possible world (state)            {S, -30, 320, 210}


Number of possible (mutually exclusive) worlds (states)
                 2 x 81 x 360 x 180
          Examples: Crossword Puzzle
• variables are words that have to be
    filled in
      • places where words can go e.g. (1
        across, 3 down)
• domains are valid English words of
    required length
•   possible worlds: all ways of assigning
    words
    • Number of English words?
    • Number of words of length k ?



    • So, how many possible worlds?
             Lecture Overview


• Recap of Lecture 11
• Intro to CSP
  - Variables/Features
  - Constraints
• Algorithms


                                Slide 16
                            Constraints
• Constraints are restrictions on the values that one or
  more variables can take
  • Unary constraint: restriction involving a single variable
  • k-ary constraint: restriction involving the domains of k
      different variables
       • we'll mainly talk about binary constraints
• Constraints can be specified by
    • giving a function that returns true when given values for
      each variable which satisfies the constraint
    • giving a list of valid domain values for each variable
      participating in the constraint
• E.g. if we have domain variables X1 and X2, both with domain
  {a, b}, we can specify the constaint “X1 and X2 must have
  different values as
    • X1 ≠ X2 or [(a, b) (b, a)]                                Slide 17
                               Examples
• Crossword Puzzle:
   • variables are words that have to be filled in
   • domains are valid English words
   • constraints: words have the same
     letters at points where they intersect




                                                     Slide 18
            Example: Map-Coloring
• Map Coloring Problem
  • variable: regions on the map
  • domains: possible colors
  • possible worlds: color assignments for each region


• how many possible worlds?




                                                         Slide 19
Example: Map-Coloring




Variables WA, NT, Q, NSW, V, SA, T
Domains Di = {red,green,blue}
Constraints: adjacent regions must have different colors
e.g., WA ≠ NT,

Or
(WA,NT) in
                  Example CSP (1)

• The delivery robot needs to schedule 5 activities,
  A,B,C,D, E

• Each activity happens at either time 1, 2, 3, 4

• Constraints
   – B cannot start at 3, C cannot start at 2 (unary constraints)
   – A cannot start at the same time as B, and B cannot start as
     the same time as C or D (binary constraints)
   – C must start before D, E must start before any other
     activity (binary constraints)
   – A must start as the same time as D (binary constraints)
Example: scheduling activities
                    Example 2
• Eight Queen problem: place 8 queens on a
  chessboard so that no queen can attack the others




• Constraints: No queens can be in the same row,
  column or diagonal
               Example 2: 8 queens
• Variables: V1 ,.. Vn.
  Vi = Row occupied by the ith queen in the ith column

• D Vi = {1,2,3,4,5,6,7,8}

• We can specify the constraints by enumerating
  explicitly, for each pair of columns, what positions
  are allowed. Ex:
   Constr(V1 , V2) = {(1,3), (1,4),..(1,8)(2,4)(2,5)…(8,6)}
          Constraint Satisfaction Problems
Definition (Constraint Satisfaction Problem)
A constraint satisfaction problem consists of
   • a set of variables

    • a domain for each variable

    • a set of constraints

Definition (model / solution)
A model of a CSP is an assignment of values
  to variables that satisfies all of the
  constraints.
            Example: Map-Coloring




Models / Solutions are complete and consistent assignments,
  e.g., WA = red, NT = green, Q = red,   NSW = green, V =
  red,SA = blue, T = green

                                                            Slide 27
Constraint Satisfaction Problem: Variants
We may want to solve the following problems using a CSP
• determine whether or not a model exists
•   find a model
•   find all of the models
•   count the number of models
•   find the best model given some model quality
    • this is now an optimization problem
•   determine whether some properties of the variables hold in all
    models
                  Solving FCSPs

The finite constraint satisfaction problem is NP-hard.
(exponential in the size of the variables’ domains)
We can:

• Try to find algorithms that work well on typical cases
  even though the worst case may be exponential

• Try to find efficient approximation algorithms
                 Next Class


• Recap of Lecture 11
• Intro to CSP
  - Variables/Features
  - Constraints
• Algorithms


                              Slide 30
         Learning Goals for today’s class

• Define possible worlds in term of variables and
  their domains.
• Compute number of possible worlds on real
  examples
• Specify constraints to represent real world
  problems differentiating between:
   • Unary and k-ary constraints
   • List vs. function format.
Verify whether a possible world satisfies a set of
  constraints (i.e., whether it is a model, a solution)
CPSC 322, Lecture 4                                       Slide 31

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:5
posted:4/26/2011
language:English
pages:30