Practicum agenttechno 3 - PowerPoint PPT Presentation

1 / 15
About This Presentation
Title:

Practicum agenttechno 3

Description:

you must construct a given number out of four other numbers ... C = {C(Ru)| i,j [1,n], C(Ru) = { b,c |b Di, c Dj, bc, i-jb-c, i-jc-b}} CSP- ERA ... – PowerPoint PPT presentation

Number of Views:27
Avg rating:3.0/5.0
Slides: 16
Provided by: BART194
Category:

less

Transcript and Presenter's Notes

Title: Practicum agenttechno 3


1
Practicum agent-techno 3
  • just some riddles

2
riddle 1
  • you must construct a given number out of four
    other numbers
  • use addition, subtraction, multiplication, and
    division operations
  • Each number must be used in the calculations
    exactly once
  • multiple solutions are possible

3
Example
  • 42 out of 2, 3, 4 and 9
  • Solution 1
  • (9 x 4) (2 x 3)
  • Solution 2
  • ((9 x 2) - 4) x 3
  • Solution 3
  • (4 (2 / 3)) x 9
  • 24 out of 1,3,4, and 6
  • ...

4
n queens problem
5
Constraint staisfaction problem (CSP)
  • consists of
  • A finite set of variables, X X1, X2, ... ,Xn
  • A domain set, containing a finite and discrete
    domain for each variable
  • D D1, D2,..., Dn, ? i?1,n, Xi ? Di
  • A constraint set, C C(R1), C(R2),... ,C(Rm),
    where each Ri is an ordered subset of the
    variables, and each constraint C(Ri) is a set of
    tuples indicating the mutually consistent values
    of the variables Ri.

6
  • n queens X X1, X2, ... ,Xn
  • n x n chessboard D D1, D2,..., Dn,
  • ?i, Di 1,2,...,n
  • Placement requirement
  • C C(Ru) ?i,j ? 1,n, C(Ru)
  • ltb,cgtb ? Di, c ? Dj, b?c, i-j?b-c, i-j?c-b

7
CSP-gtERA
  • set of variables, X set of Agents, A
  • domain set, D environment, E
  • constraint set, C set of reactive rules, R

8
initialisation
X1
X2
X3
X4
violation
9
initialisation
X1
X2
X3
X4
violation
Place a1 at (3,1)
10
initialisation
X1
X2
X3
X4
violation
Place a1 at (3,1)
11
Moves
  • Random move
  • In the beginning
  • When totally stuck
  • Least move
  • agent moves to minimum position
  • if more than one make a choice
  • Better move (not necessary)
  • agent moves to random better position
  • less efficient, but lower computational cost

12
Moves
Random move
13
Moves
Least move
Better move
14
tips
  • better-least-random move first try 'better
    move', if it doesn't found a solution (iidfas),
    try 'least move', iidfas, try 'random move'.
  • make a sketch of your agent's communication
  • does everyone has to move every step?
  • if not, who moves?
  • take care of computation time

15
Communication
  • start what happens?
  • first step?
  • step N to step N1?
  • when is it finished?
  • case M of ...then
  • else Browse 'im programmed nonsens'
  • end
Write a Comment
User Comments (0)
About PowerShow.com