Documentos de Académico
Documentos de Profesional
Documentos de Cultura
AI
Fuzzy Logic
Neural Nets
11/6/2001
Fuzzy Logic
Philosophical approach
Ontological commitment based on degree of truth
Is not a method for reasoning under uncertainty
11/6/2001
Control of a vehicle
Should slow down when close to car in front
Should speed up when far behind car in front
Fuzzy Sets
Classical set theory
An object is either in or not in the set
Membership
(Degree of Truth)
Meek
1.0
Nasty
Medium
-1
-0.5
11/6/2001
0.0
0.5
Aggressiveness
FS
1.0
FS
0.0
Units
11/6/2001
About 6
Tall
0.0
Height
11/6/2001
About 6
Tall
0.0
Height
11/6/2001
Fuzzy Rules
If our distance to the car in front is small, and the distance is
decreasing slowly, then decelerate quite hard
11/6/2001
11/6/2001
small perfect
big
v. big
brake
slow
distance
<<
<
>
>>
delta
11/6/2001
present
fast
fastest
acceleration
small perfect
big
v. big
brake
slow
distance
<<
<
>
>>
????
fast
fastest
acceleration
delta
11/6/2001
present
slow
acceleration
fast
11/6/2001
acceleration
present
acceleration
acceleration
Defuzzification Example
We have three things sets we have reason to believe we are
in, and each set covers a range of values
Two options in going from current state to a single value:
Mean of Max: Take the rule we believe most strongly, and take the
(weighted) average of its possible values
Center of Mass: Take all the rules we partially believe, and take their
weighted average
11/6/2001
Disadvantages
Sometimes results are unexpected and hard to debug
Additional computational overhead
There are other ways to get continuous acceleration
11/6/2001
References
Nguyen, H. T. and Walker, E. A. A First Course in
Fuzzy Logic, CRC Press, 1999.
Rao, V. B. and Rao, H. Y. C++ Neural Networks
and Fuzzy Logic, IGD Books Worldwide, 1995.
McCuskey, M. Fuzzy Logic for Video Games, in
Game Programming Gems, Ed. Deloura, Charles
River Media, 2000, Section 3, pp. 319-329.
11/6/2001
Neural Networks
Inspired by natural decision making structures (real nervous
systems and brains)
If you connect lots of simple decision making pieces
together, they can make more complex decisions
Compose simple functions to produce complex functions
Neural networks:
11/6/2001
Wi,j
Wi,j aj = ini
ai
ai = g(ini)
11/6/2001
Network Structure
Single perceptron can represent AND, OR not XOR
Combinations of perceptron are more powerful
11/6/2001
Dead
Sound
Low Health
Wander
Retreat
11/6/2001
Spawn
Chase
11/6/2001
0.1
Wj aj = 0.1
0
g(0.1) = 0
11/6/2001
0.6
Wj = Wj + (t-o)aj
W1 = 0.1 + 0.1(1-0)1 = 0.2
W2 = 0.6 + 0.1(1-0)0 = 0.6
After this step, try (0,1)1 example
No error, so no training
0
0.2
Wj aj = 0.6
1
g(0.6) = 0
0.6
11/6/2001
0.2
Wj aj = 0.2
0
g(0.2) = 0
11/6/2001
Disadvantages
11/6/2001
References
11/6/2001