Using Simulink | Simulink | Help Desk |
Help for SymbInfMam |
FlouLib |
This block implements a
Mamdani inference with a linguistique representation (or symbolic) of the rulebase.
The input
vector is generally the result of a symbolic fuzzification but it can also be
the output of another block producing a symbolic output (for example another SymbInfMam block),
thus allowing the chaining of rules.
The block
output is a Membership degree vector. The degrees are stored according to the
order in the file declaration. The last vector component is a number associated
with each block which allows the propagation of some information.
It
describes the rulebase and is formed as follows::
The
line "A-1 B-1 C-2 0.2" corresponds to the rule:
If Input1 is A-1 and Input2 is B-1 the Output is C-2 weight 0.2
An example
of file is as follows:
2
5
A-1 B-1 C-2
A-1 B0 C-1
A-1 B1 C0
A0 B-1 C-1
A0 B0 C0
A0 B1 C1
A1 B-1 C0
A1 B0 C1
A1 B1 C2
·
Projection
operator
The
projection operator is a T-conorm generalizing the sup.
·
Combinaison
operator
The
combinaison operator is a T-norm.
·
Premises
aggregation operator
The
premises aggregation operator is a T-norm.
The conventional T-norms and
the T-conorms have the following codes:
T-norms |
T-conorms |
1 -> Min |
-1 -> Max |
2 -> Product |
-2 -> Probabilistic Sum |
3 -> Lukasiewicz |
-3 -> Lukasiewicz |
Other T-norms or T-conorms
can be added in the UserOperators.c file of the FlouLib library.
After modification FlouLib must
be updated by the instruction: install -Update.