Is it possible to design a multiple-level inference? for example
if a is A and b is B and c is C then d is D
if e is E and f is F then g is G
if d is D and g is G then h is H
if i is I and h is H then j is J
or am I restricted to levels?
From fuzzylite 5 you can do that. However, if you want to implement a more complex algorithm for conflict resolution (or similar) you can create activation methods in fuzzylite 6 and put your logic there.
I tried it in both versions. Sets consider this:
if A is a and B is b then C is c
if D is d and E is e then F is f
if C is c and F is f then G is g
The last inference level (if C is c and F is f then G is g) does not produce any results for me neither higher levels of inference algorithm. What should I take care of?
Ι think I solved my problem. But I have two more questions:
a. In surface window I just get a black background
b. What are the criteria for implication, activation, conjuction, aggregation and defuzzifier selection?
it is not easy to find out your problem unless you post the FLL of your engine.
My guess is that you are using conjunctions in the rules (e.g., x is X *and* y is Y), and you are not specifying any conjunction operator. You need to have conjunction operators when using “and” in rules, and disjunction operators when using “or” in rules.