Question

well i’m having a hard time working with prolog (gprolog). My problem is this, I’ve the next predicate:

prob(factA,factB,X,first):-   X=<50.
prob(factA,factB,X,second):-  X>50,X=<80.
prob(factA,factB,X,none):-  X>80,X=<100.

that is factA have 50% of probability of occur, factB a 30%, and lastly 20% of non of the facts occur. Also i’ve others predicate where change the factB but not factA, also change the X intervals.

Now i have the functor:

result(X,Y,N):- random(1,100,I),
  prob(X,Y,I,N).

this predicate tells what is N for X an Y in prob. This predicate handle the case:

result(factA,factB,N).

but now i have to handle:

result(factB,factA,N).

and it will return a valid result, any idea?, i did something like this:

result(X,Y,N):- Y = factA,
     random(1,100,I),
     prob(Y,X,I,N).
            J is 100-I,
     prob(Y,X,J,N).

but this doesn’t manage well the ‘none’ result.

Somebody has any idea?

Thanks!

Answer 1

Looking at my comment above: if first is 50% (and with some optimisation if you “know” random value will be between 1 and 100):

prob(X,first):- X=<50, !.
prob(X,second):- X=<80, !.
prob(X,none).

I guess this is not what you meant since it does not depend on facts at all, so they’re not needed as arguments.

Thus, assuming factA is 50%, a bit more complex but basically the same idea as yours:

fact_prob(factA, 50).
fact_prob(factB, 30).

random_fact(X, _, I, R) :-
  fact_prob(X, XP),
  I <= XP, !,
  R = first.
random_fact(X, Y, I, R) :-
  fact_prob(X, XP),
  fact_prob(Y, YYP),
  XYP is XP + YP,
  I <= XYP, !,
  R = second.
random_fact(_, _, _, none).

Disclaimer: I never worked with gprolog, you might need to adjust details to your dialect.