Submitting abductive queries

After the abductive system is loaded and the abductive theory is imported, diagnosis tasks can be performed as abductive queries. The predicate for a query is query(+ListOfGoals, -Answer). Upon the success of the query, Answer is a tuple (As, Cs, Ns), where As is a list of (possibly non-ground) abducibles, Cs is a list of arithmetic constraints (which will be introduced in Section 3), and Ns is a list of collected dynamic denial of the form fail(ListOfUniversalVars, ListOfGoals). A dynamic denial, e.g., fail([X], [p(X, Y), q(X), r(Y)]), should be read as $\exists Y \forall X . \leftarrow p(X,Y), q(X), r(Y)$.

Let us go back to the car diagnosis example. To find out why my car does not start, you can execute the following query:

| ?- query([car_doesnt_start(mycar)], Ans).
Ans = ([has_no_fuel(mycar)],[],[fail([_A],[battery_flat(_A),lights_go_on(_A)]),fail([_B],
[has_no_fuel(_B),\+fuel_indicator_empty(_B),\+broken_indicator(_B)])]) ? ;
no
| ?-

Alternatively, if you are only interested in the collected abducibles and constraints, or if you also want to find out why a different car does not start, you can execute the queries:

| ?- query([car_doesnt_start(mycar)], (As, Cs, _)).
As = [has_no_fuel(mycar)],
Cs = [] ? ;
no
| ?- query([car_doesnt_start(yourcar)], (As, Cs, _)).
As = [battery_flat(yourcar)],
Cs = [] ? ;
As = [broken_indicator(yourcar),has_no_fuel(yourcar)],
Cs = [] ? ;
no
| ?-

Finally, if you want a diagnosis of any car, you can execute the query with non-ground goal(s), e.g.,

| ?- query([car_doesnt_start(AnyCar)], (As, Cs, _)).
As = [battery_flat(AnyCar)],
Cs = [AnyCar=/=mycar],
AnyCar=/=mycar ? ;
As = [has_no_fuel(mycar)],
Cs = [],
AnyCar = mycar ? ;
As = [broken_indicator(AnyCar),has_no_fuel(AnyCar)],
Cs = [AnyCar=/=mycar],
AnyCar=/=mycar ? ;
no
| ?-