Information Technology Reference
In-Depth Information
Problems modeled by using CO-Net has the form (O, F, Goal), in which O is a set of
Com-objects, F is a set of facts on objects, and Goal is a set consisting of goals.
The basic technique for designing deductive algorithms is the unification of facts. Based
on the kinds of facts and their structures, there will be criteria for unification proposed.
Then it produces algorithms to check the unification of two facts. For instance, when we
have two facts
fact1
and
fact2
of kinds 1-6, the unification definition of them is as
follows: fact1 and fact2 are unified they satisfy the following conditions
1.
fact1
and
fact2
have the same kind k, and
2.
fact1
=
fact2
if k = 1, 2, 6.
[
fact1
[1],
fact1
[2..nops(
fact1
)]] = [
fact2
[1],
fact2
[2..nops(
fact2
)]] if k = 6 and the
relation in
fact1
is symmetric.
lhs(
fact1
) = lhs(
fact2
) and compute(rhs(
fact1
)) = compute(rhs(
fact2
)) if k =3.
( lhs(
fact1
) = lhs(
fact2
) and rhs(
fact1
) = rhs(
fact2
)) or
( lhs(
fact1
) = rhs(
fact2
) and rhs(
fact1
) = lhs(
fact2
)) if k = 4.
evalb(simplify(expand(lhs(
fact1
)-rhs(
fact1
)- lhs(
fact2
)+rhs(
fact2
))) = 0) or
evalb(simplify(expand(lhs(
fact1
)-rhs(
fact1
)+ lhs(
fact2
)-rhs(
fact2
))) = 0) if k = 5.
To design the algorithms for reasoning methods to solve classes of problems, the
forward chaining strategy can be used with artificial intelligent techniques such as
deductive method with heuristics, deductive method with sample problems, deductive
method based on organization of solving methods for classes of frame-based problems.
To classes of frame-based problems, designing reasoning algorithms for solving them is
not very difficult. To classes of general problems, the most difficult thing is modeling
for experience, sensible reaction and intuitional human to find heuristic rules, which
were able to imitate the human thinking for solving problems. We can use Com-Nets,
CO-Nets, and their extensions to model problems; and use artificial intelligent
techniques to design algorithms for automated reasoning. For instance, a reasoning
algorithm for COKB model with sample problems can be briefly presented below.
Definition 4.1
: Given knowledge domain K = (C, H, R, Ops, Funcs, Rules),
knowledge sub-
domain of knowledge K is knowledge domain which was represented by COKB model, it
consists of components as follows
K
p
= (C
p
, H
p
, R
p
, Ops
p
, Funcs
p
, Rules
p
)
where, C
p
C, H
p
H, R
p
R, Ops
p
Ops, Funcs
p
Funcs, Rules
p
Rules.
Knowledge domain K
p
is a restriction of knowledge K.
Definition 4.2
: Given knowledge sub-domain K
p
, Sample Problem (SP) is a problem which
was represented by networks of Com-Objects on knowledge K
p
, it consists of three
components (O
p
, F
p
, Goal
p
); O
p
and F
p
contain objects and facts were specificated on
knowledge K
p
.
Definition 4.3
: A model of Computational Object Knowledge Base with Sample Problems
(COKB-SP) consists of 7 components: (C, H, R, Ops, Funcs, Rules, Sample); in which, (C, H,
R, Ops, Funcs, Rules)
is knowledge domain which presented by COKB model, the Sample
component is a set of Sample Problems of this knowledge domain.
Algorithm 4.1
: To find a solution of problem P modelled by (O,F,Goal) on knowledge K of
the form COKB-SP.
