Common Sense Reasoning in Diagnostic Systems - Efficient Decision Support Systems: Practice and Challenges from Current to Future - page 105

Having obtained analogies, the DMP may then make the final choice of the best ones. On the
basis of these facts, the facts (properties) that hold for the source O are transferred to the
receiver R.
Let us consider the steps of the functioning of the algorithm using the example from power
engineering - operation control of the nuclear power unit (see Fig. 3,4).
As a receiver R for the analogy, we take Situation 4 and as the property p , we take Close
TH11S24.
In the first step, VC =, VO =, VA = and Vp' = {Situation 1, Situation 2, Situation 3}.
Since Vp' ≠ , we go to the next step.
Determine intersections of C with R taking into account p. Add in VC only Situation 1,
because the number of common properties nRC = nR for Situation 2 and Situation 3. Since
VC ≠ , we go to the step 3.
Determine all possible sources O and go to step 4. In this case VO = {Situation 2, Situation
3}, because the Situation 1 is unique intersection for analogy.
In the fourth step, we construct only two analogies for R - Situation 4:
A1 = <Situation 2, Situation 1, Situation 4, Close TH11S24 >;
A2 = <Situation 3, Situation 1, Situation 4, Close TH11S24 >.
Add new analogies to VA and go to step 5.
The analogies obtained in step 4 ( A1, A2 ) are given to the DMP.
As a result we obtain two analogies. Choosing one of them, the DMP can transfer facts that
hold for the source of the analogy to its receiver. In this example, a new fact about the
recommendation “Supply the pump TH11D01 with boric concentrate 40g/kg caused by switching
off ACS 1 due to closing the gates TH11S24 and TH11S25” arises for Situation 4.
The methods of reasoning by analogy is more general than on the bases of cases. Analogies
are used when it is impossible to find a suitable case in a case library. The reasoning by
analogy method can be used independently from a case- based reasoning method as well as
for correction (adaptation) of the nearest to a problem situation case to form a new case for
completing a case library. Further we shall consider the case-based reasoning method and
its application.

3. Case-based reasoning
Case-based reasoning (CBR) , like analogous reasoning, is based on analogy; however, there
are certain differences in their implementation [Aamodt, 1994; Eremeev et al., 2006b].
As the practice shows, when a new problem situation arises, it is reasonable to use the
method of case-based reasoning without drawing an analogy. This is caused by the fact that
humans operate with these reasoning schemes at the first stages, when they encounter a
new unknown problem.
Case-based reasoning is an approach that allows one to solve a new problem using or
adapting a solution of a similar well-known problem. As a rule, case-based reasoning
methods include four main stages that form a CBR-cycle, the structure of which is
represented in Fig. 5.
The main stages of CBR-cycle as follows:


Retrieving the closest (most similar) case (or cases) for the situation from the case
library;

 Using the retrieved case (precedent) for solving the current problem;

 If necessary, reconsidering and adaptation of the obtained result in accordance with the

current problem;