Knowledge Representation and Validation in a Decision Support System: Introducing a Variability Modelling Technique - Efficient Decision Support Systems: Practice and Challenges from Current to Future - page 39

4.1.3 Explanation and corrective recommendation
This operation is defined (in this paper) for highlighting the sources of errors within
decision- making process. The errors happened due to dissatisfaction of constraint
dependency rules.
The general pattern that represents failure due to dissatisfaction of constraint dependency
rules is:
Decision A excludes Decision B and Decision A is selected then Decision B fails to select.
In the proposed method, there are two possibilities for the decision: decision point or choice.
Three possibilities for the exclusion constraint: choice excludes choice, choice excludes
decision point, or decision point excludes decision point. We assign the predicate notselect to
the excluded choice for preventing future select.
The following definition describes these possibilities in the form of rules:
Selection of choice n , select (n) , fails due to selection of choice x , select(x) , and assign by
notselec t predicate in three cases:
i. ∀ x,y,n:type(x,choice) ∧ select(x) ∧ type(y,decisionpoint) ∧ choiceof(y,x) ∧ type(n,choice) ∧ excludes_c_
dp(n,y) ⟹ notselect(n).
If the choice x is selected, and it belongs to the decision point y , this means y is selected
(Rule 7), and the choice n excludes the decision point y , this means n is assigned by notselect
predicate.
ii. ∀ x,y,z,n:type(x,choice) ∧ select(x) ∧ type(y,decisionpoint) ∧ type(z,decisionpoint) ∧ type(n,choice)
∧ choiceof(y,x) ∧ choiceof(z,n) excludes_dp_dp(y,z) ⟹ notselect(n).
If the choice x is selected and x belongs to the decision point y , that means y is selected (Rule
7), and if the decision point y excludes the decision point z , this means z is assigned by
notselect predicate (rule 6), and the choice n belongs to decision point z , this means n is
assigned by notselect predicate (rule 9).
iii. ∀ x,n: type(x,choice) ∧ select(x) ∧ type(n,choice) ∧ excludes_c_c(x,n) ⟹ notselect(n).
If the choice x is selected, and x excludes the choice n , which means n is assigned by
notselect predicate (rule 2).
Two examples are presented to express how the proposed method could be used for
guiding the decision maker by explanation and corrective recommendation. Example 1
shows an interactive corrective recommendation mechanism. Example 2 shows how the
proposed method validates decision maker in future based on his current selections.
Example 1
Suppose the decision maker selected decrease level as a punishment for one employee. After
that, the decision maker selects high level as training for the same employee; the system
rejects the choice and directs the decision maker to deselect decrease level first. Table 4
describes Example 1. This example represents rule (iii). The example illustrates how the
proposed method guides decision makers to solve the rejection reason.
Example 2
The decision maker asks to select the choice non promotion decision , which is excludes positive
performance decision point. The system accepts the choice and assigns the decision point
positive performance by notselect predicate to validate future selections. Table 5 describes
Example 2. The predicate notselect (positive performance) prevents the selection of its choices,
Rule 9.