Common Sense Reasoning in Diagnostic Systems - Efficient Decision Support Systems: Practice and Challenges from Current to Future - page 116

Let's estimate the obtained results.
Designate by maxd the maximal numerical value among versions (d) for all assessed
measurement points, and by CompCount a quantity of device components.
Accept in reviewing the following assessments:
1. max | J i leN d  | < CompCount. The quantity of components which are the

JEnvsx  ()

intersection result is always less than the quantity of whole device components;

2. min (| | | J  SingleNogood |) < CompCount. The quantity of components in the

JEnvsx  ()

prediction environment is always less than the quantity of the device components.
Taking into account these assessments, one can introduce a numerical assessment of the
obtained results:

0, if  JEn sxJ Sin eN d

SingleNogood
CompCount

( ) :  






ResultD(x),
maxD

if 

  min (| | | J  SingleNogood |),

JEnvsx  ()

QuanSNG x () 

if  J Envs x ( ) : SingleNogood  J

maxD  CompCount  max| J  SingleNogood |,

if  J  Envs x ():| J  SingleNogood ||  SingleNogood |



 maxD

 2* CompCount , if   J Envs x ( ): J  SingleNogood

Accordingly, for the example in the Fig. 9:
maxD = 0;
CompCount = 11;
QuanSNG(R3) = 0 + 11 - 2 = 9
QuanSNG(R1) = 0 + 11 - 2 = 9
QuanSNG(O6) = 0 + 11 - 1 = 10
QuanSNG(O5) = 0 + 2*11 = 22
QuanSNG(O4) = 0 + 11 - 1 = 10
QuanSNG(O3) = 0 + 2*11 = 22
QuanSNG(O2) = 0 + 11 - 1 = 10
QuanSNG(O1) = 0 + 2*11 = 22
The points with the greatest value of function QuanSNG ( x ) have the greatest priority of
choice. We will call the given method as SCAIEH (Supporting and Coinciding Assumptions
of Inconsistent Environment Heuristics).
One can see that the most preferable measurement points are О1, О3 and О5, one of
environment assumption of which coincides with SingleNogood . It differs from guidelines of
other choice methods, but at the same time for the given example the choice of any of these
points allows to test the most probable faulty component And2.
The developed methods of heuristic choice of the best current measurement point are
recommended to use for devices with a great quantity of components as quality of
guidelines directly depends on the quantitative difference of environments.

7.5 Practical results
Let's test the developed methods of the best measurement point choosing for the 9-bit parity
checker [Frohlich, 1998].