Civil Engineering Reference
In-Depth Information
for it and decide on adjustments or on what needs to be done to improve
the chances of total success.
Until recently, the fl ags were used entirely qualitatively as described in
Blockley and Godfrey (2000). A new algorithm developed for the purpose
of propagating the fl ags mathematically has been described recently (Block-
ley, 2008; Marashi, 2006). However certain constraints and assumptions
need to be appreciated before the method is used. The new algorithm is
implemented in two 'sister' pieces of software called PERIMETA (Perfor-
mance through Intelligent Management) and JUSTISE (Joined Up Systems
Thinking for Integrating Synergy in Engineering). An Italian fl ag is defi ned
mathematically as an interval probability - so the probability of an event
or proposition
E
has a lower bound
El
and an upper bound
Eu
. Therefore
p(E)
is an interval number as follows:
()
=
[
]
=
[
(
)
]
[9.2]
p E
El Eu
,
g
, 1
−
r
Here
g
represents the green part of the fl ag colouring an interval on the
scale [0, 1] from 0 to
g
and is a measure of the probability of the set
E
.
Likewise
r
is the red part of the fl ag from (1
−
r
) to
1
and is a measure of
the probability of the set
−
E
. The white part of the fl ag is therefore
w
=
1
−
r
and is a measure of the probability of the unknown events in the set
(E or
g
−
E)
. The white interval enables us to deal with incomplete and pos-
sibly inconsistent information which we can interpret as 'don't know'.
−
9.9
A structural example: procuring a new building
Assume we are concerned with risks to the procurement of a new building.
Since this new building is to be in a seismically active zone, we also want
to consider possible future retrofi ts. Following the process methodology
described by Blockley and Godfrey (2000), the fi rst task is to identify a
process model as shown in Fig. 9.4. In doing so, one has to draw a
system
boundary
- so we must include not just direct hard systems technical issues
but also soft systems social and political issues that may affect our building.
We use the BCIOD
+
R (Business, Customer, Integrating, Operations, Deliv-
ery
Regulatory) framework, as set out by Blockley and Godfrey (2000).
Note that, due to space restrictions, the model shown in the fi gure is incom-
plete - the intention is that the hierarchy is developed suffi ciently to illus-
trate the methodology.
Some Italian fl ags of evidence are also shown in the fi gure - remember
that they are dynamic as they change during the life of the project. The
input fl ags at the bottom of the tree can either be input directly by the
relevant process owner or derived from a performance indicator using a
value function as described by Hall
et al.
(2004). The fl ags for the higher
process are calculated using the pair-wise algorithm described earlier. This
+
Search WWH ::
Custom Search