Environmental Engineering Reference
In-Depth Information
DP1
DP = Decision pipelines (time varying portfolios)
Decision/action point
Defence
raising &
Barrier work
Defence
raising &
Barrier work
Defence
raising &
Barrier work
Defence
raising &
Barrier work
Defence
raising &
Barrier work
Flood storage,
retreat
Flood storage,
retreat
Flood storage,
retreat
Flood sto
rage,
retreat
Flood storage,
retreat
DP2
Flood storage,
retreat
Flood storage,
retreat
Flood storage,
retreat
DP3
Southend
barrier
Southend
barrier
Southend
barrier
Southend
barrier
Southend
barrier
Existing
Existing
Existing
system
Existing
system
Existing
system
Existing
Existing
Existing
system
DP4
Defence
raising &
Barrier work
DP5
Defence
raising
FRM system state at specific time
- Only key new features listed
DP6
2008
2050
2100
Time
Fig. 15.11
The performance of different strategic alternatives (represented by unique routes through the future
decisions) enable adaptive strategies to be developed that reflect future uncertainty. FRM, flood risk management.
found to create solutions that are even more
desirable. In this way its search of the solution
space is not as regimented as hill-climbing tech-
niques (allowing the search to be targeted in areas
thought to be most favourable, whilst not restrict-
ing the search based on this bias) nor as random as
Monte Carlo sampling. By combining the char-
acteristics of two different 'parent' solutions a
new solution is proposed. The best-performing
offspring are selected and recombined to develop
(hopefully) ever fitter solutions. This process is
repeated over several 'generations' (iterations)
until the maximum utility (across multiple cri-
teria) of the solution is reached.
Tools to optimize basic interventions (crest
level raising, condition grade improvements) have
now been trialled on simple flood risk manage-
ment studies (Philips. 2006; Woodward
et al. 2010). The basic building blocks of these
tools are shown in Figure 15.12 and include:
.
A description of autonomous future changes:
The future is of course unknown and there are
many uncertain influences outside of the asset
manager's control. For example (i) climate change
(UK Climate Programme 2009, UKCP09) provides
a probabilistic description of potential future cli-
mates than can be readily utilized within the
optimization process); (ii) asset deterioration [in
the absence of management or reduced manage-
ment. Expert-based deterioration curves are typ-
ically used to describe deterioration from one
condition grade to the next (Simm et al. 2008).
Process-based statistical models are starting to
emerge with the ability to model asset time-de-
pendent processes using Markov processes, such
as the Poisson or the gamma process (Buijs
et al. 2005)]; (iii) central budgetary change; and
(iv)
socioeconomic
change
and
floodplain
development.
.
An ability to incorporate multiple (competing)
objectives: Flood risk management takes place in
a world of many competing demands. Optimiza-
tion allows these to be explicitly described as ob-
jective functions, for example (i) maximise
economic benefits; (ii) minimize whole-life costs;
(iii) minimize loss of life; (iv) maximize environ-
mental enhancements, etc. whilst, for example,
reflecting budgetary constraints.
