Optimizing
system's performance
Because parameters
are not known a priori but rather have to be optimized, simulation is
driven by an external optimization procedure, where the objective
is the maximisation of system's performance.
A
performance index of the system is introduced as a measure of the
efficiency of a particular policy management, which is totally specified
through the reservoir operation rules. Therefore, the
control variables of the optimization model are just the two parameters
of each reservoir.
The performance index
is expressed
in terms of reliability, yield or economy. The
objective function can have one of the following formulations:
 |
minimization
of the mean failure probability of the system, for a given set of
operational targets |
|
maximization
of the total annual withdrawal, for a given reliability level |
|
minimization
of the mean water transportation cost, for a given set of operations
targets and a given reliability level |
The objective function
is evaluated via the simulation process. Given that the optimization problem
is strongly nonlinear, advanced techniques are used, particularly the
multi-start downhill simplex method (Nelder and Mead, 1965) and
the shuffled complex evolution method (Duan et al., 1992).
|
Home
page 
greek pages
