Information Technology Reference
In-Depth Information
which results in large memory requirement and long computing time. Thus, research-
ers have diverted their attention to meta-heuristic algorithms such as GA [31], SA
[32], and HS [33].
To test the algorithm performance, a four-dam system in Figure 5, first proposed
by Larson [34], is considered.
Fig. 5. Schematic of four-dam system
The objective of the multiple dam scheduling optimization is to maximize the
benefits from hydroelectric power generation and irrigation while satisfying all opera-
tional constraints such as range of water releases, range of dam storages, and mass
conservation between inflows and outflows. Boundary conditions such as initial stor-
ages and final storages must also be satisfied.
Although the total number of candidate schedules is 6.87 × 10 34 , HS found five dif-
ferent global optima with the identical maximal benefits after 35,000 evaluations, tak-
ing 46 seconds on a desktop with Intel Celeron 1.8GHz CPU [33]. However, GA
(with binary, gray, & real-value representations; tournament selection; one-point,
two-point, and uniform crossovers; and uniform and modified uniform mutations)
only reached a near-optimum after the same amount of evaluations [31].
4 Parameter Calibration of Flood Routing Model
A flood is a water overflow that submerges land [35]. Flooding results from the situa-
tion when the volume of a water body, such as a river or lake, exceeds the total capac-
ity of its bounds. Because the flood causes many problems, such as infrastructure
Search WWH ::




Custom Search