PROJC01092

• Contract/PO #: C153625
• Contractor Name: Dames and Moore
• Water District: 0
• Bill Number: XB 99-999
• Loan Projects - Doc Type: Contract Documents

.,

()

Simulation models are also known as descriptive models. The user instructs the model
in how to proceed, a situation which is ide:1lly suited to playing .what if' games. Simulation
. models can be further classified into the following subcategories:

· Physically based models. These models actually represent the physical processes.
An example of this type of model is flow routing by directly solving the equations
of motion.

· Conceptually based models. These types of models, oiten referred to as empirical
models, approximate the actual physical process by simple empirical mathematical
relationships. The parameters of these models can usually be related to the
physical characteristics of the system.

· System theoretic models. System theoretic models are also known as .black box.
or input/output models. Examples of a model of this type would be a unit
hydro graph routing, or an ARIMA model.

· Other models of processes or regulations, such as water law, are not easily
expressed in a mathematical fonn.

Optimization models are also known as prescriptive models, since they present the
user with an action plan that will maximize or minimize some single or multiple objectives.
Three broad subclasses of optimization models include:

.

· Linear programming models. In this type of model constraints and objective
functions are linear. They are commonly solved by a simplex algorithm or one
of its derivatives.

· Non-linear programming models. Here, constraints and/or the objective function
are non-linear. One type of non-linear programming problem is the maximization
of hydropower generation.

· Dynamic programming models. In this type of model have an Optlmlzation
problem is solved by sequentially decomposing a problem into a series of
dependent simpler problems. Dynamic programming and its derivatives are
powerful tools in solving non-linear optimization problems although the size of the
problem may be limited.

Analysis models are used for evaluation of either historical data, or simulation of
optimization model results. There are at least three subclasses of analysis models:

· Statistical analysis. Given the current state of science, it is impossible to forecast
with accuracy most water processes. Random processes are analyzed by
probabilistic and statistical techniques. Therefore, statistical analysis tools must
be included in the CRDSS.

.

..1_