Chapter
5
In the previous chapters the emphasis has been placed on finding an optimal
plan for water resources development. However, other aspects of the optimal plan besides the values of the objective function and the variables
must be taken into account by planners. For example, all plans must be
executed in a dynamic environment in which changes in inputs, costs, and
objectives occur because of related technological, political, and social
changes. If a planner is to select an appropriate series of decisions in time,
he must be aware of the consequences of such changes on the optimal
plan(s). In fact, it often is more important for the planner to know how the
physical system will perform under different operating conditions than it is
for him to devise an optimal plan for a single postulated set of prices,
inputs, and demands.
Methods of incorporating the appropriate features of variability and
fluctuations into the planning include deterministic techniques, such as
"worst case" methods and sensitivity analysis, as well as stochastic approaches. To accommodate the prospects of uncertainty, in this chapter
we first examine the sensitivity of planning decisions by application of the
algorithm previously used in Chapter 4 to solve the deterministic river
basin management problem. We then make a few remarks concerning the
stochastic approach to water resources planning, even though this topic is
essentially outside the scope of this book.
5.1. Sensitivity Analysis
A sensitivity analysis examines the effect of small perturbations in model
inputs and model parameters on the values of the model outputs and the
130
5
In the previous chapters the emphasis has been placed on finding an optimal
plan for water resources development. However, other aspects of the optimal plan besides the values of the objective function and the variables
must be taken into account by planners. For example, all plans must be
executed in a dynamic environment in which changes in inputs, costs, and
objectives occur because of related technological, political, and social
changes. If a planner is to select an appropriate series of decisions in time,
he must be aware of the consequences of such changes on the optimal
plan(s). In fact, it often is more important for the planner to know how the
physical system will perform under different operating conditions than it is
for him to devise an optimal plan for a single postulated set of prices,
inputs, and demands.
Methods of incorporating the appropriate features of variability and
fluctuations into the planning include deterministic techniques, such as
"worst case" methods and sensitivity analysis, as well as stochastic approaches. To accommodate the prospects of uncertainty, in this chapter
we first examine the sensitivity of planning decisions by application of the
algorithm previously used in Chapter 4 to solve the deterministic river
basin management problem. We then make a few remarks concerning the
stochastic approach to water resources planning, even though this topic is
essentially outside the scope of this book.
5.1. Sensitivity Analysis
A sensitivity analysis examines the effect of small perturbations in model
inputs and model parameters on the values of the model outputs and the
130