Ronald M. North

An Operational Approach to Multiple Objective Decision Making for Public Water Resources Projects Using Integer Goal Programming

Traditional decision-making techniques, such as linear programming, can only optimize one criterion. A variation of linear programming called integer goal programming is proposed as a quantitative approach to integrate multiple objectives in the selection and scheduling of federally funded water resources projects. The deviations from target levels of both economic and environmental objectives may be minimized by using integer goal programming in the evaluation of projects under the proposed multiple objectives. Integer goal programming has the added advantages of permitting decision makers to specify and to change the priorities given each goal and to select only complete rather than partial projects in multiple period funding of multiple objective projects.

The Measurement of Economic Uncertainty in Public Water Resource Development

The existing benefit-cost criteria for evaluating water resource projects are deterministic and therefore incomplete, since the uncertainty inherent in project outcomes is not considered. A Monte-Carlo simulation approach is used to generate a mean and standard deviation for the benefits, costs, benefit-cost ratio, and net present value for the controversial Spewrell Bluff Project. Subjective estimates defining probability distributions of project benefits and costs were obtained from the Corps of Engineers. A project selection process that includes probability considerations in the benefit-cost criteria is recommended and several approaches for including uncertainty as a variable are suggested.

Risk Analyses Applicable to Water Resources Program and Project Planning and Evaluation

Water resources project planning and evaluation involve engineering and economic types of analyses; and both are valid only to the extent that future events are correctly estimated. Each assumption, each conclusion is dependent on the analysts’ ability to estimate future events. Regardless of the analysts’ experience or the magnitude of the data base, all estimates of future events are inherently uncertain.