An implicit stochastic process is compared to possible alternative analytical processes. An implicit stochastic process refers to a sequence of steps consisting of streamflow synthesis, deterministic optimization, and multivariate analysis of deterministic optimization results. The comparison is incomplete in that the performance of the multivariate analysis of results is not within the scope of this paper. Some formulations based on the method of Dantzig‐Wolfe decomposition are shown to be reasonably accurate representations of a nonlinear multireservoir deterministic optimization problem. One of these methods is shown to consume only moderate amounts of computer time for a problem of distinctly nontrivial size. The research issues yet to be resolved in achieving the overall objective (i.e., a multireservoir operation rule for flow regulation and energy production) are identified.

Linear programing formulations, dynamic programing formulations, and policy iteration models have all been used to solve for optimal operating rules for a single stochastic reservoir. In this paper, quantitative guidelines which discriminate between dynamic programing models and policy iteration models are developed. A judgment that stochastic dynamic programing models should be the preferred algorithm is reached.

The basic structure of the Techcom planning methodology is reviewed. In the context of this system a weighting methodology is developed to measure preferences regarding the attainment of postulated societal goals. The selected methodology, the Metfessel general allocation test, is applied to the hierarchies of three Techcom goals relating to aesthetics, econ6mics, and recreation. The weighting methodology is tested with answers to a mail questionnaire randomly distributed throughout Arizona and New Mexico. Examples of the resulting preference weights are presented, and the use of such measures by decision makers discussed.

For a given production rate from a proposed well field, well construction and replacement costs decrease with fewer and larger wells spaced farther apart, while pumping lift costs increase. A method for minimizing the sum of these two costs uses a two‐dimensional uniform grid of wells to select a combination of well spacing and well capacity that minimizes costs for a given demand. Uncertainty in field permeability at future well sites is treated by averaging the possible outcomes over all wells in accord with a probability density function for permeability values. The method assumes an annual cycle whereby the water table declines owing to pumping in a dry season and is recharged to its initial state during a wet season.

The problem considered in this paper i s the detemnination of the operating p o l i c y over t h e of a network of water reservoirs or dams arranged according t o an arbitrary topology. If streamflow was a known quantity, the problem could be formlated as a k g e nonlinear pro constraints as a m i n i m ) and so ved by the two methods proposed i n this paper. Since spwmflow is