Preferential short-range reservoir control

Abstract

From the standpoint of real-time reservoir operation, the multipurpose control problem may be reduced to a dual purpose problem of (1) short-range control, which aims at reducing high flows and (2) longrange control, which aims at augmenting low flows and distributing stored water after the flood has receded. A decision framework for short-range reservoir control is formulated under three postulates: (1) The input to the control model is a stochastic forecast of the reservoir inflow process. (2) The control process is guided by a preference criterion which reflects the reservoir manager's value judgments concerning preferences over operating attributes, trade-offs between reservoir purposes, and attitude toward risk. (3) The long-range control is imbedded into the short-range control through the attribute space of the preference criterion, which allows for explicit consideration of the trade-offs between reservoir purposes, and through the state space and time domain of the control process, which allows for maintaining the continuity of the control. This investigation focuses on development of a preference criterion and on formulation of a control model. The preference crite- Huais developed within the framework of utility theory. The value judgments of the reservoir manager are quantified in terms of a two-attribute disutility function. It is argued that minimization of expected disutility is a plausible and well motivated criterion for reservoir control under uncertainty. A suitable disutility model is developed. The case of a group decision maker is analyzed in depth, and a methodology for obtaining a group disutility function is proposed. Some principles and techniques for assessing disutility functions are advocated; they are motivated by results of psychological research in human decision behavior, and are further supported by experimental evidence. Results of assessment of the reservoir control disutility function for several single and group decision makers are presented. The reservoir control process is conceptualized in the form of two sub-processes: (1) Forecast-Strategy Process, which is modeled as an open-loop feedback controller and (2) Control Process, which is modeled as a truncated Markovian adaptive controller. The optimal control strategy is selected on the basis of the preference criterion. A set of measures of effectiveness is proposed for evaluating the past performance of the controller. Computational aspects of the control model are analyzed. Certain monotonicity properties of the optimal control strategy are proven, and two suboptimal control strategies are derived: (1) partial open-loop strategy and (2) naive/partial open-loop strategy.