Some capacity expansion problems encountered in water resources development are studied. Models yielding least-cost shedules (i.e. optimal selection and sequencing of potential projects) in order to meet a stipulated increase in water demand are first discussed. However, recognizing uncertainties in the future demand, these ought to be taken into account. In order to do this, a new general dynamic programming algorithm has been developed. Depending on how detailed we are able to state the preferences against shortage, chance-constraints or penalty functions are included in the optimization. The use of the algorithm, as an aid in the current planning, is illustrated by solving an example under deterministic as well as stochastic assumptions about the future demand.
Research Article|June 01 1977