MULTIOBJECTIVE DYNAMIC-PROGRAMMING - CLASSIC PROBLEM REDRESSED

The problem of noncommensurable multiple‐objective functions in water resources and potential methods for solving such multiobjective problems are discussed. A new method, namely, multiobjective dynamic programing (MODP), that borrows from conventional dynamic programing is developed. It generates the entire noninferior solution set of the multiobjective problem, and it can be used to generate the trade‐off ratios between objectives. The trade‐off ratios are actually a by‐product of the MODP method and are available explicitly between all objectives. The noninferior solution set is obtained solely in terms of the objectives rather than in terms of the individual decisions that influence each objective, thus facilitating the decision maker's task in choosing the preferred solution. The Reid‐Vemuri multi‐objective problem in water resources was chosen as an example and was successfully solved using MODP. Copyright 1979 by the American Geophysical Union.