Application of the modified fibonacci search-based search beam method to derive a nondominated release policy for a multi-purpose reservoir

Abstract

One of the most serious problems of water resources development planning is the lack of acceptable procedures to deal with conflicts among different purposes at the early conceptual planning stage of any water resources system. As far as the consideration of formulating planning strategies is concerned, a new interactive multi-objective decision making (MODM) is approach, the search beam method (SBM) has been conceived. This model is expected to provide a reliable systematic procedure to search for the nondominated solutions along prespecified lines representing hypothetical search beams (SB) in the objective space. The proposed SBM consists of two major parts, an optimization phase and a subsequent evaluation phase. Incremental dynamic programming (IDP), one of the best known approximative techniques of dynamic programming (DP), is applied in the optimization phase. The ‘modified Fibonacci search’ is heuristically used to seek, in a successive search for the intersection with the Pareto optimal surface along each SB. The viability of the SBM is demonstrated by applying it to an existing single unit reservoir system. This reservoir is assumed to accommodate two conflicting objectives: maximizing hydropower generation and maximizing water surface elevation to sustain accessibility of the navigation structures and to enable low head pumping of irrigation water.