Multiple reservoirs management using genetic algorithm

摘要

An advanced methodolgy is proposed that can provide an optimal reservoir management strategy for effective or beneficial use of water stored in an interconnected multiple reservoirs system. Each individual reservoir is regarded as a lumped impoundment, thus its time-varying storage being represented by the water balance equation in the form of an ordinary differential equation. The reservoirs are also assumed to be operated in compliance with a combination of both the individual prior determined rules (i.e., rule curves) and the interrelated rule that allows regulatory release or transfer of redundancies to the downstream reservoir. A simulation/optimization model is thus developed that optimizes the regulated flows (i.e.,the proportionals) and computes the resulting individual storages. In search of an optimal storage regulation for a given reservoirs system,the genetic algorithm is employed to find the optimal proportionals in the context of minimizing an appropriately defined fitness function. Some demonstrative operations of the model for a fictitious system suggest that with the successful convergence of the genetic algorithm the proposed mothodology enables complete determination of the regulated flows which can make maximum use of the impounded water resources.