Calculation of multi-objective optimal tradeoffs between environmental flows and hydropower generation

Reservoir operation commonly involves tradeoffs between the objective of meeting downstream environmental flows and hydropower generation that may call for high reservoir level to generate hydropower more efficiently. Fixed length gene genetic programming (FLGGP) is an optimization tool that searches the decision space of an optimization problem employing a variety of functions, mathematical operators, and variables. This paper develops the FLGGP as a multi-objective optimization method and compares its performance with that of the nondominated sorting genetic algorithm (NSGA)-II in calculating multi-objective reservoir operational rules. Results show the multi-objective FLGGP decreased (improved) 5.64 and 39.5% the environmental flow and hydropower generation objectives, respectively, relative to the NSGA-II's values.