Speeding Up Many-Objective Optimization of Hydro-Generator Parameters

Multi-objective evolutionary algorithms (MOEAs) are broadly accepted methods in multi-criteria decision-making. Numerous advanced genetic optimization algorithms compute the complicated sets of Pareto-optimal solution. However, with hard-to-calculate objectives, the computation becomes intractable. This paper presents the approach to accelerate the solution of a multi-objective optimization problem, by the example of hydro-generator parameters. Using clustering and regression analysis allow us to reduce the sample size and optimization time, which is especially relevant for problems where each individual requires a long computation time. This paper will describe the underlying genetic algorithm, clustering analysis, regression analysis, and some initial results.