©ALGORITMA GENETIK UNTUK MENINGKATKAN KINERJA MODEL TANGKI STANDAR PADA ANALISA TRANSFORMASI DATA HUJAN MENJADI DATA ALIRAN SUNGAI

Abstract

Fundamental weakness of the tank model application is so much value parameters must firstbe defined simultaneously before the model was applied. This condition causes tank models areconsidered not efficient to solve practical problems. This research is an attempt to improve theperformance of Standard Tank Model that can be applied more effectively, especially for thetransformation of climate data into the stream data. The discussion focused on efforts to completethe system of equations in standard tank model using genetic algorithms for optimization parameters,so that the resulting equation system can determine the appropriate model parameters automaticallyat a watershed in the study. Standard tank model is a system composed tank 4 series and has 17parameters. Results of research on the Konto Watershed and the Lekso Watershed show thatStandard Tank Model-based Genetic Algorithm can present relationships very well climate data andstreams data. At the maximum generation value of 500 obtained root mean square error (RMSE) of0.241 m3/sec for the Konto Watershed and the Lekso Watershed of 0.30 m3/sec.Keywords: genetic algorithm, a standard tank model, optimization, parameters