Implementasi Metode Dual Response Dengan Jaringan Saraf Tiruan Untuk Memprediksi Overcapacity Tempat Pembuangan Akhir Di Kota Manado

Abstract

The dual response method is a method used to get maximum prediction results on artificial neural networks to be able to predict the next period. The purpose of this study is to predict the mass amount of waste that goes to the landfill for the next one year and the time of excess capacity at the Sumompo landfill based on historical data on the mass of waste that goes to the landfill and the area of ​​the Sumompo landfill.                                           

   The initial stage carried out in this study was to take historical data on the mass of waste that entered the Sumompo landfill at the Manado City Environmental Service then from the data searched for the average value and standard deviation of the historical data obtained, then the data was normalized to binary sigmoid form. with a range of 0.1 to 0.9. After that, the data is trained to recognize patterns through the training and testing process, then proceed to the simulation process to get predictive results.

            Prediction results from research using the "trainlm" training function with network architecture 12-8-3-2 obtained the mass of waste that enters the final disposal site from September 2021 to August 2022 are: 1796172 tons, 1703016 tons, 1553125 tons, 1627118 tons , 1627118 tons, 1552924 tons, 1703036 tons, 1553150 tons, 1705416 tons, 1855288 tons, 1695870 tons, 1776243 tons. The prediction results show that the highest waste mass will occur in June 2022 with a pile height of 17 meters. Overcapacity of the final disposal site is predicted to occur in March 2022 where the landfill has exceeded the limit for the remaining volume of the landfill, which is 30,000,000 cubic meters of waste.

.

Keywords : Prediction, dual response, neural network