Pemodelan Jaringan Saraf Tiruan Pertumbuhan Sawi Pagoda pada Sistem Hidroponik

Abstract

Pemodelan Jaringan Saraf Tiruan Pertumbuhan Sawi Pagoda pada Sistem Hidroponik. Dibimbing oleh HERRY SUHARDIYANTO. Sawi pagoda merupakan sayuran yang kaya akan nutrisi. Sawi pagoda umumnya dibudidayakan secara konvensional. Namun berkurangnya lahan pertanian menjadi sebuah kendala. Budidaya di dalam greenhouse dengan sistem hidroponik dan root zone cooling (RZC) membantu mengoptimalkan produktivitas tanaman. Oleh karena itu, dilakukan penelitian ini untuk mempelajari hubungan antara pertumbuhan sawi pagoda terhadap faktor-faktor lingkungan. Selama percobaan, suhu harian rata-rata di dalam greenhouse berkisar antara 33.1C hingga 39.1C. RZC telah dapat menjaga suhu perakaran dalam rentang 24.4C hingga 26.7C. Parameter lingkungan yang diamati meliputi suhu udara di dalam greenhouse, radiasi matahari, kelembapan relatif udara di dalam greenhouse, dan suhu perakaran. Parameter tanaman meliputi bobot segar, bobot kering, jumlah daun, dan luas permukaan daun. Hubungan parameter-parameter tersebut telah dapat diterangkan sesuai model ANN yang telah dikembangkan yaitu menggunakan algoritma backpropagation. Model ANN terbaik memiliki momentum 0.9, learning rate 0.1, dan arsitektur 11-2-2. Evaluasi model ANN menghasilkan nilai R2 sebesar 0.98800 dan nilai RMSE bobot segar dan bobot kering secara berturut-turut sebesar 3.54870 dan 0.24627. Nilai-nilai tersebut menunjukkan bahwa model ANN dapat memprediksi bobot segar H+2 dan bobot kering H+2 dengan sangat baik.

Artificial Neural Network Modelling of Tatsoi Growth on Hydroponic System. Supervised by HERRY SUHARDIYANTO. Tatsoi is a vegetable that is rich in nutrition. Tatsoi is usually cultivated conventionally. However, the decreasing agricultural land has become a problem. Cultivation in a greenhouse with a hydroponic and root zone cooling (RZC) system helps optimize the plant’s productivity. Therefore, this research was performed to study the correlation between tatsoi’s growth and environmental factors. During the experiment, the daily average air temperatures inside the greenhouse were between 33.1C and 39.1C. RZC kept the root zone temperature between 24.4C and 26.7C. The observed environmental factors were outside air temperature, solar radiation, relative humidity of inside air, and root zone temperature. The plant’s growth parameters were fresh weight, dry weight, number of leaves, and leaf area. The correlation between these parameters has been explained through the ANN model developed using the backpropagation algorithm. The best ANN model figures were 0.9 momentum, 0,1 learning rate, and 11-2-2 as the architecture. The model evaluation produced an R2 value of 0.98800 and fresh weight and dry weight RMSEs of 3.54870 and 0.24627, respectively. These values show that the model could predict fresh weight D+2 and dry weight D+2 very well.