GROWTH SIMULATION OF ALFALFA CUTTINGS IN VITRO BY KALMAN FILTER NEURAL NETWORK

In order to investigate the combined effects of CO₂ and sucrose on the growth of alfalfa cuttings, the growth model using Kalman filter neural network was developed. Experiments were conducted on the growth of alfalfa cuttings under different conditions in combination of concentrations of CO₂ and sucrose. Growth data were collected 21 days after treatment. Then those data were used as training data for the neural training.

The simulated growth parameters such as dry weight, leaf number and root initiation of alfalfa cuttings calculated by the trained neural network showed good agreement with the experimental data. This result verified the satisfactory performance of the neural network model for simulating the behavior of the plantlet growth system.

The combined effects of CO₂ and sucrose on the growth of alfalfa cuttings were disclosed illustratively by this neural network model. The estimated optimum condition to increase the dry weight is about 2 000 ppm CO₂ with 0% sucrose.