Abstract Two sorts of optimal control techniques, the modern control theory based on optimal regulator and the intelligent control based on neural network and genetic algorithm, were applied to the control of two physiological processes of plant, the water status to light intensity and the net photosynthetic rate to the drainage and supply of the nutrient solution in hydroponics, respectively. In the modern control, the ARMA model of the water status to light intensity was identified by using least squares method. The Kalman filter was used to estimate state variables. The covariance of the Kalman filter was decided by simulation. In the experiment, the ratio of the weight matrices were varied between 0.1 to 10.0. The optimal control performances showed good agreement with the simulation, and it showed good results for control of water status of plants. In the intelligent control, on the other hand, the neural network was used for the identification of the net photosynthetic rate to the drainage and supply of the solution, and the genetic algorithm was used for the search of optimal value through the model simulation. The three-layer neural network made it possible to identify the nonlinear behavior of the net photosynthetic rate to the drainage and supply. The use of genetic algorithm allowed to quickly search the optimal value of the drainage and supply times which maximizes the net photosynthetic rate through the model-simulation. The optimal value could be obtained within 5-generation when both rates of crossover and mutation take high values as 0.8 and 0.8, respectively. It was found that the optimal regulator is useful for the control of the plant water status to light intensity, and the combination of neural network and genetic algorithm is effective for the control of such complex process as net photosynthetic rate to the drainage of the solution
[1]
Tetsuo Morimoto,et al.
Identification of Hydroponics in an Advanced Control System of the Greenhouse
,
1992
.
[2]
John J. Grefenstette,et al.
Optimization of Control Parameters for Genetic Algorithms
,
1986,
IEEE Transactions on Systems, Man, and Cybernetics.
[3]
Osamu Kitani,et al.
Measurement and Control of Water Status of a Plant (Part 2)
,
1993
.
[4]
Stephen A. Billings,et al.
Non-linear system identification using neural networks
,
1990
.
[5]
Yasushi Hashimoto,et al.
Identification and Control of Hydroponic System in Greenhouses
,
1985
.
[6]
E. J. van Henten,et al.
Optimal control of greenhouse climate.
,
1991
.
[7]
H.-J. Tantau,et al.
Optimal control for plant production in greenhouses
,
1991
.
[8]
Geoffrey E. Hinton,et al.
Learning representations by back-propagating errors
,
1986,
Nature.
[9]
D. E. Goldberg,et al.
Genetic Algorithms in Search
,
1989
.
[10]
Yasushi Hashimoto.
Computer integrated plant growth factory for agriculture and horticulture
,
1991
.