Dynamic Optimization of Watering for Qualitative Improvement of Satsuma Mandarin Using Intelligent Control Techniques

To improve the quality of Satsuma mandarin during the developing and maturing stages, an optimal watering operation was investigated through identification and optimization using intelligent control techniques. Dynamic changes in the sugar content and the citric acid of Satsuma mandarin, as affected by rainfall and sunshine duration, was first identified using neural networks, and then an optimal operation for watering (rainfall management) that maximizes the sugar content and that minimizes the citric acid of Satsuma mandarin was sought through simulation of the identified neural-network model using genetic algorithms. An optimal solution obtained here was an operation that increases the amount of watering during the first stage and then markedly reduces it during the latter half stage of the maturing stage. From simulation, the maximization of the amount of watering during the first stage resulted in a significant decrease in the citric acid, and the minimization during the latter half stage led to a marked increase in the sugar content. A mulching technique with plastic film and a sprinkling operation allow controlling the amount of rainfall. This operation suggests a better way for supplying the water to Satsuma mandarin and improving the quality. Copyright@2005 IFAC