Measurement and modelling of the water requirement of some greenhouse crops with artificial neural networks and genetic algorithm

Crop evapotranspiration is the most important parameter for management of irrigation systems in greenhouses. This study was conducted to determine the evapotranspiration of cucumber, tomato and peppers, using micro–lysimeter during seven months in a greenhouse located in central region of Iran. Reference evapotranspiration estimated using drainage lysimeters and the water balance of soil micro–lysimeters was determined using the gravimetric method. To find the relationship between meteorological data and crops height with crops evapotranspiration, artificial neural networks (ANNs) and genetic algorithms–ANNs (GA–ANNs) were used. The results indicated that both models had a quite good agreement with the actual evapotranspiration of crops, but the GA–ANNs model will respond better than the ANNs model.

Keywords: greenhouses, crop evapotranspiration, cucumbers, tomatos, peppers, artificial neural networks, ANNs, genetic algorithms, GAs, micro–lysimeters, irrigation management, Iran, drainage lysimeters, water balance, greenhouse crops