Abstract
This paper presents the results of a study of a strategic management for estimation of the water demand in a paddy field throughout the operation period of the water management forecasting system developed by the authors using fuzzy inference methodology.
On water management practice in a paddy field, the administrator uses a certain strategy to either increase or decrease the amount of water demand, taking into account not only previous day's water demand and environmental conditions, such as the growing stage of the rice, daily average temperature and weather, but also the change in the farm management practices, requirements of the farmers and so on.This process is mostly influenced by intuitional and creative judgement of the administrator.In order to grasp the structure of strategy, index W which can measure grade of the strategy in quantity was newly introduced into the forecasting system.This was achieved by applying the idea of Truth Value of Fuzzy Set.Then the property of W value was searched using the results obtained from the system operation in the field. Moreover, the mechanism ofinferring W value was studied using the Neuro Network, and it was confirmed that this model is useful for estimating the most suitable management strategy.
Based on this fundamental study, the authors developed a new type of system with a function inferring suitable management strategy using the Neoru Network as a sub-system within the water management forecasting system.As a result of operation in the paddy field, it was observed that the new type of system can estimate water demand of paddy fields with a better correlation to the actual water demand, in comparsion with the old forecasting system.
