LPWM Expert: An expert system for water management during land preparation in a paddy estate in Malaysia

Abstract

Seberang Perak paddy estate, Malaysia, which practices intensive mechanized farming still uses the traditional approach in decision-making. Land preparation water management, the critical process to be completed within scheduled duration, needs better and quick management decisions for many alternative scenarios. A method proposed to encapsulate specific knowledge available with domain experts and generated through modeling to an expert system (LPWM Expert) is outlined. The LPWM expert consists a database, a modal base, a knowledge base and a user interface. Knowledge was gathered through discussions and interviews with domain experts. Collected quantitative data were used in modeling canal flows and water balance to extract knowledge for different possible scenarios. Knowledge base represent extracted knowledge as rules. All the rules in IF-THEN structure and syntax are verified with the help of wxCLIPS debugging capability and results generated by the LPWM expert are validated with the domain experts. The expert system proposes decisions for many combinations of scenarios considering all the possible variations in rain, irrigation water supply, secondary blocks, sub-estates, cropping seasons and cropping intensity. The LPWM expert is user friendly and efficient where the outputs are supported with graphics.