In this paper, we propose a novel technique to implement nutrient solution control using genetic algorithm based on mamdani fuzzy inference system (FIS) that grades the nutrient solution control. A novel FIS grading system has been developed for this purpose based on expert guidance from agricultural scientists of Murugappa Chettiar Research Centre (MCRC), Chennai, India. The designed FIS is used as fitness function to execute genetic algorithm which optimizes the control system parameters of nutrient preparing unit periodically and thus maintaining the quality of solution. The proposed technique has been implemented using Matlab and the results of algorithm are validated by simulating a virtual nutrient control unit using Labview. The performance of the system is compared with classical nutrient control unit designed using genetic algorithm using absolute error as fitness function.
The paper is structured as follows: Section II gives a literature survey of nutrient solution control techniques. Section III gives a brief description of a typical nutrient control unit and the experimental setup used in this paper. Section IV gives description of the algorithm and detailed analysis of FIS rule base developed using expert opinion. Section V discusses the results and a performance comparison of proposed technique against classical absolute error based technique is presented. Section VI concludes the paper with future scope of this research work.