Prediction of Dissolved Oxygen in Shrimp Pond using Dolphin Glow Worm Optimization Based Radial Basis Function Neural Network

The invent of culture fisheries in India has brought a significant increase in the export revenue for India in recent years. Shrimp farming in culture fisheries faces several challenges due to the improper water quality management in the shrimp pond. As a result, the productivity gets a significant hit and the operational costs increases. Dissolved Oxygen (DO) is the vital water quality parameter which influences the growth rate and also plays a pivotal role in the mortality of the shrimps when not maintained at optimal levels in the water. The DO is tested from the water samples in the pond in laboratories far from the actual shrimp pond which is time consuming.This research study proposes an Internet of Things (IoT) based real time monitoring of the water quality parameters using a solar powered floating buoy which monitors the critical water quality parameters and displays them in an interactive graphical user interface to the farmers which provides insights to farmers for taking necessary corrective actions.The water quality data collected from the pond will serve as an input to the Dolphin Glow Worm Optimized Radial Basis Function Neural Networks (DGO-RBFNN) for prediction of DO in the shrimp pond. The DGO-RBFNN overcomes the problem of selection of hidden neurons in the network faced by the conventional machine learning algorithms and reduces the error rate of the classifier. The proposed DGO-RBFNN outperforms the conventional Multilayer Perceptron Neural Networks (MLPNN), Radial Basis Function Neural network(RBFNN) and Fuzzy Based RBFNN in terms of Mean absolute error, Root mean square error and Willmott index of agreement.