Design of Hybrid Soft Computing Techniques for Estimation of Suspended Sediment Yield in Krishna River, India

This paper explores the soft-computing approaches for estimating suspended sediment yield (SSY). In the management of water resources, the SSY estimation is an important issue. The SSY estimation is essential for getting the information about mass balancing between the land and ocean.  The traditional methods for measuring the SSY require larger magnitudes of time and significant financial investments. Also, the SSY depends on numerous variables and their internal relationship which are extremely nonlinear and complex in nature. Thus, traditional methods are not capable to handle the complex nonlinear sedimentation behaviors and unable to accurate estimation of SSY. The multilayer perceptron (MLP) artificial neural network (ANN)-based genetic algorithm (GA) model is used for SSY prediction which resolves complex sedimentations problems. In this proposed model, the GA optimized all ANN’s model parameters simultaneously. The input functional parameters that impact the SSY in the Krishna River are water discharge and water level. This paper presents artificial intelligence-based sediment yield estimation algorithms at Waddepally gauge station in Krishna River, India. The GA is used for the optimization of the performance of ANN in accurately estimating the SSY. The hybrid GA-based ANN (GA-ANN) has produced most accurate and efficient results for estimation of SSY in Krishna River.