New optimization methods for designing rain stations network using new neural network, election, and whale optimization algorithms by combining the Kriging method.

Abstract

In many studies of water and hydrological sources, estimation of the spatial distribution of precipitation based on point data recorded in rain gauge stations is of particular importance. The purpose of this paper is to optimize the network of rain gauge stations in the Sistan and Baluchestan Province with respect to the variance of Kriging and topography estimation in the region and to maintain or reduce the number of stations in the region (without incurring additional costs). A new neural network algorithm has been presented in the present study to determine the optimum rain gauge stations. In this study, a new method of meta-heuristic optimization algorithm based on biological neural systems and artificial neural networks (ANNs) has been proposed. The proposed method is called a neural network algorithm (NNA) and has been developed based on unique structure (ANNs). In order to evaluate the proposed method, the election and whale algorithms have been used. The election algorithm is a repetitive algorithm that works with a set of known solutions as a population, and the whale optimization algorithm is derived from the new nature based on the special bubble hunting strategy used by the vultures. The results showed that 22 stations of the existing network had no significant effect on rainfall estimation in the province and their removal to the optimal network is suggested. Therefore, the remaining 27 stations can be effective in optimizing the rain gauge network. The results of comparing the abovementioned algorithms showed that the neural network algorithm with a mean error of 0.06 mm has a higher ability to optimize the rain gauges than blue whale and election algorithms.