Compressive strength prediction of structural lightweight concrete by applying the recurrent neural network‐fractional order bat optimization algorithm

Abstract

AbstractNeural networks methods produce an analysis in various fields. The Neural Network is frequently employed for developing statistical methods for nonlinear systems since, the Neural Networks use the simulation benefits of complicated behavior in several problems. In this study, the Recurrent Neural Network and Recurrent Neural Network based on Fractional Order Bat Optimization Algorithm is utilized for predicting the compressive strength of Lightweight Concrete mixes in 3, 7, 14, and 28 days curing. Recurrent Neural Network has been used for having a comparison. The inputs to the network for the compressive strength are made of eight variables: sand, lightweight fine aggregate, water to cement ratio, silica fume utilized in solution, Lightweight course aggregate, silica fume utilized supplementary to cement, curing period, and superplasticizer. According to the results, the Recurrent Neural Network based on Fractional Order Bat Optimization Algorithm has predicted precise outcomes and trained so fast in comparison with RNN method.by considering the outcomes, the Neural Networks methods are satisfactory mechanisms to evaluate the LWC compressive strength. This can decrease the price and reduce the time of processes in these problems.