Estimating the in-plane lateral resistance of reinforced log wall employing soft modelling techniques

Abstract

The popularity of log houses has been on the rise in numerous regions worldwide. In the context of log construction, the stability of log walls is notably influenced by the friction existing between the layers of logs and the openings designated for windows and doors. This study endeavors to comprehensively evaluate the lateral resistance of log walls through an extensive parametric analysis utilizing finite element (FE) methods. To construct a robust dataset, a total of 71 distinct samples were generated employing FE analysis, where the shuffled frog-leaping algorithm (SFLA) was incorporated in conjunction with a feed-forward (FF) neural network. Within this framework, the accuracy of the SFLA-based informational model was juxtaposed against that of an artificial neural network (ANN) coupled with particle swarm optimization (PSO), genetic algorithm (GA), and statistical models including multiple linear regression (MLR).