Smart Home AI Model (SHAM): Optimizing Switches and Sockets space in Residential Building during Planning/Construction through Multilayer Perceptron Estimation

Abstract

Abstract This study presents the development and evaluation of a Multilayer Perceptron (MLP) model for estimating the optimal number of switches and sockets space in houses not designed by professional engineers. Utilizing data from 145 houses across diverse cities, the study employs the Back Propagation Levenberg-Marquardt (BPLM) technique for MLP model estimation The SHAM is AI based multilayer perceptron model which recommends optimum number of switches and sockets space by considering factors like topography, type of property, size of property, type of occupancy. This model predicts the most suitable combination of switches and sockets space for a newly planned house. Utilizing data from 145 houses across diverse cities, the MLP architecture selection was based on performance metrics such as Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE) from testing and training data. The statistical analysis identified the 26-9-1 MLP architecture with a minimum MAPE value of 0.2572. Subsequently, this selected neural network model was applied to predict the optimal number of switches and sockets space. The study demonstrates the efficacy of the developed Smart Home AI Model (SHAM), proving valuable for both unprofessional and professional civil engineers during the planning stage. The implementation of the AI-based SHAM model is discussed, highlighting its adaptability to get optimum number of switches and sockets space without a detailed designing of electrical circuits. Results from real-world simulations demonstrate the effectiveness of the model in achieving energy savings, efficient construction and user satisfaction compared to traditional, non-optimized setups.