Estimating Heating Load in Residential Buildings Using Multi-Verse Optimizer, Self-Organizing Self-Adaptive, and Vortex Search Neural-Evolutionary Techniques

Abstract

Using ANN algorithms to address optimization problems has substantially benefited recent research. This study assessed the heating load (HL) of residential buildings’ heating, ventilating, and air conditioning (HVAC) systems. Multi-layer perceptron (MLP) neural network is utilized in association with the MVO (multi-verse optimizer), VSA (vortex search algorithm), and SOSA (self-organizing self-adaptive) algorithms to solve the computational challenges compounded by the model’s complexity. In a dataset that includes independent factors like overall height and glazing area, orientation, wall area, compactness, and the distribution of glazing area, HL is a goal factor. It was revealed that metaheuristic ensembles based on the MVOMLP and VSAMLP metaheuristics had a solid ability to recognize non-linear relationships between these variables. In terms of performance, the MVO-MLP model was considered superior to the VSA-MLP and SOSA-MLP models.