Bio-inspired cyber-physical systems for adaptive building performance: a simulation study integrating AI-driven control and digital twin concepts

Abstract

Purpose This study develops a simulation-based cyber-physical framework that integrates bio-inspired adaptation and artificial intelligence (AI) within a digital twin environment to improve building energy efficiency, indoor environmental performance, and adaptability. Design/methodology/approach A scientometric and narrative review was conducted to examine trends in intelligent building control, followed by the development of a MATLAB/Simulink simulation. The framework combines bio-inspired thermoregulation logic, AI-based predictive control, and a simulation-based feedback loop for adaptive building performance assessment. It was evaluated across temperate, tropical, and arid climates using representative building configurations under these climate conditions. Findings The AI-augmented system achieved the best performance, reducing HVAC energy use by up to 27% in tropical climates (from 15.2 to 11.1 kWh/day) and improving the temperature-based comfort proxy by reducing deviation from the setpoint from 2.9°C to 1.3°C. It also showed faster thermal recovery (25–30 min) and stable learning convergence, with mean squared error decreasing from 0.09 to 0.005 over 50 epochs. Scientometric analysis indicates rapid growth in this field between 2016 and 2024, with AI, sustainability and IoT as dominant themes. Research limitations/implications Results are based on simulation using standardised parameters. Real-world validation is required. Practical implications The framework supports adaptive, energy-efficient building control using biologically inspired and data-driven strategies. Originality/value The study presents an integrated simulation framework combining bio-inspired control, AI, and digital twin modelling for adaptive building performance.