VARFIS: A Hybrid Neuro-Fuzzy Model for Intelligent Microclimate Control in Black Soldier Fly Farming Systems

Maintaining optimal microclimate conditions is essential for Black Soldier Fly (BSF) cultivation, yet traditional systems often struggle with dynamic environmental changes. This study proposes the Vector Autoregressive-Fuzzy Inference System (VARFIS), a hybrid model combining Vector Autoregression (VAR) and Adaptive Neuro-Fuzzy Inference System (ANFIS), to enhance temperature and humidity control in BSF insectariums. VARFIS adapts to uncertainty using probabilistic learning, achieving a 48% reduction in prediction error (MAPE = 1.36%) and high accuracy (R² = 0.9695), outperforming standalone VAR and ANFIS models. The model effectively captures daily climate fluctuations, improving larval growth efficiency and waste conversion. However, it remains limited in handling extreme events such as sudden heatwaves or humidity spikes, indicating the need for enhancements like adaptive fuzzy rule tuning and integration of physical constraints. VARFIS presents a scalable solution for intelligent microclimate management, supporting sustainable insect farming and circular economy goals. This work contributes to precision agriculture by offering data-driven tools for resilient environmental control.