Effective functional dossier categorisation is important for maximising system adeptness and averting some sins in the Industrial Internet of Things (IIoT). The feature of preparation subsets is a common restraint on usual ensemble methods like bagging. To overcome this, we support a mixture strategy that connects Genetic Algorithms (GA) and Teaching Learning-Based Optimisation (TLBO), exceptionally built to correct the preparation subset option in IIoT requests. This plan helps the ensemble’s adaptability and elasticity in a variety of industrialised frameworks by dynamically regulating bag sizes and utilising an appropriateness function that carefully balances veracity and variety. Explainable AI (XAI) approaches are contained in the ensemble design to support obvious and understandable visions into the accountable process, therefore promoting transparency and confidence in robotic IIoT schemes. Experiments on original-planet IIoT datasets show that our approach outperforms normal ensemble methods in terms of veracity and explainable indicators, which are essential for making educated decisions in mechanical backgrounds.