Automatic production planning using hybrid semi-supervised learning and optimization algorithms

The Industrial Hub of Manaus faces a challenging reality, marked by recurring logistical issues and supply shortages especially during the dry season. In this context, ensuring the continuity of production operations requires more than just planning; it demands intelligence and adaptability. This article presents the development of a recommendation system that leverages artificial intelligence and machine learning to swiftly suggest alternative production models when materials are unavailable. The solution combines various techniques, including similarity-based clustering (k-means and PCA), collaborative filtering, and linear programming. It was tested using real data from a local manufacturing plant in a simulated environment. The results were highly positive: there was a reduction in material shortages, an increase in continuous production time, and improved adherence to production plans. More than just numbers, this study represents a concrete advancement in dealing with uncertainties and harnessing technology for faster, safer, and well-informed decision-making. The proposed solution not only addresses current challenges but also points toward a more flexible, intelligent, and responsive approach to modern industrial planning.