Hirdaramani

This project aims to enhance production planning in Sri Lanka’s garment industry by developing a hybrid machine learning model that accurately predicts daily production efficiency. The model integrates ensemble learning techniques (such as XGBoost and Random Forest) with time-series forecasting using LSTM networks to capture both operational and temporal patterns. Evaluated on real-world data from three factories, the model demonstrated superior performance in minimizing forecasting errors compared to traditional methods. The outcomes support data-driven decision-making in manufacturing, helping reduce overproduction, optimize resource allocation, and increase competitiveness in dynamic apparel markets.

Presented at:

SLAAI International Conference on Artificial Intelligence (SLAAI-ICAI 2024)

http://dx.doi.org/10.1109/SLAAI-ICAI63667.2024.10844970

International Conference on Information Technology Research (ICITR 2024)

http://dx.doi.org/10.1109/ICITR64794.2024.10857803