Sky-ZeroCarbon: A Sustainable Machine-Learning Framework for in-Flight Catering
International Symposium for Production Research, ISPR 2025, İstanbul, Türkiye, 9 - 11 Ekim 2025, ss.833-839, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1007/978-3-032-22784-3_71
- Basıldığı Şehir: İstanbul
- Basıldığı Ülke: Türkiye
- Sayfa Sayıları: ss.833-839
- Anahtar Kelimeler: Carbon Emissions, Food Waste, Machine Learning
- Maltepe Üniversitesi Adresli: Evet
Özet
Airline catering traditionally relies on large safety buffers of food to ensure that every passenger receives the meal of choice. While comforting for travellers, this practice imposes a hidden environmental cost: surplus meals increase aircraft weight, which raises fuel consumption and leads to unnecessary carbon emissions and food waste. Sky-ZeroCarbon is a data-driven framework that integrates machine-learning based demand forecasting with sustainability analysis to optimize in-flight catering. Using flight and passenger data for roughly 180 000 flights we train Random Forest and LightGBM models to predict meal demand with 92% accuracy, reducing food waste by 60%. The right-sizing of meals prevents around 120 kg of surplus food per flight and saves about 250 kg CO2, corresponding to annual savings of 20 000 t of carbon dioxide for a large carrier. We derive a simple relationship between excess meal weight and fuel consumption—each 100 kg of unnecessary meals requires 8–10 kg of extra fuel and emits 25–30 kg CO2—and convert meal reductions into greenhouse-gas benefits. This work demonstrates that integrating predictive modelling with sustainability metrics enables airlines to meet operational needs while advancing the United Nations Sustainable Development Goals on responsible consumption, climate action and healthy lives.