Inventory management is critical in aviation to ensure that essential supplies, including catering, passenger comfort items, and emergency equipment are available for smooth operations. Mismanagement can lead to delays, increased costs, and operational inefficiencies. Machine learning (ML) offers the potential to analyze historical data, predict inventory needs, and streamline restocking processes, ultimately improving turnaround times and reducing waste.
Objectives:
- Develop a ML-based framework for forecasting inventory needs, optimizing restocking schedules & minimizing delays in aviation operations.
- Identify key factors influencing inventory usage patterns.
- Test and validate predictive models using real or synthetic aviation inventory datasets.
- Provide actionable insights to reduce costs and ensure availability of critical items.
Research Questions:
- How can ML models predict inventory needs for catering, comfort, and emergency supplies based on flight and passenger data?
- What features (e.g., passenger count, flight duration) significantly impact inventory usage?
Tasks:
- Data Collection: Acquire historical inventory usage data, flight schedules, and passenger details. Use open datasets or create synthetic data reflecting inventory trends in aviation.
- Data Preprocessing: Clean and preprocess data (handling missing values, normalizing features). Perform exploratory data analysis (EDA) to identify trends and correlations.
- Feature Engineering: Identify influential features such as flight duration, seasonality, passenger count, and regional preferences. Develop derived features like average consumption per passenger.
- Model Development: Apply supervised ML models like: Regression models (for quantity forecasting) & Classification models (to flag critical inventory shortages). Compare with unsupervised learning for anomaly detection.
- Testing: Use historical data for model training and recent data for validation.
- Evaluation Metrics: Accuracy and precision of predictions. Reduction in overstock/understock incidents. Time savings in restocking processes.
Time duration: 6 Months
Contact:
Dr. Thomas Feuerle, t.feuerle@tu-braunschweig.de
Parth Purohit, parth-yogeshbhai.purohit@tu-braunschweig.de