1686. How many is too many toasties? Optimising fresh food loading for inflight retail
Invited abstract in session MA-42: Sustainable food supply chains, stream Circular & Sustainable Supply Chains.
Monday, 8:30-10:00Room: Newlyn GR.02
Authors (first author is the speaker)
| 1. | Craig Poku
|
| Datasparq | |
| 2. | Paul Sharkey
|
| Datasparq | |
| 3. | Matthew Sawkins
|
Abstract
From production to consumption, food is one of the world’s biggest waste sources. Staggeringly, the UN estimates that one-fifth of food produced is thrown away. For airlines, fresh food sales are a crucial revenue stream for inflight retail, making it essential to strike the right balance between maximising profit, minimising food waste, and avoiding stockouts.
This talk presents the solution developed for a low-cost airline to optimise fresh food loading predictions. Previously, their business practices led to extreme overloading or underloading, resulting in missed profit opportunities and waste levels as high as 42%. Our solution consisted of two key elements: a forecaster that uses features such as past sales and flight characteristics to predict demand of a given food item; and a stock optimiser to find the optimal number of loadings that will maximise profit while constraining for both food waste and stockouts.
During offline evaluation, we observed a 5% profit uplift across the airline’s network while decreasing food waste by 19% and stockouts by 9%. This and further live tests demonstrated that our model is competitive in terms of profit, while enhancing sustainable practices. Although our client initially told us to focus on just profit uplift, we showed that beyond financial gains, using demand forecasting improves sustainability by reducing food waste and improving supply chain efficiency.
Keywords
- Supply Chain Management
- Sustainable Development
- Airline Applications
Status: accepted
Back to the list of papers