DEMAND FORECASTING & ASSET REPOSITIONING
The industrial tank container division of a large shipping company wanted to reduce idle tank time, increase demand fulfillment to better serve customers.
Challenge
Struggling with missed sales opportunities in certain parts of the world and idle/excess capacity in others, the client wanted to:
- Forecasts Demand at Depot Level & Increase Forecast Accuracy
- Provide guidance(partially automate) asset repositioning planning
Approach
To enable depot-level forecasting – gaps in the system were filled using predictive ML techniques. This input was then used to create demand forecasts for weekly shipments per different tank types for each of the 1800 depots. The output of the demand forecasting models is being used to create the tank repositioning planning, which helps ensure the physical availability of tanks at the right place at the right time. During COVID-19, event modeling was applied on-demand models to simulate the impact of changes in demand and create alternative business planning scenarios.
Methodologies
- MRD Orchestrator with ML for Ops Probs
Results
- Increase Forecast Accuracy from 30% to 60%