Optimizing Supply Chain Using ML Forecasting
BUSINESS CHALLENGES
The Retail giant has hundreds of stores across the country. To fulfil ‘the supply demand for those stores, they have several sorting facilities and distribution warehouses scattered across the county. The Client wanted to optimize their supply chain and optimize the use of labor, truck waiting time and optimize the truck load capacity, along the different routes connecting their distribution centers and retail stores.
SOLUTION
Our ML based Demand forecasting solution involved predicting the number of incoming and outgoing trucks in and out of the distribution warehouses and retail stores. This was done using ML-based demand forecasting techniques. At a high level, it consists of exploratory data analysis, cleansing, and data imputation, creation of multiple ML models including SARIMAX, FB Prophet, and XGBoost, training those models and producing the forecasts, and visualizing and presenting those forecasts to business users. Our solution has the following distinct features.
- Comprehensive EDA and use case context understanding.
- Pre-processing of the raw data to make the quality of data better.
- Experimentation with multiple models and utilize the best- performing ML Model.
- Prediction up to 90+%
- Predict from 4 weeks to 54 weeks
- Considering Multiple External factors, Seasonality and
Creation of visualization Creation of dashboards to present data to the stakeholders
BUSINESS IMPACT
1
The client was able to save cost on labor and optimize their utilization.
2
Optimize the space utilization in the Distribution warehouses.
3
Optimize the trucks’ movement between distribution warehouses and retail stores.
4
Better planning for truck and equipment future needs.
5
Better contractual agreement with third-party transportation vendors
