Using Al and simulation tools in supply chain

The success of the initial daily forecasting activities at Coolies regional outlets (refer to Activity 2 HERE) has helped them coordinate agreements with suppliers that significantly minimized the risk of overstocking/understocking faced by these stores. Now, Coolies is considering a reconfiguration of its supply chain network and wishes to implement new agreements accordingly with its suppliers for high demand non-perishable items (i.e., rice and pasta). It must first understand and forecast the total demand for these products in order to enter the right contract with suppliers, and therefore must use aggregated demand for its future forecasts. While there has been success with the daily forecasting approach, this method is more suitable to implement in day to day operations. Focusing on network reconfiguration and contract design, Coolies has decided to employ aggregated Weekly forecasting approach rather than daily to save costs and improve coordination. Before they proceed with aggregated weekly forecasts, however, they must establish the error in the forecasting method to determine whether the switch in forecasting approach for both SKUs is justified. Using the historical demand data for rice and pasta provided HERE, run the Forecasting Al Plus module and submit your answer to the following questions:
• Q1: Considering the historical demand data from August 1, 2019 to December 31, 2020 and using the forecasting Al Plus module of the Log-Hub add-in, forecast the aggregated weekly demand for January 1st, 2021 to March 31st, 2021 for each product. Provide the visualization for each forecast and briefly explain your observations.
• Q2: For each product, calculate the error in the weekly forecast (i.e., the weekly forecasted value between January 1st, 2021 and March 31st, 2021 minus actual total demand value for each week in the same period). Calculate the weekly MAPE of the forecasts for each product (refer to Measures of Forecast Error concept). Briefly explain your observations.
• Q3: If the Weekly MAPE value obtained in the previous step is less than 50%, use weekly aggregation for that SKU to forecast the months of April to June 2021, based on actual demand data from August 1, 2019 to March 31st, 2021. Otherwise, change the aggregation basis in the Forecasting Al module to Day and use daily aggregation if MAPE for an SKU is 50% or greater. Briefly explain your observations.