Inventory Planning - How to Optimize Inventory Policies

Balance inventory to hedge forecast risk and protect service levels while weighing cost tradeoffs from ordering larger batches, reduced transportation costs, and merchandising visibility across global and local warehouses.

Balance inventory levels by weighing holding costs and obsolescence risk against warehouse space and marketing effects. Weigh service levels against forecast error and bulk discounts.

Defines and contrasts cycle service level, item fill rate, on shelf availability, and order fill rate, showing how to measure them—often in Excel—and their impact on inventory policy.

Explore five cost buckets in supply chain planning—purchasing, transaction, holding, exploration, and shortages costs—and how variable, fixed, and change over time costs shape inventory policies.

The lecture analyzes holding and transaction costs in inventory management, covering warehouse finance, insurance, obsolescence, shelf life, shortages, and the roles of cycle and safety stock.

Define on-hand inventory as physically available units in your store, warehouse, or production facility, for inventory planning. Compute net inventory as on-hand plus in-transit minus backorders.

Explore how inventory policies split into static and dynamic approaches, and continuous versus periodic ordering, with trade-offs between forecast-driven decisions, fixed quantities, and exact quantities.

Learn periodic review and up-to-level inventory policies to decide when and how much to order, position stock downstream or upstream, and balance grouping with non-fixed quantities and slower reaction.

Examine periodic review and up-to-level policies with static and dynamic examples, stocking milk to three liters every Friday and using four-week forecasts to set net inventory to level s.

Learn about continuous review inventory policies with reorder point and fixed quantity orders; trigger points, dynamic reorder points tied to forecast, and the trade-offs of reactive ordering with suppliers.

Explore how combining fixed quantity with a periodic review and a reorder point enables single-supplier consolidation and logistics optimization, while noting the optimization challenges and potential shortages.

Define the risk horizon as the wait for an order, including lead time, and relate it to demand forecast coverage and the up-to level for periodic versus continuous review.

Compute forecast coverage by setting the order-up-to level to the sum of the next three months' forecasts, and compare with simple multipliers to show the impact of promotions and seasonality.

Compare three inventory policies via a data-driven cost simulation with real demand and forecast; the dynamic policy yields the lowest total cost, best fill rate, and lower shortage costs.

Simulate an inventory policy in Python using numpy arrays, managing historical and future demand, starting inventory, orders, and holding and shortage costs.

Refine the simulation by adjusting the code to future scenarios, assuming demand equals forecast, and compute sales as the minimum of forecast and starting inventory, highlighting potential shortages.

Simulate the hybrid policy by setting order up to level as the sum of the next three months' forecast plus a safety stock of 100 pieces, and generate a plot.

Simulate a dynamic inventory policy by computing the order up to level using a floor window and summing the forecast for the next three periods.

Explore why the safety stock formula relies on lead time, the review period, and forecast error, and how normality, independence, and constant over time variability limit its practicality.

Discover why safety stock uses forecast rmse, not demand standard deviation, since rmse compares a forecast created before demand to actual demand and captures trends and seasonality.

Evaluate forecast error with RMSE versus demand standard deviation to optimize safety stock, considering seasonality, promotions, and forecast-driven policies in inventory planning.

Evaluate independence versus three-month grouping in safety stock models by computing the RMSE over three consecutive months, revealing potential 1–2% inventory reductions with higher modeling complexity.