📦 Warehouse Inventory Lab: EOQ, Safety Stock and Reorder Point

The science

Continuous-style (s, Q) policy: when the inventory position (on hand + on order - backlog) drops to the reorder point s, order a fixed batch Q. Daily demand is Normal(mu, sigma); unmet demand is backordered. The order arrives after the lead time L.

  • Economic order quantity: Q* = sqrt(2 K D / h), minimising K D/Q + h Q/2 (D = 365 mu per year).
  • Lead-time demand has mean mu L and std. dev. sigma_L = sigma sqrt(L); reorder point s = mu L + z sigma_L + u, safety stock z sigma_L. Because the shelf is checked once a day, the position overshoots the trigger by about u = (mu^2 + sigma^2)/(2 mu) on average, so that allowance u is added to s.
  • Probability of no stock-out in a cycle: Phi(z). Fill rate: 1 - sigma_L G(z) / Q with the normal loss function G(z) = phi(z) - z (1 - Phi(z)).
  • Annual cost: K D/Q + h (Q/2 + z sigma_L) + pi (D/Q) sigma_L G(z).

Honest simplifications: demand is integer-rounded normal and independent day to day; the stock is reviewed once a day (handled by the allowance u, itself an approximation); one supplier with a fixed lead time; the average-inventory term Q/2 + safety stock ignores the backorder correction; costs are illustrative. The sweep button simulates 10 years per order size and the dots should hug the closed-form curve.