How to Calculate Reorder Point & Safety Stock: QR Inventory Model

Given stockout costs of 150 per unit and carrying costs of 50 per unit, determine optimal safety stock level by comparing total costs at different safety stock levels: (1) For 100 units: Carrying cost = 100 × 50 = 5,000, Stockout cost = 0, Total = 5,000, (2) For 80 units: Carrying cost = 80 × 50 = 4,000, Stockout cost = 20 × 150 × 0.02 = 60, Total = 4,060, (3) For 50 units: Carrying cost = 50 × 50 = 2,500, Stockout cost = 50 × 150 × 0.05 = 375, Total = 2,875, (4) For 20 units: Carrying cost = 20 × 50 = 1,000, Stockout cost = 80 × 150 × 0.10 = 1,200, Total = 2,200, (5) For 0 units: Carrying cost = 0, Stockout cost = 100 × 150 × 0.33 = 4,950, Total = 4,950. The optimal safety stock is 20 units with minimum total cost of 2,200.

Stockout occurs when demand cannot be fulfilled due to insufficient inventory. Safety stock is the level of stock maintained in excess of expected consumption to protect against unexpected demand or supply delays. The optimal safety stock level is determined by balancing carrying cost (cost of holding inventory) against stockout cost (cost of lost sales or customer dissatisfaction). The goal is to minimize total cost (carrying cost + stockout cost). Given: Maximum stockout = 100 units, Probability = 2%, Stockout cost = 150 per unit. Carrying cost = 50 per unit. Safety stock options: 0, 20, 50, 80, 100 units. Total cost = Carrying Cost + Expected Stockout Cost. The optimal safety stock is determined by finding the level that minimizes total cost.

Stock out occurs when inventory is depleted, causing inability to fulfill orders. This results in financial losses (lost contribution margin) and non-financial losses (customer dissatisfaction). Safety stock provides buffer against demand or supply uncertainties. The trade-off is: higher safety stock reduces stockout risk but increases carrying costs. Expected stockout cost is calculated by multiplying stockout cost at each level by its probability and summing results. Stockout cost = Stockout Quantity × Contribution Margin per Unit. The optimal safety stock level is found where total costs (carrying cost + expected stockout cost) are minimized. In the example, 20 units of safety stock yielded minimum total cost of ₹2,140, demonstrating that the optimal level is not necessarily the highest level.

This section covers the EOQ model when stockouts are permitted. The total cost includes ordering cost, holding cost, and stockout cost. The stockout cost includes lost sales, penalty payments, customer dissatisfaction, and production disruption. The optimal order quantity with stockouts is: Q* = √(2 × D × CO × (CS + CP) / (CS × CP)), where CP is the stockout cost per unit. The instructor emphasizes that allowing stockouts increases the optimal order quantity. A critical requirement is that all costs must be expressed in the same time units before applying the formula. This extension shows how the basic model can be adapted for different business scenarios.

This segment covers EOQ calculation using formula √(2 × Annual Demand × Ordering Cost / Carrying Cost). With annual demand of 54,000 batteries, ordering cost of 9,000, and carrying cost of 37.5% of purchase price (300 per unit), EOQ is 800 units. Safety stock and reorder point calculation with risk consideration involves: (1) Determining desired service level (e.g., 85% for 15% risk), (2) Using delivery time probability distribution to find minimum lead time for desired service level, (3) Calculating reorder point as daily usage × minimum lead time for desired service level, (4) Calculating safety stock as buffer beyond average lead time usage. For 15% risk of stockout with 70% chance of 6-day delivery and 15% chance of 7-day delivery, orders should be placed 7 days before stockout. With daily usage of 150 units, reorder point is 1,050 units and safety stock is 150 units.

Probabilistic inventory models manage uncertainty in demand and lead time using probabilities and statistics. The continuous review policy monitors inventory constantly and triggers orders at the reorder point. Three scenarios exist: (1) variable demand with constant lead time, (2) constant demand with variable lead time, and (3) both variable. In the variable demand scenario (200 units/day average, 80 units standard deviation, 18-day lead time), optimal order quantity is 6,000 units, reorder point is 4,035 units, and safety stock is 435 units. In the variable lead time scenario (8-day standard deviation), safety stock increases to 20,500 units, reorder point to 5,650 units, and maximum inventory to 80,500 units. Annual costs rise from 1,630,000 to 1,681,000 soles. Higher variability requires larger safety buffers to maintain service levels.

Three specialized models handle situations with variability: (1) When only demand varies: ROP = D̄×LT + Z×σD×√LT; (2) When only lead time varies: ROP = D×L̄T + Z×D×σLT; (3) When both vary: ROP = D̄×L̄T + Z×√(L̄T×σD² + D̄²×σLT²). Each model uses the normal distribution assumption and requires finding the appropriate Z-score from service level tables. The choice of model depends on which parameters (demand, lead time, or both) exhibit variability in the specific inventory situation.

This comprehensive segment develops advanced stochastic inventory frameworks for managing random lead times and uncertain demand. The (R, Y) policy triggers fixed quantity Y orders when inventory reaches reorder point R, accommodating both random lead times and variable demand. Three core assumptions govern this model: unfilled demand is backlogged, no more than one outstanding order exists, and demand distributions remain stationary. The total cost function integrates setup costs (KD/Y), holding costs (H × I), and shortage costs (PD × S), where I = Y/2 + E[max(R - X, 0)] represents average inventory. Expected shortage cost per cycle equals ∫(x-R)f(x)dx from R to infinity. Setting partial derivatives to zero yields two interdependent equations that resist analytical solution, requiring iterative numerical methods. The algorithm initializes with Y₁ = √(2KD/H) and R₀ = 0, alternating computations until convergence. This progression from deterministic foundations to sophisticated stochastic optimization illustrates the evolution of inventory management theory toward practical decision-making tools for uncertain supply chain environments.

Inventory simulation models systems with random lead times and demand. Parameters include: Reorder point (4 units), Lead time distribution (1 day: 50%, 2 days: 30%, 3 days: 20%), Daily demand distribution (1 unit: 10%, 2 units: 30%, 3 units: 40%, 4 units: 20%). Random number ranges map to these probabilities: Lead time ranges are 0-49 (1 day), 50-79 (2 days), 80-99 (3 days). Demand ranges are 0-9 (1 unit), 10-39 (2 units), 40-79 (3 units), 80-99 (4 units). Simulation tracks: Opening Inventory, Simulated Demand, Closing Inventory, Order Placement, Lead Time, Replenishment Received. Order placement rule: If Closing Inventory < Reorder Point, place order. Lost sales occur when demand exceeds available inventory.

In inventory management with random demand and lead times, uncertainty is measured by combining demand variability (standard deviation σ_D) and lead time variability (standard deviation σ_L) through the standard deviation of lead time demand (σ_DL), calculated using the formula that accounts for both the average demand rate and lead time duration, providing a unified measure of risk that enables better inventory decision-making under uncertainty.

This section covers safety stock and reorder point calculations. Safety stock is calculated by multiplying daily consumption by the safety period: Safety Stock = Daily Consumption × Safety Period. Maximum stock is calculated by adding safety stock to minimum stock: Maximum Stock = Minimum Stock + Safety Stock. Reorder point is calculated by multiplying daily consumption by the lead time: Reorder Point = Daily Consumption × Lead Time. When inventory levels reach the reorder point, a new order should be placed. These calculations ensure continuous operations while maintaining optimal inventory levels.

This segment demonstrates the complete calculation of Safety Stock and Reorder Point. The Safety Stock formula is: Z × √[(Lead Time × Demand Variance) + (Demand² × Lead Time Variance)]. Using Z=1.04, lead time=15 days, demand=18 units/day, demand variance=0.15, and lead time variance=3, the calculation yields approximately 57 units. The Reorder Point formula is: (Average Daily Demand × Lead Time) + Safety Stock = (18 × 15) + 57 = 327 units. These calculations ensure the company maintains adequate inventory to meet demand during supplier lead times.

Safety stock is calculated using the formula: Safety Stock = NORMINV(service level, average demand, standard deviation) × √(lead time) + (average demand × lead time) - average demand. This formula accounts for demand variability and lead time uncertainty. The NORMINV function returns the z-score corresponding to the desired service level. Reorder point is calculated as: Reorder Point = (Average Demand × Lead Time) + Safety Stock. This ensures that when stock reaches this level, enough inventory will be available to cover demand during the lead time plus the safety buffer. The reorder point is also the minimum stock level that triggers ordering.

Safety stock is the extra inventory held to protect against uncertainties in demand or supply. The formula involves calculating the difference between maximum and minimum consumption rates, then multiplying by the minimum lead time. The reorder point determines when to place new orders, calculated as normal consumption multiplied by the maximum reorder period. For emergency purchases, calculations consider the gap between normal and maximum consumption rates. The stock level calculation determines current inventory position using: Stock Level = Reorder Quantity + Reorder Level - Minimum Consumption. These formulas help businesses maintain optimal inventory levels and prevent stockouts by accounting for variability in demand and supply lead times.

Safety Stock is a buffer inventory level maintained to protect against supplier delays or demand uncertainties. Companies never want inventory to fall below this minimum level. The Reorder Point is the inventory level at which a new order must be placed, calculated by multiplying lead time by daily demand. For example, with a 5-day lead time and 25 kg daily demand, the reorder point is 125 kg. When inventory reaches this level, the company must order to avoid stockouts. These concepts ensure continuous operations despite supply chain uncertainties.

Multi-echelon inventory optimization leverages the risk pooling principle to reduce total system safety stock by strategically relocating safety stock from upstream locations to customer-facing echelons; simulation studies demonstrate that pushing all safety stock to regional distribution centers can achieve the same or higher customer service levels (99%) while reducing total safety stock by approximately 30%, though this counterintuitive approach faces resistance from managers who perceive it as reducing their internal service levels despite the overall inventory reduction benefits.

Multi-Echelon Inventory Optimisation (MEIO) is an advanced supply chain technique that optimizes safety stock placement across multiple echelons (suppliers, manufacturers, distribution centers, and customers) simultaneously, rather than treating each node independently as in single-echelon approaches. The core principle is that safety stock coverage at each location depends on the service time commitments made by upstream and downstream nodes, with the total coverage across all stocking points equaling the maximum lead time at the customer-facing node. MEIO leverages the risk-pooling effect, where consolidating safety stock downstream reduces total inventory requirements compared to decentralized holding. The optimization process uses five key demand statistics (mean, standard deviation, non-zero mean, standard deviation excluding zeros, and inter-demand interval mean) to classify demand patterns (smooth, erratic, or lumpy) and select appropriate probability distributions, ultimately minimizing total holding costs while maintaining target service levels.

Multi-echelon inventory optimization addresses coordination challenges across supply chain stages. The Stochastic Service Model (Clark-Scarf) treats each node as a news vendor but struggles with complex networks. The Guaranteed Service Model (Graves-Williams) uses dynamic programming for tree-structured networks by bounding demand above (typically 3 standard deviations). For unsolved problem classes, Stockpile implements brute force enumeration (testing all base stock level combinations) and coordinate descent (optimizing one variable at a time), though these are computationally intensive. The package supports serial, assembly, and general tree networks, with approximately 70 built-in textbook instances. Simulation validates analytical solutions by comparing optimized costs against simulated outcomes.

Service time defines the fulfillment time when one stage requests another stage's product. Unlike lead time, service time is strategically determined and need not correlate with a stage's own processing or review time. A stage might quote zero service time (immediate availability) while still holding safety stock. The optimization objective minimizes total expected safety stock cost: Σ(α × C_i × Z × √(τ_i)), where τ_i = S_i^out + T_i + R_i - S_i^in. This framework enables coordinated inventory management across the entire supply chain network rather than treating stages independently.

Multi-echelon optimization recommends inventory targets holistically across the entire network, recognizing that inventory pooling reduces total requirements. When recommending safety stock at one location, the algorithm simultaneously considers all other locations, understanding that upstream positioning enables risk pooling to serve multiple downstream points. Mathematical foundations include WMAPE for forecast error quantification and EOQ balancing ordering against holding costs. Larger order quantities can reduce safety stock requirements by providing extended coverage. Service level measurement distinguishes probability of no stockout (cycle service level) from fill rate (proportion of demand satisfied immediately). Strategic positioning favors upstream locations for greater flexibility responding to regional demand fluctuations while minimizing total capital investment.