Order-Up-To Level

The inventory position is the key metric monitored, not just on-hand stock. It is calculated as:

• On-hand inventory: Physical units in the warehouse
• + On-order units: Quantities already requested from suppliers but not yet received
• - Backorders: Customer demand that is committed but unfilled

The order quantity is simply S minus current inventory position, ensuring the position returns to the target level regardless of demand spikes.

Unlike a reorder point system that only covers lead time demand, the Order-Up-To Level must protect against variability over the protection interval:

• Review Period (R): The fixed time between inventory checks
• Lead Time (L): The supplier's delivery time after order placement
• Total Exposure: R + L

This means the system must absorb demand uncertainty for the entire review cycle plus replenishment lag, requiring a higher safety stock than continuous review systems for the same service level.

The optimal Order-Up-To Level is derived from the probability distribution of demand over the protection interval:

• Mean demand: Expected consumption during R + L
• Standard deviation: Variability of demand during R + L
• Target service level: Desired probability of no stockout (e.g., 95% cycle service level)

The formula S = μ + z × σ uses the z-score from the standard normal distribution corresponding to the target service level, ensuring the safety stock component is statistically calibrated.

A defining characteristic of the Order-Up-To policy is that order sizes are inherently variable:

• In high-demand periods, large orders are placed to restore the position to S
• In low-demand periods, small orders or even no orders are generated
• This variability can create the bullwhip effect upstream if suppliers are not prepared

This contrasts with fixed-order-quantity policies like EOQ, where batch sizes remain constant but the timing between orders fluctuates.

In a multi-echelon inventory optimization (MEIO) context, the Order-Up-To Level at each node is not calculated in isolation:

• A retail store's S level directly impacts the demand pattern seen by the regional distribution center
• The distribution center's S level must account for the aggregated variability of all downstream locations
• Guaranteed Service Models (GSM) and Stochastic Service Models (SSM) use different assumptions about how upstream stockouts propagate to determine optimal S levels across the network

Real-world deployment requires defining the review rhythm and exception handling:

• Review frequency: Daily, weekly, or monthly cycles aligned with supplier order windows
• Min-max constraints: Business rules that override the calculated S level for shelf-life limits or warehouse capacity
• Demand sensing integration: Short-term POS data can dynamically adjust S levels between formal review cycles
• Promotional overrides: Planned marketing events require temporary S level increases to avoid stockouts during demand surges

The foundational inventory control framework where stock status is checked at fixed, regular intervals rather than continuously. The Order-Up-To Level is the target used in this system.

• Review Period (R): The fixed time between inventory checks.
• Mechanism: At each review, an order is placed to raise the inventory position to the Order-Up-To Level.
• Contrast: Unlike a continuous review system that triggers an order when stock hits a reorder point, this policy only acts at scheduled times.

The real-time metric that the Order-Up-To Level is compared against to calculate the replenishment quantity. It is not just on-hand stock.

• Formula: On-Hand Inventory + On-Order Units - Backorders.
• Critical Distinction: Failing to include pipeline stock (units already ordered but not yet received) leads to double-ordering and excess inventory.
• Calculation: The order quantity at each review is exactly Order-Up-To Level - Current Inventory Position.

The risk horizon that the Order-Up-To Level must cover. It is the sum of the review period and the replenishment lead time.

• Formula: Review Period (R) + Lead Time (L).
• Why it matters: A stockout can only be corrected at the next review. The order placed now must cover demand until the subsequent order arrives.
• Example: If you review weekly (R=7 days) and lead time is 3 days (L=3), the Order-Up-To Level must cover 10 days of demand variability.

A special case of the Order-Up-To Level policy where the review period is continuous. An order is placed for exactly one unit every time a demand occurs.

• Relationship: The Base-Stock Level is functionally an Order-Up-To Level with R=0.
• Use Case: Ideal for high-value, slow-moving spare parts where continuous monitoring is feasible and batch ordering costs are negligible.
• Key Difference: No batching of demand occurs between reviews, eliminating the review period risk.

The component of the Order-Up-To Level that buffers against variability. The target level is decomposed into cycle stock and safety stock.

• Cycle Stock: Covers expected demand during the protection interval (R+L) * Avg Demand.
• Safety Stock: Covers demand variability, calculated as Z * σ * √(R+L), where Z is the service level factor and σ is demand standard deviation.
• Optimization: Dynamic safety stock algorithms continuously recalculate this buffer based on real-time demand sensing.

The inventory trigger used in continuous review systems, serving as the counterpart to the Order-Up-To Level in periodic systems.

• Mechanism: When inventory position drops to the ROP, a fixed quantity (Q) is ordered.
• Contrast with OUT: ROP uses a fixed order quantity (Q-system), while OUT uses a variable order quantity to reach a fixed target.
• Selection Driver: ROP is preferred when review costs are low (automated systems); OUT is preferred when ordering costs are high and coordination is needed.