Supply Chain Planning and Demand Forecasting

Chapter 9
Inventory, Service and Replenishment Planning

Post Graduate Program in Logistics
Supply Chain Planning and Demand Forecasting • Post Graduate

9. Inventory, Service and Replenishment Planning

Introduction: From Forecasts to Inventory Decisions

Demand forecasting does not end when a forecast is published. Its managerial value is realised only when the forecast is translated into decisions about inventory, replenishment, capacity and customer service. Every forecast contains uncertainty. Consequently, supply chain planning must determine not only how much demand is expected, but also how much inventory is required to protect customers against forecast error, lead-time variation and supply disruption.

Inventory planning is therefore a balancing problem. Excess inventory increases working capital, storage cost, insurance, obsolescence and markdown exposure. Insufficient inventory creates stockouts, lost sales, production interruptions and damaged customer relationships. The appropriate decision depends on the economics and strategic importance of the item, the variability of demand and supply, the replenishment lead time, and the service promise made to customers.

This chapter develops an integrated approach to inventory, service and replenishment planning. It connects forecast uncertainty with safety stock and service levels, explains ABC–XYZ segmentation, and examines economic order quantity, reorder-point, periodic-review, base-stock and multi-echelon policies. The emphasis is not on applying formulas mechanically, but on understanding the assumptions, trade-offs and managerial judgement behind them.

Learning Objectives

After studying this chapter, the learner should be able to:

1. Explain the relationship between demand uncertainty, inventory investment and customer service.

2. Distinguish cycle stock, safety stock, anticipation stock, pipeline stock and decoupling stock.

3. Apply ABC–XYZ analysis to differentiate inventory policies.

4. Calculate EOQ, reorder points, safety stock and periodic-review order quantities.

5. Compare continuous-review, periodic-review and base-stock replenishment systems.

6. Explain the logic of multi-echelon inventory planning and risk pooling.

7. Evaluate service-level, working-capital and operational trade-offs.

8. Identify implementation risks, technology requirements and governance considerations.

The Inventory–Service–Working Capital Relationship

Inventory serves several purposes. Cycle stock supports normal replenishment cycles; safety stock protects against uncertainty; anticipation stock is built before predictable peaks, promotions or supply interruptions; pipeline stock is in transit; and decoupling stock separates dependent activities so that a disruption in one operation does not immediately stop another.

A useful distinction is between inventory position and on-hand inventory. Inventory position is usually calculated as:

Inventory position = On-hand inventory + On-order inventory − Backorders

A replenishment decision based only on physical stock can be misleading when a large quantity is already in transit or when unfulfilled orders have accumulated.

The cost of inventory is broader than the purchase price. Relevant costs include ordering or setup cost, holding cost, shortage cost, transportation cost, handling cost and the opportunity cost of capital. Holding cost is often expressed as a percentage of annual inventory value and includes financing, insurance, storage, deterioration and obsolescence. In capital-intensive supply chains, inventory reduction can release significant working capital, but aggressive reduction may expose the organisation to stockouts.

Customer service has several meanings. Cycle service level is the probability that no stockout occurs during a replenishment cycle. Fill rate is the proportion of total demand supplied immediately from available stock. These measures are related but not identical. A system may avoid stockouts in most cycles yet experience substantial shortages during the few cycles in which a stockout occurs. Conversely, it may have a high fill rate while experiencing frequent but small stockouts. The selected measure must therefore match the commercial promise and operational reality.

Service levels should not automatically be set at 99 or 100 per cent. The final increments of service are usually expensive because they require disproportionately more safety stock. The appropriate target depends on margin, customer criticality, substitution possibilities, contractual penalties and the cost of shortage. A spare part supporting a safety-critical asset may justify a very high service target, while a low-margin, easily substituted retail item may not.

Forecast Uncertainty and Safety Stock

Safety stock exists because actual demand and replenishment lead time differ from their expected values. If demand during lead time were perfectly known, cycle stock alone would be sufficient. In practice, planners must estimate the distribution of demand during the protection period.

For a simple model with independent, normally distributed demand and constant lead time, safety stock may be estimated as:

Safety stock = z × σLT

where z is the standard-normal factor associated with the desired service level and σLT is the standard deviation of demand during lead time. Typical z-values are approximately 1.28 for 90 per cent, 1.645 for 95 per cent and 2.33 for 99 per cent cycle service.

If weekly demand has standard deviation σd and lead time is L weeks, then, assuming independent weekly demand:

σLT = σd√L

When both demand and lead time vary, a commonly used approximation is:

σLT = √(Lσd² + d̄²σL²)

where d̄ is average demand per period, σd is the standard deviation of demand per period, L is average lead time and σL is the standard deviation of lead time. This formula assumes independence between demand and lead time. If high demand systematically causes longer lead times, the covariance must be considered; ignoring it understates risk.

Forecast error should be measured at the level and horizon at which inventory decisions are made. Mean absolute deviation, mean absolute percentage error and root mean squared error can describe forecast performance, but no single measure directly determines safety stock. The standard deviation of forecast error during the replenishment lead time is more directly relevant. Historical demand variability should not be confused with forecast error: a stable but poorly forecast demand stream and a naturally volatile but accurately forecast stream require different treatment.

Safety stock can also be calculated using a target fill rate rather than a cycle service level. Fill-rate calculations require an estimate of expected shortage per cycle, often derived from the loss function of the demand distribution. In practice, organisations should validate statistical calculations against actual stockout history, demand intermittency and supply reliability.

Segmentation: Why One Policy Cannot Fit Every Item

The product portfolio may contain thousands of items with different values, demand patterns and strategic roles. Applying one replenishment rule to all items produces either excessive control effort or inappropriate service levels. Segmentation provides a disciplined basis for differentiated policies.

ABC analysis classifies items according to annual consumption value, commonly calculated as annual demand multiplied by unit cost. A-items account for a relatively large share of value and therefore merit close review, accurate data, frequent cycle counting and senior attention. B-items receive moderate control, while C-items often require simplified processes, larger review intervals or automated ordering. ABC categories are managerial conventions, not universal laws; boundaries should reflect the organisation’s portfolio.

XYZ analysis classifies items according to demand predictability. X-items have stable and forecastable demand, Y-items show moderate variation or seasonality, and Z-items have irregular, intermittent or highly unpredictable demand. A high-value Z-item presents a different challenge from a low-value X-item. The first may require postponement, supplier collaboration, risk pooling or criticality-based stocking; the second may be suitable for automated replenishment.

Combining the two dimensions creates a nine-cell matrix. AX items are valuable and predictable, making them suitable for close planning and precise continuous review. AZ items are valuable but difficult to forecast and should not be managed solely by increasing safety stock. Their policy may include make-to-order supply, customer allocation, substitution, strategic sourcing or executive approval. CX items are inexpensive and predictable, making automatic replenishment practical. CZ items may be purchased in response to demand, rationalised, pooled or managed with broad service commitments.

ABC–XYZ should be supplemented by other dimensions. Criticality, lead time, supplier risk, shelf life, product lifecycle, substitutability and regulatory importance may override a purely financial classification. For example, a low-value medical component can deserve more protection than its annual consumption value suggests.

Replenishment Policies and Their Logic

Economic Order Quantity

The economic order quantity model determines an order size that balances ordering cost and holding cost. Under the classical assumptions of constant demand, fixed ordering cost, constant lead time, no quantity discounts and immediate replenishment, the model is:

EOQ = √(2DS/H)

where D is annual demand, S is ordering or setup cost per order and H is annual holding cost per unit.

The model implies that annual ordering cost decreases as order quantity increases, while annual holding cost increases. EOQ minimises their combined cost. It does not determine when to order, nor does it automatically account for service-level requirements, capacity constraints, supplier minimums, truckload economics or price breaks.

For example, if annual demand is 24,000 units, ordering cost is 50 monetary units per order and annual holding cost is 4 monetary units per unit, EOQ is approximately 775 units. If the supplier’s minimum order is 1,000 units, the organisation must compare the total relevant cost and operational consequences rather than blindly applying 775.

EOQ is best treated as a benchmark. In modern supply chains, frequent replenishment, quantity discounts, shared transportation and supplier agreements often require an adapted total-cost model.

Continuous-Review Reorder-Point Systems

In a continuous-review, or (Q, R), system, inventory is monitored continuously. A fixed quantity Q is ordered when inventory position reaches the reorder point R. The basic reorder point is:

R = Expected demand during lead time + Safety stock

or, under constant average demand d̄ and lead time L:

R = d̄L + SS

The system is responsive and well suited to high-value, critical or fast-moving items for which stock information is available in real time. Its effectiveness depends on accurate transaction recording, reliable lead-time data and clear treatment of backorders, returns and inventory in transit.

Periodic-Review Systems

Under a periodic-review, or (P, S), system, inventory is reviewed at fixed intervals P. At each review, the order raises inventory position to an order-up-to level S. The protection period is the review interval plus lead time. Therefore:

S = Expected demand during (P + L) + Safety stock for (P + L)

The order quantity is:

Order quantity = S − Inventory position

Periodic review permits order consolidation and reduces administrative effort. It is useful where suppliers visit on fixed schedules or where several products are ordered together. Its disadvantage is that demand occurring between reviews cannot trigger immediate replenishment, so it generally requires more protection than an equivalent continuous-review system.

Base-Stock Policies

A base-stock policy attempts to maintain inventory position at a specified level after each demand event. The base-stock level generally covers expected demand over the replenishment lead time plus protection against uncertainty. Base-stock logic is particularly useful in high-frequency, short-cycle environments and in systems with rapid electronic signals. It can be implemented through point-of-sale data, automatic replenishment or vendor-managed inventory.

The policy is conceptually simple, but it can amplify upstream volatility if customer demand signals are distorted by promotions, order batching or artificial minimum quantities. Effective implementation requires visibility and disciplined ordering behaviour.

Min–Max and Practical Variants

Many organisations use a min–max policy. When inventory position falls below a minimum, the system orders enough to reach a maximum. This is operationally convenient and can approximate a periodic-review or reorder-point policy. However, the minimum and maximum values must be reviewed when demand, lead time, pack sizes or service requirements change. Static parameters are a common source of excess inventory and stockouts.

Worked Application: Reorder Point and Periodic Review

Consider a distribution centre supplying a product with average daily demand of 80 units and a daily standard deviation of 20 units. Supplier lead time is five days and is constant. The organisation wants a 95 per cent cycle service level, with z = 1.645.

Lead-time demand is 80 × 5 = 400 units. The standard deviation during lead time is 20√5, or approximately 44.7 units. Safety stock is 1.645 × 44.7, approximately 74 units. The reorder point is therefore about 474 units.

If the same product is reviewed every seven days, the protection period becomes 12 days. Expected demand is 80 × 12 = 960 units. The standard deviation is 20√12, or approximately 69.3 units. Safety stock is about 114 units, producing an order-up-to level of approximately 1,074 units. The periodic-review policy carries more protection because the firm must cover demand during the seven-day period before the next review.

This example assumes normal demand, independent observations and reliable data. If demand is intermittent, lead time varies or the service measure is fill rate, a different model may be more appropriate.

Multi-Echelon Replenishment and Risk Pooling

In a multi-echelon supply chain, inventory is held at several connected levels, such as suppliers, plants, regional distribution centres and stores. Optimising each location independently can produce excessive total inventory because every echelon protects itself against the same uncertainty. Multi-echelon planning seeks to position inventory across the network so that total cost and service objectives are achieved jointly.

A central distribution centre may hold strategic inventory while downstream locations carry only the stock needed for rapid response. Centralisation can create risk pooling: independent demand variability is aggregated, and the relative variability of total demand may decline. However, pooling is not always beneficial. Longer outbound delivery times, transportation constraints, product customisation and correlated regional demand can offset the advantage.

The placement of safety stock depends on the customer response requirement and the time available for upstream replenishment. If customers require same-day delivery, inventory must be positioned near them. If customers accept several days, stock can be concentrated upstream. The concept of echelon inventory measures stock at one stage together with downstream stock, enabling more coherent network decisions.

Multi-echelon optimisation requires reliable lead-time, demand, capacity and service data. It also requires governance: one party must own the network objective rather than allowing each location to maximise its own performance metric.

From Models to Managerial Decisions

Quantitative policies are useful only when aligned with commercial and operational realities. Managers should evaluate the following questions:

What service promise is being protected: cycle service, fill rate, response time or availability?
Is the item substitutable, critical, perishable or subject to obsolescence?
Which uncertainty dominates: demand, lead time, supply quantity or transportation?
Can the supplier provide smaller, faster or more reliable replenishment?
Is inventory visible and transaction data trustworthy?
Does the policy fit order multiples, truck capacity and warehouse constraints?
Who bears the cost of inventory and who benefits from improved service?

Service-level differentiation is commercially important. Strategic customers or critical production lines may justify higher targets, but the decision should be explicit. A high target applied indiscriminately creates working-capital pressure without necessarily improving customer value.

Inventory parameters should be reviewed through a structured process. Demand history and forecast error are analysed; lead-time performance is measured using actual receipt dates; service targets are agreed; parameters are calculated; exceptions are reviewed; and outcomes are monitored. Parameter governance should include version control, approval thresholds and documented overrides.

Technology, Analytics and Automation

Enterprise resource planning and advanced planning systems can automate inventory calculations, purchase requisitions, exception alerts and supplier collaboration. Point-of-sale data, radio-frequency identification, warehouse management systems and transport tracking improve inventory visibility. Machine-learning models can estimate nonlinear demand patterns, promotion effects and lead-time risk, but they do not eliminate the need for policy judgement.

Useful analytical capabilities include probabilistic forecasting, intermittent-demand methods, simulation, inventory optimisation and digital twins. Simulation is particularly valuable when assumptions such as normal demand, independent locations or constant lead time are unrealistic. A model can test alternative service targets, review intervals, supplier reliability and network structures before implementation.

Automation introduces risks. Poor master data can cause systematic over-ordering; algorithmic recommendations may reproduce historical bias; parameter changes may be opaque; and users may follow a system recommendation without understanding its assumptions. Human oversight, audit trails, exception management and clear accountability remain essential.

Challenges, Risks and Common Mistakes

A frequent mistake is adding safety stock without diagnosing the source of uncertainty. If variability results from unreliable suppliers, poor forecast processes or unstable production, operational improvement may be cheaper than permanent inventory. Another mistake is using average lead time when the upper tail of lead-time performance drives stockouts.

Current Developments

Several developments are reshaping replenishment planning. Supplier collaboration and shared forecasts improve visibility of upstream constraints. Control towers combine demand, inventory and logistics signals to identify emerging risks. Probabilistic planning increasingly replaces single-point forecasts with demand distributions and scenario ranges. Sustainability considerations are also entering inventory decisions: excess stock may create waste, while emergency shipments increase emissions. The best policy therefore considers financial, service, resilience and environmental consequences together.

Resilience does not mean maximising inventory everywhere. It may involve dual sourcing, postponement, modular product design, strategic buffers, flexible capacity and alternative transportation. Inventory is one resilience instrument among several, and its value depends on the nature and duration of the disruption.

Illustrative Case: Northstar Industrial Components

Background

Northstar Industrial Components is a fictional manufacturer of equipment used by food-processing plants. It stocks approximately 3,500 spare parts across a central warehouse and four regional depots. Customers value rapid delivery because a missing component can stop a production line.

Situation and Problem

Northstar’s reported overall fill rate was 96 per cent, but working capital had risen by 22 per cent in two years. The regional depots held overlapping safety stock, while several high-value parts still experienced stockouts. Planners used a fixed three-week lead time even though actual supplier lead times ranged from two to seven weeks. Sales staff also placed large end-of-month orders that distorted demand history.

Analysis

The company first separated cycle stock, safety stock and obsolete inventory. It measured forecast error at the replenishment horizon and calculated supplier lead-time distributions rather than averages alone. ABC–XYZ analysis identified high-value, stable items that could be planned accurately, high-value intermittent items requiring criticality assessment, and low-value predictable items suitable for automation.

The analysis showed that stockouts were concentrated in AZ and BZ items, not in the largest-volume products. Increasing safety stock for every item would have raised cost without solving the underlying problem. For critical AZ parts, the company assessed failure consequences, supplier alternatives and repair possibilities. For noncritical Z items, it introduced customer-order-based replenishment and substitution rules.

Approach and Solution

Northstar implemented continuous review for critical AX and AY items, periodic review for low-value C items, and a base-stock arrangement for fast-moving standard parts. It recalculated reorder points using forecast-error variability and observed lead-time variation. Regional depots were assigned target service levels based on customer response requirements, while selected slow-moving parts were pooled at the central warehouse.

Supplier scorecards began tracking lead-time variability, confirmed-date reliability and minimum-order compliance. Sales and operations planning introduced promotion and project-order overrides so that unusual demand was not treated as normal recurring demand. A monthly parameter-governance meeting reviewed exceptions and obsolete stock.

Outcome

In this illustrative case, Northstar reduced duplicated regional buffers and improved availability for critical parts without applying a blanket increase in inventory. The main improvement came from differentiating policies, correcting lead-time assumptions and pooling selected items. The company also recognised that the reported fill rate had hidden the severity of shortages on a small set of critical parts.

Learning Points

The case demonstrates five principles. First, aggregate service metrics can conceal important item-level failures. Second, forecast uncertainty and lead-time uncertainty must be modelled together. Third, ABC–XYZ segmentation is a starting point, not a complete policy. Fourth, network-level optimisation is superior to isolated depot decisions. Finally, inventory improvement requires data governance and cross-functional behaviour change, not only new formulas.

Best-Practice Checklist

Effective inventory and replenishment planning generally includes the following practices:

1. Define the service measure and customer promise before calculating stock.

2. Use forecast-error distributions at the relevant planning horizon.

3. Measure actual lead-time variability and supplier reliability.

4. Segment items by value, predictability, criticality and lifecycle.

5. Match review frequency and replenishment logic to segment characteristics.

6. Optimise inventory across the network rather than by location alone.

7. Include order multiples, capacity, shelf life and obsolescence in decisions.

8. Review parameters after structural changes, not only on a calendar basis.

9. Monitor both service and working-capital outcomes.

10. Maintain human oversight of automated recommendations.

Chapter Summary

Inventory planning converts uncertain forecasts into decisions about stock, service and replenishment. Cycle stock supports normal ordering cycles, while safety stock protects against demand and lead-time uncertainty. Service levels must be defined carefully because cycle service and fill rate measure different outcomes, and very high service targets can require disproportionately large investment.

EOQ balances ordering and holding costs but does not determine timing or guarantee service. Reorder-point systems trigger fixed quantities when inventory position reaches a threshold; periodic-review systems raise stock to an order-up-to level after scheduled reviews; and base-stock policies seek to maintain a target inventory position. ABC–XYZ analysis supports differentiated policies, while multi-echelon planning positions inventory across the network and exploits risk pooling.

Successful application depends on accurate data, realistic assumptions, supplier collaboration, parameter governance and managerial judgement. Technology and analytics can improve visibility and automation, but they do not replace criticality assessment or ethical responsibility. The central managerial task is to achieve the required customer service with the least total supply-chain risk and working-capital burden.

Key Terms

ABC analysis: Classification of inventory according to annual consumption value or another value-based measure.

Base-stock policy: A replenishment policy that maintains inventory position near a specified target level.

Cycle service level: Probability that no stockout occurs during a replenishment cycle.

Cycle stock: Inventory used to satisfy expected demand between replenishment orders.

Echelon inventory: Inventory at a location together with inventory positioned downstream from that location.

Economic order quantity (EOQ): Order quantity that balances relevant ordering and holding costs under stated assumptions.

Fill rate: Proportion of total demand supplied immediately from available inventory.

Inventory position: On-hand stock plus on-order stock minus backorders.

Lead-time demand: Demand expected during the replenishment lead time.

Pipeline stock: Inventory currently in transit between supply-chain stages.

Periodic-review system: Replenishment system in which inventory is reviewed at fixed time intervals.

Reorder point: Inventory-position level that triggers replenishment in a continuous-review system.

Risk pooling: Reduction in relative variability achieved by aggregating demand or inventory across locations or products.

Safety stock: Inventory held to protect against uncertainty in demand, lead time or supply.

Service level: Agreed probability or proportion describing inventory availability to customers.

Stockout: Situation in which available inventory cannot meet a demand requirement.

XYZ analysis: Classification based on demand predictability or variability.

Working capital: Funds committed to operating assets, including inventory, that support business activity.

Review Questions

Short Answer Questions

1. Distinguish cycle stock from safety stock.

2. Why are cycle service level and fill rate not identical?

3. What information is included in inventory position?

4. State the basic EOQ formula and identify its variables.

5. Why can a high-value Z item require a policy different from a high-value X item?

Long Answer Questions

1. Explain how forecast error and lead-time variability influence safety-stock requirements.

2. Compare continuous-review, periodic-review and base-stock replenishment systems.

3. Discuss how ABC–XYZ analysis can be extended using criticality, lifecycle and supplier-risk information.

4. Explain the advantages and limitations of multi-echelon inventory planning.

5. Discuss the governance, data and behavioural conditions required for successful automated replenishment.

Analytical / Application Questions

1. Daily demand averages 100 units with a standard deviation of 25 units. Lead time is four days and is constant. Calculate lead-time demand, safety stock and reorder point for a 95 per cent cycle service level using z = 1.645.

2. A firm reviews inventory every 10 days, and supplier lead time is six days. Explain why its order-up-to level must protect against 16 days of demand rather than only six days. What operational disadvantage does this create compared with continuous review?

3. A low-cost spare part has intermittent demand but is essential to prevent a production shutdown. Recommend an inventory policy and justify why ordinary ABC classification alone would be inadequate.

22. Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes cycle service level from fill rate?

A. Cycle service level measures the probability of no stockout in a cycle, while fill rate measures the proportion of demand supplied immediately.

B. Cycle service level measures inventory value, while fill rate measures ordering cost.

C. Cycle service level applies only to suppliers, while fill rate applies only to customers.

D. They are two names for exactly the same measure.

Correct Answer: A. Cycle service level measures the probability of no stockout in a cycle, while fill rate measures the proportion of demand supplied immediately.

Question 2

Average daily demand is 80 units, lead time is five days, and safety stock is 74 units. What is the reorder point?

A. 154 units

B. 400 units

C. 474 units

D. 800 units

Correct Answer: C. 474 units

Question 3

In a periodic-review system, why is the protection period normally equal to review interval plus lead time?

A. Because holding cost is calculated twice.

B. Because demand must be covered until the next review and through subsequent replenishment lead time.

C. Because periodic systems do not use safety stock.

D. Because suppliers always require two deliveries.

Correct Answer: B. Because demand must be covered until the next review and through subsequent replenishment lead time.

Question 4

Which item is most likely to be classified as AZ in an ABC–XYZ matrix?

A. Low-value and highly predictable

B. High-value and highly predictable

C. Low-value and moderately seasonal

D. High-value and highly unpredictable

Correct Answer: D. High-value and highly unpredictable

Question 5

What is the principal purpose of EOQ?

A. To determine the customer service level

B. To balance ordering cost and holding cost under specified assumptions

C. To calculate forecast error

D. To determine supplier lead-time variability

Correct Answer: B. To balance ordering cost and holding cost under specified assumptions

Question 6

If weekly demand standard deviation is 20 units and lead time is four weeks, assuming independent demand, what is the lead-time standard deviation?

A. 20 units

B. 40 units

C. 80 units

D. 160 units

Correct Answer: B. 40 units

Question 7

Which action most directly addresses safety stock caused by unreliable supplier lead times?

A. Use only average demand in the forecast

B. Measure and model lead-time variability and improve supplier reliability

C. Reduce the service target for every item

D. Ignore inventory in transit

Correct Answer: B. Measure and model lead-time variability and improve supplier reliability

Question 8

What is a major potential benefit of multi-echelon inventory planning?

A. It guarantees zero stockouts at no cost.

B. It eliminates the need for demand forecasts.

C. It can reduce duplicated buffers and position stock using network-wide risk pooling.

D. It requires every location to hold identical safety stock.

Correct Answer: C. It can reduce duplicated buffers and position stock using network-wide risk pooling.

Question 9

Why may a normal-distribution safety-stock model be inappropriate for intermittent demand?

A. Intermittent demand contains many zero-demand periods and irregular order sizes.

B. Intermittent demand always has constant lead time.

C. Normal models cannot calculate any average demand.

D. Intermittent demand has no financial value.

Correct Answer: A. Intermittent demand contains many zero-demand periods and irregular order sizes.

Question 10

Which is the best governance practice for automated replenishment parameters?

A. Allow parameters to remain unchanged permanently.

B. Review service, demand, lead-time and exception data with documented approvals.

C. Permit sales staff to override every recommendation without records.

D. Optimise warehouse utilisation without monitoring customer service.

Correct Answer: B. Review service, demand, lead-time and exception data with documented approvals.

Supply Chain Planning and Demand Forecasting
Chapter 9 of 15

Chapter 9 Inventory, Service and Replenishment Planning