Skip to content
Back to search results

Inventory Routing

Introduced by
Coelho, L. C., Cordeau, J. F., & Laporte, G.
Year
2014
Type
Optimization problem
Family
Distribution and Logistics
Font size

What does it do?

The Inventory Routing Problem (IRP) is an optimization problem that jointly determines inventory levels at customer locations, delivery routes, and shipment quantities to minimize total logistics and inventory holding costs.

Rather than treating inventory management and vehicle routing as separate decisions, IRP recognizes that they are interdependent: larger shipments reduce routing costs but increase inventory holding costs, and vice versa. IRP is solved using mixed-integer programming, heuristics, and metaheuristics, and is a cornerstone of vendor-managed inventory (VMI) programs.

The intuition behind the method

Imagine a beverage distributor supplying 50 convenience stores. Traditional approaches answer two questions separately: (1) How often should each store be replenished? (2) What route minimizes driving distance? These separate decisions are suboptimal. IRP instead asks: 'What shipment frequency and quantity for each store, combined with efficient routing, minimizes total cost (storage + transportation)?' If you visit Store A frequently with small quantities, you incur high transportation costs but low inventory. If you visit infrequently with large shipments, storage costs spike. IRP balances these trade-offs jointly. For the distributor, this might mean visiting Store A every 2 weeks with 200 units (rather than weekly with 100 units) if the route efficiency gain outweighs the inventory increase.

History

IRP emerged in the 1980s–1990s in the operations research community as an extension of vehicle routing problems (VRP) and inventory management. Seminal work by Bell et al. (1983) and Campbell et al. (1998) formalized the problem and developed solution algorithms. IRP gained practical prominence in the 2000s with vendor-managed inventory adoption in retail supply chains. Today, major beverage and consumer goods companies (Coca-Cola, Procter & Gamble, Nestlé) use IRP-based systems for route and inventory planning.

When to use it

IRP is valuable for industries with significant joint inventory-routing costs: fast-moving consumer goods (beverages, snacks, deliveries), spare parts distribution, fuel delivery, and rental fleet management. Apply IRP when demand is relatively stable (enabling multi-period planning), customer locations are fixed, and transportation cost is significant. Less effective in highly dynamic environments or when products are bulky (full-truckload constraints limit flexibility).

Limitations and considerations

  • Requires accurate demand forecasting; forecast errors propagate to suboptimal routing and inventory decisions
  • Assumes relatively static customer base and stable demand patterns; high volatility or frequent customer additions complicate planning
  • IRP is computationally complex; practical algorithms are heuristics that may not find globally optimal solutions
  • Real-world constraints (vehicle availability, driver regulations, customer delivery windows) may not be fully captured in the model

Source conditions do not replace independent scientific review or validation by an analysis engine.

Strengths

  • Reduces total logistics costs by optimizing the trade-off between inventory holding and transportation
  • Improves service levels by enabling more frequent, smaller replenishments without proportional cost increases
  • Enhances customer satisfaction through reduced stock-out risk and lower shelf age (fresher products)
  • Enables data-driven decision-making: systematic evaluation of 'how often and how much' replaces intuitive rules
  • Provides foundation for vendor-managed inventory programs, strengthening customer relationships

Applications

  • A soft drink distributor applies IRP to optimize routes for 200 retailers across a metropolitan area, reducing transportation costs by 15 percent while improving delivery frequency from weekly to bi-weekly without stock-outs
  • A pharmaceutical wholesaler uses IRP to balance inventory at 500 pharmacies, reducing average inventory by 10 percent while maintaining 98 percent in-stock performance
  • A heating oil supplier applies IRP to plan deliveries for 1,000 residential customers, using weather forecasts and historical consumption patterns to optimize delivery routes and tank fill-ups
  • A food distribution company uses IRP to manage fresh product delivery to grocery chains, optimizing for product freshness, shelf space constraints, and transportation efficiency

How it works

How it works

1

Forecast customer demand over the planning horizon (typically 1–4 weeks). Gather data on transportation costs, vehicle capacity, customer inventory capacity, and holding costs.

2

Determine inventory constraints: maximum stock levels at each customer (shelf space, safety regulations), minimum order quantities, and desired service levels.

3

Decide which customers to visit on each day/route, the quantity to deliver to each, and the sequence (routing) to minimize distance or time.

4

Implement the plan: deliver shipments, record actual demand, update inventory levels.

5

Use optimization algorithms (mixed-integer programming, genetic algorithms, ant colony optimization) to find a near-optimal plan. Iterate with updated forecasts.

Common pitfalls

  • Over-relying on a static IRP plan without frequent updating; real demand deviations require plan adjustment to remain cost-effective
  • Ignoring customer preferences; even an optimal plan that violates customer expectations (e.g., specific delivery days) will fail
  • Optimizing cost at the expense of service; excessive inventory lows or missed deliveries erode customer goodwill and increase churn

Frequently asked questions

How is IRP different from vendor-managed inventory (VMI)?

VMI is a business arrangement where the supplier manages inventory at customer locations; the supplier decides when and how much to deliver. IRP is the optimization problem underlying VMI: determining the best delivery quantities and routes given the VMI arrangement. IRP is the technical solution; VMI is the business model.

What is the planning horizon for IRP?

Typically 1–4 weeks, long enough to smooth short-term demand variability but short enough that forecasts remain accurate and customer needs don't change dramatically. Longer horizons (e.g., monthly) may reduce transportation cost but increase inventory; balance is needed.

Can IRP handle multiple products per delivery?

Yes. Modern IRP formulations explicitly include product dimensions, constraints (e.g., temperature control), and customer product preferences. Vehicles may be shared across product categories or dedicated. Algorithms adjust complexity accordingly.

How do we validate an IRP solution before implementation?

Simulate the plan: project inventory levels given demand uncertainty, check for stock-outs, measure total cost. Conduct sensitivity analysis: how does cost change if demand is 10 percent higher? If a customer refuses delivery on a certain day? Robust solutions perform well under variation.

Related methods

Sources

  1. Coelho, L. C., Cordeau, J. F., & Laporte, G. (2014). Thirty years of inventory routing. Transportation Research Part B: Methodological, 55, 28-67.

    Open DOI reference
    Copied.
  2. Campbell, A. M., & Savelsbergh, M. W. (2011). Vehicle routing with time windows for inventory routing problems. Operations Research, 59(2), 500-515.

    Open publication reference
    Copied.