434. Backorders

434.1. EOQ with planned backorders

Relax one dimension from basic EOQ: excess demand is no longer forbidden — stockouts are allowed at penalty rate per unit per year. Now there are two decisions: order quantity and maximum backorder .

434.1.1. Setup

New variables (beyond basic EOQ):

The inventory profile in each cycle:

Two area integrals over a cycle, divided by the cycle length :

434.1.2. Cost model

Total relevant cost (drop the constant purchase-cost term):

Two changes from basic EOQ TRC, both highlighted in red: the holding term now sees instead of (less average on-hand), plus a new shortage term.

434.1.3. Derive first (FOC in )

Holding fixed, take :

Set to zero, multiply by :

So the backlog is the share of the order quantity.

434.1.4. Derive (FOC in , after substituting )

Take and set to zero (algebra for the holding term uses the product rule):

Multiply by :

Substitute so :

Take the square root:

The first factor is the basic EOQ; the second is a multiplier that grows the order when backorders are cheap (small ).

434.1.5. Final formulas

Sanity check: as (backorders prohibitively expensive), the multiplier and — recovers basic EOQ. As (backorders are free), and .

Example

Given (shared EOQ params + a backorder penalty):

  • Annual demand: units/year
  • Order cost: = $50 / order
  • Holding cost: = $2 / unit / year
  • Backorder penalty: = $8 / unit / year

Step 1 — backlog share

So the optimal backlog is 20% of the order quantity.

Step 2 — order quantity

Step 3 — backlog and cost

Compare to basic EOQ on the same inputs:

  • Basic EOQ: , $1549
  • With backorders: (larger), , $1386 (10% lower)

Allowing planned backorders is worthwhile here because is only — backorders cost a bit more than holding but not so much that we want zero backlog.