1.
Economic Order Quantity (EOQ) Model
The assumption is that the demand (D) is known and uniform. Parameters needed for the
analysis are: ordering cost (S), inventory holding cost (H), purchasing price (p) (can be
omitted from the analysis). The decision is to find the optimal order quantity Q0.
2DS Q
Q0 = Length of order cycle =
H D
Q D D
Total Cost = Dp + H+ S Number of orders =
2 Q Q
2. Continuous Review: ROP Policy ≈ (Q,R) Model
Demand rate is random and normally distributed with average µ and standard deviation σ;
lead time is L, which can be known or random. We assume service level is 1 - α.
Structure of the optimal policy: order Q units when on hand inventory drops to R.
Order quantity Q is either a fixed value provided, or can be obtained from the EOQ analysis
where D is replaced by the average demand. Reorder point R can be obtained as follows:
R = µ dLT + z1−α σ dLT ,
where µ dLT and σ dLT are , respectively, the average and standard deviation of demand
during lead time, and be found from the table below.
Cases µ dLT σ dLT
Constant Lead Time µL σ L
Variable Lead Time µ.Ave(L) σ 2 . Ave(L) + µ 2 . Var(L)
The Safety Stock (SS), Average Inventory (I) of this system are:
Q
SS = z1−α σ dLT , I = SS +
2
Assuming the shortage cost of K, the annual total cost (TC) of this inventory system can be
obtained as follows:
TC = Ordering + Holding + Shortage Cost
!D$ !Q $ !D$
= # &S + # + SS & H + # & α K
"Q% "2 % "Q%
* If the shortage cost is assessed on the number of items short per year (instead of the
number of shortage cycles), then the last term above should be replaced by [E(N) K] where
E(N) is obtained from Section 4.2 of this document.
3. Periodic Review: Base Stock Policy ≈ (r,B) Model
Demand rate is random and uniformly distributed with average µ and standard deviation σ.
The review period is r and lead time is L. We assume service level is 1 - α.
Structure of the optimal policy: at each review point order enough such that inventory
position reaches base stock level B.
Order quantity is random with an average of µr. Base stock B can be obtained as follows:
B = µ (r + L) + z1−α σ r + L
The Safety Stock (SS), Average Inventory (I) of this system are:
rµ
SS = z1−α σ r + L , I = SS +
2
Assuming the shortage cost of K, the annual total cost (TC) of this inventory system can be
obtained as follows:
TC = Ordering + Holding + Shortage Cost
!1$ ! rµ $ !1$
= # &S + # + SS & H + # & α K
"r% "2 % "r%
* The above formula assumes that r is based on years. Otherwise 1/r should be replaced by
the number of review periods per year.
4. Service Level
4.1. Type 1 (1 - α):
Can be found using the Normal Distribution Table.
Excel formulas are:
z1- α = [Link](1- α, 0, 1) , 1 - α = [Link](z, 0, 1, 1)
4.2. Type 2 ≈ Annual Service Level ≈ Fill Rate (SLannual):
E(n) = expected number of units short per order cycle,
E(N) = expected number of units short per year. Then,
D
E(n) = E(z) σ dLT , E(N ) = E(n)
Q
where E(z) can be found from the table. Alternatively you can use the following Excel
formula to find E(z):
E(z) = [Link](z, 0, 1, 0) – z * ([Link](z, 0, 1, 1))
Finally, the annual service level can be found as follows:
E(z) σ dLT E(n) E(N )
SLannual = 1− = 1− = 1−
Q Q D
5. Single Period (Newsvendor) Model
Let the underage (shortage) cost Cu be the cost of procuring one too few items, and the
overage (excess) cost Co be the cost of procuring of one too many items. For instance if the
product has a sales revenue of r, purchase cost of c, and salvage revenue of s, then
𝐶" = 𝑟 − 𝑐, and 𝐶) = 𝑐 − 𝑠
Structure of the optimal policy: order Q* units such that
,-
Service Level = 𝛼 =
,- .,/
To find the optimal order quantity we consider two cases.
5.1. Discrete Demand Distributions:
Demand takes values of Di with probability pi for i = 1, …, n, where D1 < … < Dn. In this
case, simply form the cumulative probability distribution table and find a demand value for
which the cumulative probability is at least 𝐶" (𝐶" + 𝐶) ). That is, the optimal order
quantity is Q* = DJ where J is the smallest index for which 𝑝4 + ⋯ + 𝑝6 ≥ 𝐶" (𝐶" + 𝐶) ).
Furthermore, for a given order quantity Q, the expected profit can be found as follows
n
Expected Profit = ∑ pi "#r ( min {Q, Di }) − cQ + s (Q − min {Q, Di })$%
i=1
5.2. Continuous Demand Distributions:
Demand is random and uniformly distributed with average µ and standard deviation σ. In
this case the optimal order quantity is
𝑄∗ = 𝜇 + 𝑧 ,- . 𝜎
,- .,/
Alternatively you can use the following excel formula to find the optimal Q
,-
𝑄 ∗ =[Link] ( , 𝜇, 𝜎)
,- .,/
6. Supply Contracts Cost Formulas (Discrete Demand Dist.)
Let r be the distributor’s sales revenue, and c the manufacturer’s production cost. The item
has a salvage value of s. Suppose demand takes values of Di with probability pi for i = 1, …,
n, where D1 < … < Dn.
6.1. Supplier on MTO, Distributor on MTS, Sequential Optimization:
Let w be the wholesale price, and let Q represent the distributor order quantity. Then
n
Distibutor Expected Profit = ∑ pi "#r ( min {Q, Di }) − wQ + s (Q − min {Q, Di })$%
i=1
Supplier Expected Profit = ( w − c ) Q
6.2. Supplier on MTO, Distributor on MTS, Buyback Contract:
Let w and b be the wholesale and buyback prices, respectively. Then
n
Distibutor Expected Profit = ∑ pi "#r ( min {Q, Di }) − wQ + b (Q − min {Q, Di })$%
i=1
n
Supplier Expected Profit = ( w − c ) Q + ∑ pi "#( s − b) (Q − min {Q, Di })$%
i=1
6.3. Supplier on MTO, Distributor on MTS, Revenue Sharing:
Let w be the (new) wholesale price, and α fraction of revenue received by the manufacturer.
n
Distibutor Expected Profit = ∑ pi "#( r − α r ) ( min {Q, Di }) − wQ + s (Q − min {Q, Di })$%
i=1
n
Supplier Expected Profit = ( w − c ) Q + ∑ pi "#α r ( min {Q, Di })$%
i=1
6.4. Supplier on MTS, Distributor on MTO, Sequential Optimization:
Let w be the wholesale price, and let Q represent the supplier production quantity. Then
n
Distibutor Expected Profit = ∑ pi "#( r − w ) ( min {Q, Di })$%
i=1
n
Supplier Expected Profit = ∑ pi "#w ( min {Q, Di }) − cQ + s (Q − min {Q, Di })$%
i=1
6.5. Supplier on MTS, Distributor on MTO, Payback Contract:
Let w and b be the wholesale and payback prices, respectively. Then
n
Distibutor Expected Profit = ∑ pi "#( r − w ) ( min {Q, Di }) + ( s − b) (Q − min {Q, Di })$%
i=1
n
Supplier Expected Profit = ∑ pi "#w ( min {Q, Di }) − cQ + b (Q − min {Q, Di })$%
i=1
6.6. Supplier on MTS, Distributor on MTO, Cost Sharing Contract:
Let w be the (new) wholesale price and α fraction of the prod. cost paid by the distributor.
n
Distibutor Expected Profit = −α cQ + ∑ pi "#( r − w ) ( min {Q, Di })$%
i=1
n
Supplier Expected Profit = ∑ pi "#w ( min {Q, Di }) − (c − α c)Q + s (Q − min {Q, Di })$%
i=1
