Scma 444 - Exam 2 Cheat Sheet
Scma 444 - Exam 2 Cheat Sheet
When to Reorder?: In a continuous review policy, a reorder is triggered whenever the inventory position reaches a specific limit. When the demand rate D is constant: Reorder point = Lead time x Demand rate ROP = L x D
ReOrder Point:: Continuous review Policy - also known as (𝑄, 𝑅) policy, a reorder is triggered whenever the inventory position reaches a specific limit. When the demand rate D is constant: Reorder point = Lead time x Demand rate [ROP = L x D] - When
Demand during LT is uncertain, ROP = LxD is not good enough -> use[ ROP = LxD + SS] (safety stock). ROP under Uncertainty: Demand or lead time uncertainty creates the possibility that demand will be greater than available supply Safety stock -
Stock that is held in excess of expected demand due to variable demand and/or lead time. It can be positive or negative depending on the relationship between excess and stockout costs The service level is the probability that the there will be no
stockout within a time interval. [ROP = Expected demand during LT + SS] *Desired service level in the distribution of demand during lead time determines SS (?). How Much Safety Stock? The amount of safety stock that is appropriate for a given
situation depends upon: 1. Demand and lead time variability 2. The desired service level * Service Level: The probability that demand will not exceed supply during lead time = = 100% - Stock-out risk. [ ROP = Expected demand during LT + zL WHERE z =
# std devs associated w/ target cycle Service level & L= std dev of demand during LT ] \ Safety Inv & Service Level: The service level is measured by the probability that the actual leadtime demand will not exceed the reorder point. [SL = Prob(LT demand
<= ROP)… SS = ROP-DL = z x L & z = (ROP-DL )/ L ] Cumulative Standard Normal Distribution - Finding z value given a service level: If we want 85% service level: From z table -> z = 1.035 Then SS can be obtained SS =z *𝜎𝐿 Finding service level given z
value: For example, if z = 1.28, SL = 0.8997 ≈ 0.90 Finding SL given ROP: So to find the service level for a given reorder point, calculate: [𝑧 = 𝑅𝑂𝑃−𝐷𝐿 / 𝜎𝐿] and look up z in the table on the right.\ Excel Functions: Given ROP, to find the associated service
Level =NORM.DIST(ROP, DL , 𝜎𝐿 , 1) For instance, suppose ROP = 112.8, DL = 100 and 𝜎𝐿=10. What is SL? SL = NORM.DIST(112.8, 100, 10, 1) ≈0.9 Given SL, to find ROP =NORM.INV(SL, DL , 𝜎𝐿 ) For instance, target SL = 0.9, DL = 100 and 𝜎𝐿=10.
What is ROP? ROP = NORM.INV(0.9, 100, 10, 1) ≈112.8 Calculation of Lead-time Demand Variability:: Variability in replenishment leadtime demand can result from variability in the demand rate, variability in the replenishment leadtime, or both.
Replenishment leadtime demand standard deviation is calculate as: Idnetifying the distribution of customer demand during LT when LT and Periodic Demand are uncertain, demand during LT is normally distrubted w/ a mean of DL and std dev of 𝜎𝐿 => [
𝜎𝐿 = √𝐿𝜎𝐷2 + 𝐷 2𝑆𝐿2 where D: average demand per period : 𝜎𝐷 is the standard deviation of the demand per period L: Average lead time for replenishment S L: standard deviation of the lead time. ]\\ Pooling Efficiency Through Aggregation: Physical
Centralization Physical centralization (vs. decentralization) of inventory to serve customers eliminates the possibility of stock imbalance and reduces the coefficient of variation of replenishment leadtime demand. SSC = z x √𝑁 x𝜎𝐿 (assuming each
warehouse has the same D and 𝜎𝐿𝑇𝐷 ) \\ SC Goals: (1) reduce supplier LT (L) -> if lt decreases by a factor of k the required safety inventory decreases by a factor of √k, (2)reduce underlying uncertainty of demand ( 𝜎𝐷 ) -> if uncertainty 𝜎𝐷 is
reduced by a factor of k the required safety inventory also decreases by a factor of k.
Insights on EOQ Robustness of EOQ: Notice that the annual cost is very flat in the neighborhood of the optimal order quantity. So, even if H, S, D, etc. are just estimated, as long as they are reasonably close, the true
optimal order quantity should be close to the calculated Q*. Purchase Cost (CxD) does not change with the order size that we choose, so it is constant here. (it does not affect EOQ which is a result of trade off between
ordering and holding costs) If ordering cost is larger than holding cost, we know that the order size is probably too small and we order too frequently. If order cost is less than holding cost, we know that the order size
is probably too large and we old too many inventories on average. At EOQ, Holding cost = Ordering cost If demand increases by a factor of 𝑘, Q* increases by a factor of 𝑘 . That is, when your demand increases by a
factor k, you don’t need k times more inventory. Price Discounts: In the case of quantity discounts (price incentives to purchase large quantities), the unit cost, C, varies by order quantity and is a crucial element in
determining the EOQ. Quantity discount policies: With an all unit quantity discount policy, the buyer receives a discount on all units purchased whenever the quantity purchased exceeds a certain threshold. EXAMPLE:
With price discounts using an all quantity discount policy, the following two-step process determines the EOQ: Beginning with lowest price, calculate the EOQ for each price level until a feasible EOQ is found. It is
feasible if it lies in the order quantity range corresponding to its price. If the first feasible EOQ found is for the lowest price level, this quantity is best. Otherwise, calculate the total cost for the first feasible EOQ and for the
larger price break quantity at each lower price level. The quantity with the lowest total cost is optimal. *if an order placed is at least as large as qi but smaller than qi+1, each unit is obtained at a cost of Ci.. unit cost
decreases as qty ordered increases* \\\ With an incremental (marginal) quantity discount policy, a buyer receives a discount only on additional units purchased above a certain threshold value. Define 𝑉𝑖 as the cost of 𝑞𝑖
units i.e. ordering up to the entire (𝑖 − 1)-th discount block 𝑉𝑖 = 𝐶0 (𝑞1 − 𝑞0) + 𝐶1 (𝑞2 − 𝑞1) + ⋯ + 𝐶𝑖−1(𝑞𝑖 − 𝑞𝑖−1) The cost of 𝑄 units in the range between 𝑞𝑖 and 𝑞𝑖+1 is 𝑉𝑖 + (𝑄 – 𝑞𝑖)𝐶𝑖 that is, the cost of all the units of
previous discount blocks and a portion of the 𝑖-th block.. The costs are Annual order cost =(𝐷/𝑄)*𝑆 Annual holding cost = (𝑉𝑖 + (𝑄 – 𝑞𝑖) 𝐶𝑖)( ℎ/2).. Annual material cost = (𝐷/𝑄)* (𝑉𝑖 + (𝑄 – 𝑞𝑖) 𝐶𝑖) Note that the only
difference from the original costs is that the per unit cost 𝐶 is now (𝑉𝑖+ 𝑄 −𝑞𝑖 𝐶𝑖)/Q. As Usual, the total annual cost is the sum of these three components: 𝑇𝐶𝑖 = (𝐷/𝑄)*𝑆 + (𝑉𝑖 +( 𝑄 – 𝑞𝑖)*𝐶𝑖)*ℎ)/2 + (𝐷/𝑄)* (𝑉𝑖 + (𝑄 –
𝑞𝑖)*𝐶𝑖). THEN you have to recalculate the optimal lot size for price Ci Qi = SQRROOT([2D( S+Vi – qi*Ci)] / h*Ci) **basically use this calculation for each option, then pick the best option {or just pray this shit isn’t on the
test}* Ordering with Multiple Products: Three approaches (Aggregation in general works better when joint setup cost 𝑆 is large relative to 𝑠i)
1. No aggregation Products are ordered independently *use normal EOQ on each product individually*
2. Complete aggregation Products are always ordered jointly Pro: Easy to administer and may reduce costs
Con: May not be effective to order low-demand and high-demand items at the same frequency
Deriving the formula: It is easier to rederive the formula in terms of the order frequency 𝑛 since we now have a lot size 𝑄𝑖 for each item.. Let 𝑆𝑐 be the total fixed order cost: 𝑆𝑐 = 𝑆 + ∑𝑖∈𝑃 𝑠𝑖 (p = set of products). Annual
𝑄 ∑𝑖∈𝑃 𝐷𝑖 𝐻𝑖 ∑𝑖∈𝑃 𝐷𝑖 𝐻𝑖 ∑𝑖∈𝑃 𝐷𝑖 𝐻𝑖
ordering cost : 𝑆𝑐 𝑛 Annual holding cost : (recall 𝑛 = 𝐷𝑖 /𝑄𝑖 ) ∑𝑖∈𝑃 ( 𝑖) 𝐻𝑖 = Total annual cost: 𝑆𝑐 𝑛 + Taking derivative of 𝑆𝑐 𝑛 + w.r.t. to 𝑛 and setting it to zero yields the optimal frequency
2 2𝑛 2𝑛 2𝑛
∑𝑖∈𝑃 𝐷𝑖 𝐻𝑖
𝑛∗ = √ Lot sizes for each item may be obtained from 𝑄𝑖 = 𝐷𝑖 /𝑛
2𝑆𝑐
𝑖
∗
th order of product 𝑖 ∗ . To calculate this frequency 𝑚𝑖 , first calculate frequency of the most ordered product 𝑖 now considering the rest of