Professional Documents
Culture Documents
• Deterministic Demand
1. Static Demand
2. Dynamic Demand
• Probabilistic Demand
1. Stationary Demand
2. Non-Stationary Demand
INVENTORY MODELS
• Types of Inventory Models
1. Deterministic Demand Inventory Models
- Annual Demand is known priory
2. Probabilistic Demand Inventory Models
- Stochastic Demand with known Distribution
- Stochastic Demand with unknown Distribution
1. DETERMINISTIC DEMAND INVENTORY MODELS
𝑇 𝐿
𝑡
Fig. 1: Inventory level vs. Time [3]
1. DETERMINISTIC DEMAND INVENTORY MODELS
Economic Order Quantity
𝑆
𝑇
Fig 2: Inventory level vs. Time [4]
2. PROBABILISTIC DEMAND INVENTORY MODELS
EO
Q
• Total cost (q, D) = h *max (0, q-D) + b *max (0, D-q) Mini
mum
Cost
• Total cost (q) = E [h *max (0, q-D) + b*max (0, D-q)]
Stocking
Quantity
• Total cost (q) = Q*
• ∈ ⊆ Sϵ
• where , is known as right one sided
and left one sided derivatives respectively
• Where is relative regret
• ; where at ,
𝒆𝒙𝒑 (𝛍=𝟏𝟎𝟎) 𝑼
( 𝑨=𝟎 , 𝑩=𝟏𝟎𝟎 )
MINIMUM REQUIRED SAMPLE SIZE USING VARIOUS
PROBABILISTIC BOUND FOR NORMAL DISTRIBUTION
Sample Size for various Distributions using SAA Approach
Normal Distribution
Uniform Distribution
• Shortage cost = units short in that month * (selling price of the product -
Ordering
Cost
cost price of the product)
∗ Stocking
• Purchase cost = purchase cost/unit * number of units purchased in the 𝑄 Quantity
Fill orders
Order units of stock
Stock in hand
Stock in hand
Demand Region
(c) Fill Demand
Location
Inventory Position
Mobile Module
•Initialization:
initialize in period t = and a counting n = 0
Main Step Suppose for some i ∈ {1,……..,m} for each subsequent period t =
• Case 1 : let j = min{ k : denote the lifetime of the oldest inventory in period
t if no outdating occur in period t-1nand set
• Case 2 in the period t-1 if outdating occurs then;
(i) set if i =1 and set n:= n+1
(ii) Else set
Cycle gradient :
Figure 7 :u(t ) is updated for product lifetime 3
CYCLIC UPDATE POLICY 2 (CUP2):
• }
• Case 1: at the end of day t see the outdating ( and lost sale demand is realized. If in period t outdating is occur i.e.
earlier
• Case2: at the end of day t see the outdating ( and lost sale demand is realized. If in period t lost sales occur earlier
then outdating i.e
• Case3: at the end of day t see the outdating ( and lost sale demand is realized. If in period t lost sales lost sales and
outdating occur simultaneously i.e.
COMPARISON OF COST BETWEEN DYNAMIC SAA
POLICY AND CYCLE UPDATE POLICY 1
exponential demand distribution : 300 products per day Demand mean 300 , Range 800
COMPARISON OF COST BETWEEN CYCLE
UPDATE POLICY 1 AND CYCLE UPDATE POLICY 2
exponential demand distribution : 300 products per day Demand mean 300 , Range 800
COMPARISON OF COST BETWEEN CYCLE
UPDATE POLICY 1 AND CYCLE UPDATE POLICY 3
exponential demand distribution : 300 products per day Demand mean 300 , Range 800
TIME
HEURISTIC TO MANAGE PERISHABLE INVENTORY WITH
POSITIVE LEAD TIME AND TIME VARYING DEMAND
(BASE POLICY WITH LEAD TIME) POLICY (A) [9]
• If
Then = integer [
DEMAND WITHDRAWAL POLICIES[9]
• First
in first out (FIFO) policy for r = 1,………m , 1.on day t the amount of withdrawal
}
• Last in first out (LIFO) policy for r = m, m-1 ,…….1.on day t the amount of withdrawal
}
• Inventory replenishment decisions: for r = 2,….,m
}
=-
Demand withdrawal by FIFO policy
}
POLICY (B) CALCULATE ON OF OPTIMAL ORDER QUANTITY
WITH DYNAMIC SAA POLICY[9][1]
• If
Then = integer [
MODEL TO CALCULATE OPTIMAL ORDER QUANTITY CAPTURING
LEAD TIME FOR POISSON DISTRIBUTED DEMAND[8]:
• Optimal
order quantity according to this model which will capture poisson demand and that can
minimize cost over long time can be given as:
• Here denotes the probability of wasting an item and calculated in following way:
• Here derived order quantity also depends upon expected shortage and which is calculated as
formulated below
• ES = [lam*T(1-
STEPS FOR POISSON POLICY[8]:
•• Guess the initial value of S recommended to start initially with mean of
past demand
• For Poisson demand mean demand = lam*T = lam*L/m set
• Apply this condition until
• Parameters
Underage cost, b = pselling - ppurchasing + Cgoodwill
Overage cost , h = ppurchasing + Cdisposal + Cholding – Vsalvage
Cholding = 1, Cdisposal – Vsalvage = 2 , ppurchasing = 8 , pselling = 10 ,Cgoodwill =10
b = 10-8+10 = 12 , h = 8 +2+1 = 11
COMPARION OF CUP SAA AND POISSION DEMAND DERIVED POLICY TOTAL COST
REFERENCES:
• Levi, Retsef, Georgia Perakis, and Joline Uichanco. "The data-driven newsvendor
problem: new bounds and insights." Operations Research 63.6 (2015): [1294-1306].
• Natarajan, Karthik, Melvyn Sim, and Joline Uichanco. "Asymmetry and ambiguity in
newsvendor models." Management Science 64.7 (2017): [3146-3167].
• Zhang, Huanan. Data-Driven Algorithms for Stochastic Supply Chain Systems
Approximation and Online Learning. Diss. (Industrial and Operations Engineering) in
The University of Michigan (2017).
• Taha, Hamdy A. Operations research: an introduction. Pearson Education India, (2013):
[427-481], [531-548], [641-663].
REFERENCES:
• Nahmias, Steven, and Tava Lennon Olsen. Production and operations analysis. Waveland Press,
(2015). [52-90], [198-313], [449-453].
• Conover, W. J. "Practical nonparametric Statistics. New York: Wiley." 584 p (1999): [81-86].
• Scarf, Herbert. "A min-max solution of an inventory problem." Studies in the mathematical
theory of inventory and production No. P-910. RAND CORP SANTA MONICA CALIF, (1957).
• Lowalekar, H., Nilakantan, R., & Ravichandran, N. (2016). Analysis of an order-up-to-level
policy for perishables with random issuing. Journal of the Operational Research Society, 67,
483-505.
REFERENCES:
• Broekmeulen, R.A. and Van Donselaar, K.H., 2009. A heuristic to manage perishable
inventory with batch ordering, positive lead-times, and time-varying demand. Computers &
Operations Research, 36(11), pp.3013-3018.
• Pauls-Worm, K.G.J., 2016. Inventory control for a perishable product with non-stationary
demand (Doctoral dissertation, Wageningen University).