analysis of supply contracts with total minimum commitment yehuda bassok and ravi anupindi presented...
Post on 20-Dec-2015
215 views
TRANSCRIPT
Analysis of Supply Contracts with Total Minimum Commitment
Yehuda Bassok and Ravi Anupindi
presented by
Zeynep YILDIZ
Introduction
Traditional Review Models No restrictions on purchase quantities Most flexible contracts
In practice… Restrictions on buyer by commitments
In general, Common in electronics industry Family of products
expressed in terms of a minimum amount of money to purchase products of the family
Introduction
In this paper, Single-product periodic review The buyer side
Total minimum quantity to be purchased over the planning horizon
Flexibility to place any order in any period The supplier side
Price discounts applied to all units purchased (price discount scheme for commitments)
Introduction – Basic Differences
Stochastic environment, whereas most of the quantity discount and total quantity commitment literature assumes deterministic environment
Buyer makes a commitment a priori to purchase a minimum quantity
Introduction – Basic Contributions
Identifies the notion of a minimum commitment over the horizon in a stochastic environment
Identifies the structure of the optimal purchasing policy given the commitment and shows its simple and easy to calculate structure
The simplicity of the optimal policy structure enables us to evaluate and compare different contracts and to choose the best one
Model and Analysis – Assumptions
Distribution of demands is known i.i.d.r.v Deliveries are instantaneous Unsatisfied demand is backlogged Setup costs are negligible Purchasing, holding, and shortage costs
incurred by the buyer are proportional to quantities and stationary over time
Salvage value is 0 High penalty for not keeping the commitment Periods numbered backward
Model and Analysis
Actions taken by the buyer At the beginning of each period the inventory and the
remaining commitment quantity are observed Orders are placed Demand is satisfied as much as possible Excess inventory is backlogged to the next period
Optimal policy that minimize costs for the buyer characterized in terms of The order-up-to levels of the finite horizon version of the
standard newsboy problem with the discounted purchase cost
The order-up-to level of a single-period standard newsboy problem with zero purchase cost
Model and Analysis
Ct(It,Kt,Qt) = cQt + L(It + Qt) + EDt{C*
t-1(It-1,Kt-1)}
Purchasingcost
Holding & Shortage cost
Total expected costFrom period t
Through 1
Optimal cost from period t-1 through 1
Ct*(It,Kt) = minQt
Ct(It, Kt, Qt)
C0(I0,0) 0
where It-1 = It + Qt - t
Kt-1= (Kt - Qt)+
Model and Analysis
The Single Period Problem Assume K10
Min C1 (I1, K1, Q1) = cQ1 + L(I1+Q1)
s.t. Q1≥ K1
Cost function is convex in Q1
S1 is obtained from standard newsboy problem
Q1 =S1-I1 if S1-I1≥K1
K1 o.w
Model and Analysis
Proposition 1
1. The function C1*(I1,K1) is convex with
respect to I1,K1.
2. The function C2(I2,K2,Q2) is convex in I2,K2,Q2.
Model and Analysis
Q2 < K2 (K1>0) Constrained problem in last period
Q2 ≥ K2 (K1=0) Standard newsboy problem in last period
Model and Analysis
(P1)
MaxQ2 {cQ2+L(I2+Q2)+E {C1
*(I1,K1)}}
s.t. Q2<K2
Q2≥0. (P2)
MaxQ2 {cQ2+L(I2+Q2)+E {C1
*(I1,0)}}
s.t. Q2≥K2
Model and Analysis
The structure of the optimal policy Until the cumulative purchases exceed the total
commitment quantity (say this happened in period t), follow a base stock policy with order-up-to level (SM) until period t+1
After the commitment has been met, follow a base stock policy for the rest of horizon as in a standard newsboy problem with order-up-to levels of St-1,…,S1; in period t
Model and Analysis
Solution structure for unconstrained version of (P1) and (P2)
Proposition 21. The unconstrained solution of (P1) is order-up-to
SM, that is, Q2*=(SM-I2)+, where SM=F-1(/+h).
2. The unconstrained solution of (P2) is order-up-to S2 where S2 is the optimal order-up-to level of the two period standard newsboy problem.
3. S2≤ SM
Model and Analysis
Proposition 31. Assuming that I2≤SM, one and only one of the
following conditions holds:1. Problem (P1) has and unconstrained optimal solution that
is feasible. In this case, this solution is also the optimal solution of Problem (P).
2. Problem (P2) has an unconstrained optimal solution that is feasible. In this case, this solution is also the optimal solution of Problem (P)
3. Neither problem (P1) nor (P2) has an optimal unconstrained solution that is feasible. In this case the optimal solution of Problem (P) is Q2
*=K2.
2. If I2≥SM then Q2*=0.
Model and Analysis
The N-Period Problem Proposition 41. The function Ct
*(It,Kt) is convex with respect to It, Kt for t=1,…,N-1.
2. The cost function Ct(It,Kt,Qt) is convex in It, Kt, Qt for t=1,…,N.
Proposition 5At every period t, t=2,…,N, if Kt>0 then there are two critical levels SM and St such that
Computational Study
Discounted contracted price against market price with no restrictions
Percentage savings in total costs Parameters
Demand is normal and mean = 100 Coefficient of variation of demand = 0.125, 0.25, 0.5, 1.0 Percentage discount = 5%, 10%, 15% Market price per unit = $1.00 Penalty cost per unit = $25, $40, $50 Holding cost per unit = 25% of purchasing price Number of periods = 10
C programming language