View
3
Download
0
Category
Preview:
Citation preview
Research ArticleInventory Control by Multiple Service Levels underUnreliable Supplying Condition
Byungsoo Na,1 Jinpyo Lee,2 and Hyung Jun Ahn2
1Division of Business Administration, Korea University, Sejong, Republic of Korea2College of Business Administration, Hongik University, Seoul, Republic of Korea
Correspondence should be addressed to Jinpyo Lee; jinpyo.lee@hongik.ac.kr
Received 28 March 2016; Accepted 31 July 2016
Academic Editor: Francisco Gordillo
Copyright Β© 2016 Byungsoo Na et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
We consider an inventory system where there is random demand from customers as well as unreliable supplying capacity fromsupplier. In many real-world cases, supplier might fail to satisfy the amount of order from retailers or producers so that onlypartial proportion of order is satisfied or even fail to deliver all of the order. Moreover, recently a concern regarding unreliablesupplying capacity has been increasing since the globalizationmakes the retailer or producer face the extended supply network withcomplicated and risky supplying capacity. Also, we consider two classified customers, of which one is willing to pay extra charge forexpedited delivery service but the other is not reluctant to delay the delivery without any extra charge. We show that there exists anoptimal threshold for inventory and price for each service level in the following sense: if the inventory level is less than the predeter-mined threshold, then the retailer or producer needs to order up to the threshold level and offer threshold price corresponding to ser-vice level. Otherwise, the retailer does not need to order.The risk of stockout due to unreliable supplying capacity can bemitigated bythe dynamic pricing and inventory control with multiple service levels.
1. Introduction
Coordination of dynamic inventory and pricing control hasbeen used as main strategy for many companies such asAmazon, Dell, and J. C. Penny [1]. In many traditional worksfor inventory control problem, how to maximize the profitand control the uncertainty of demand from customers hasbeen a stark issue, in which the optimal policy has been pro-vided using either ordering quantity or pricing of each prod-uct. Even some works show the optimal inventory controlpolicy using multiple pricing on single product dependingon the service levels. However, in addition to the randomdemand from customers, it is not well addressed that mostfirms experience the unreliable supplying capacity fromsupplier, which is another randomness carried from supplier.For example, suppose that Amazon.com orders 100 books toa publisher. However, the publisher might fail to fulfill theorder and it may deliver only some of the ordered books(i.e., 80 books) for its own supplying capacity problem. Somerecent works address this unreliable supply issue which is
incorporated with pricing on the product and replenishmentdecisions but only addresses the pricing policy on a productwith a single-service level [2]. However, as shown in manyonline stores, each product is delivered with more than asingle-service level where faster service can be suggestedto the customer willing to pay an extra charge. So, how tobalance the demand and supply by incorporating multiplepricing corresponding to the service level is an importantissue faced by many firms, especially online firms. In thispaper, we show that there can be an optimal policy toan inventory control problem in which each product ispriced depending on the service level and the supplyingcapacity from the supplier is also random in addition to therandomness of demand from the customer.
2. Literatures Review
Two streams of literatures are related to this paper:(1) The inventory/production or pricing control problem
under reliable supplying capacity.
Hindawi Publishing CorporationMathematical Problems in EngineeringVolume 2016, Article ID 1534580, 10 pageshttp://dx.doi.org/10.1155/2016/1534580
2 Mathematical Problems in Engineering
(2) The inventory/production or pricing control problemunder unreliable supplying capacity.
There are many literatures that study the inventory orpricing control problem under reliable supplying capacity.Reference [1] gives a comprehensive survey on this problem,in which only a single-service level is considered. Refer-ences [3, 4] address a single-period problem under riskneutrality, which is addressed using Newsvendor Model.Reference [5] suggests optimal replenishment and pricingpolicy of a single-period model under risk aversion. Formultiperiod system, when total ordering cost is a linearfunction of ordering quantity, a base stock list price isshown to be optimal for single-service level in [6] and formultiple service level in [7]. References [8, 9] show that,for multiperiod system, there is an optimal inventory andpricing policy when a fixed ordering cost is considered andbacklogging is allowed. Reference [10] provides an optimalinventory and pricing policy for multiperiod system when afixed ordering cost is considered and lost-sales are allowed.Reference [11] addresses joint inventory and pricing modelwhere there is a single item with stochastic demand subjectto reference effects and the random demand is a functionof the current price and the reference price is acting as abenchmark with which customers compare the current price.Reference [12] suggests the joint pricing and inventorymodelfor a stochastic inventory system with perishable products,where the inventory system under random demand andreliable supplying capacity is modeled by a continuous-timestochastic differential equation. Reference [13] considers jointpricing and inventory replenishment decision problem overan infinite horizon where sequentially arriving customers areforward-looking to the price of a product sold by a seller. Intheir model, so-called strategic customers wait and monitorprices offered from the seller and then anticipate a lowerfuture price. Reference [14] addresses a joint pricing andproduction decision problem for perishable items sold toprice-sensitive customers, assuming that shortages are notallowed.
The other stream of literature relevant to our paper isthe inventory/production or pricing control problem underunreliable supplying capacity. Reference [15] surveys theliterature on how to quantitatively determine lot sizes whenproduction or procurement yields are uncertain. Reference[16] addresses an inventory control model for a periodicreview with unreliable production capacity, random yields,and uncertain demand but does not consider the dynamicpricing by considering the price as exogenous and thusminimizing the total discounted expected costs which areproduction, holding, and shortage costs. References [17, 18]address the joint inventory and pricing decision problemwith random demand and unreliable suppling capacity, inwhich there are no multiple service levels. Reference [19]addresses the inventory (but no pricing) decision modelfor production-inventory systems in which the stock candeteriorate with time and supplying capacity from multiplesuppliers is unreliable. Reference [20] investigates an optimalinventory strategy for a risk-averse retailer facing demanduncertainty and unreliable supply in which the price is given
and is exogenous. Reference [21] considers a risk-neutralmonopolist manufacturer ordering a key component fromseveral suppliers using a single-period model, in which somesuppliers might face risks of complete supplying disruptions,which are called unreliable suppliers. Reference [22] suggestsoptimal sourcing strategies without pricing decision whenthere are a perfectly reliable supplier and an unreliablesupplier.
3. Assumptions and Notations
In this paper, the following notations are used:
(1) π: per unit marginal cost,(2) ππ ,π‘: price charged for regular service in period π‘,
(3) ππΈ,π‘: price charged for express service in period π‘,
(4) ππ‘= ππΈ,π‘
βππ ,π‘: extra charge for the express service in
period π‘ on [0, π],(5) ππ‘: random uncertainty term of demand which had a
known distribution,(6) π·
πΈ,π‘(ππ ,π‘
, ππΈ,π‘
): demand for the express service,(7) π·π ,π‘
(ππ ,π‘
): demand for the regular service,(8) π₯π‘: inventory level at the beginning of each period π‘
before ordering,(9) π¦π‘: inventory level at the beginning of each period π‘
after ordering,(10) β
π‘(πΌ): inventory cost (holding or backlogging) at the
end of each period π‘.
Assumption 1. Backlogging is allowed.
Assumption 2. Replenishment after ordering becomes avail-able instantaneously.
Traditionally, in operations research literatures, the cus-tomerβs demand is assumed to be decreasing concave functionor simply a decreasing linear function with the price as aninput variable. Also, the expected demand value is assumed tobe finite and strictly decreasing as the price increases, which isgenerally acceptable and reasonable.This decreasing propertyof demand can be negative if the value of price is sufficientlylarge. Thus, we need to make another assumption such thata feasible price should be selected on the restricted range inorder to be nonnegative valued [6, 9, 23, 24].
Assumption 3. In each period π‘ = 1, 2, . . . , π, the demandfunction for regular service π·
π ,π‘(ππ ,π‘
, ππΈ,π‘
) = ππ ,π‘
(ππ‘), where
ππ ,π‘
(ππ‘) is given by π
π ,π‘(ππΈ,π‘
β ππ ,π‘
), and nondecreasinglinear function in π
π‘(=ππΈ,π‘
β ππ ,π‘
) β [0, π]. The demandfunction for express service π·
πΈ,π‘(ππ ,π‘
, ππΈ,π‘
) = ππΈ,π‘
(ππ ,π‘
, ππ‘),
where ππΈ,π‘
(ππ ,π‘
, ππ‘) is given by π(π
π ,π‘, ππ‘) β π·
π ,π‘(ππ ,π‘
, ππΈ,π‘
) =
π(ππ ,π‘
, ππ‘) β ππ ,π‘
(ππ‘), which is nonincreasing linear function
in ππ ,π‘
β [ππ , ππ ] and π
π‘β [0, π]. Moreover, π(π
π ,π‘, ππ‘) is the
possible maximum demand and nonincreasing and linearfunction of π
π ,π‘β [π, π
π ]. Moreover, values for π and π
π are
taken such that π(ππ , ππ‘) βπ·π ,π‘
(ππ ,π‘
, ππΈ,π‘
) is nonnegative withprobability 1.
Mathematical Problems in Engineering 3
Assumption 3 came up from the following insight. Thenumber of customers using the express service woulddecrease as the price difference π
π‘between regular and
express service increases. For example, when you try to buysome product from an online store, the higher the pricedifference between regular shipping and express shippingservice are, the more you are reluctant to select the expressshipping service. Thus, as the price difference π
π‘between
regular and express service increases, the customers who arereluctant to select express service will become the customersfor regular service.
Assumption 4. The revenue in each period π‘ = 1, 2, . . . , π isgiven by
ππ ,π‘
πΈ [π·π ,π‘
(ππ ,π‘
, ππΈ,π‘
)] + ππΈ,π‘
πΈ [π·πΈ,π‘
(ππ ,π‘
, ππΈ,π‘
)]
= ππ ,π‘
πΈ [ππ ,π‘
(ππ‘)] + π
πΈ,π‘πΈ [ππΈ,π‘
(ππ ,π‘
, ππ‘)]
(1)
and is finite and concave for ππ ,π‘
β [ππ , ππ ] and π
π‘β [0, π],
where ππ‘= ππΈ,π‘
β ππ ,π‘.
Assumption 5. π»π‘(πΌ) = πΈ[β
π‘(πΌ)] is convex in πΌ and π»
π‘(0) = 0
in each π‘ = 1, . . . , π.
Assumption 6. limπ¦βΒ±β
π»π‘(π¦ β π·
πΈ,π‘(ππ ,π‘
, ππΈ,π‘
)) =
limπ¦βΒ±β
(ππ¦+π»π‘(π¦βπ·
πΈ,π‘(ππ ,π‘
, ππΈ,π‘
))) = β for anyππ ,π‘
β [ππ ,
ππ ] and π
π‘β [0, π].
4. Mathematical Formulation
We consider an additive demand model, where the demanduncertainty for regular and express service is represented byan additive random noise π
π‘, respectively,
π·π ,π‘
(ππ ,π‘
, ππΈ,π‘
) = ππ ,π‘
(ππ‘) = D
π ,π‘(ππ‘) + ππ‘,
π·πΈ,π‘
(ππ ,π‘
, ππΈ,π‘
) = ππΈ,π‘
(ππ ,π‘
, ππ‘) = D
πΈ,π‘(ππ ,π‘
, ππ‘) + ππ‘.
(2)
We assume that ππ‘has mean zero and support [π, π]. For
any feasible choice of ππ ,π‘
β [ππ , ππ ] and π
πΈ,π‘β [ππΈ, ππΈ],
the demand is positive with probability one and the averagedemands πΈ[π·
π ,π‘(ππ ,π‘
, ππΈ,π‘
)] and πΈ[π·πΈ,π‘
(ππ ,π‘
, ππΈ,π‘
)] are finite.Given the inventory level πΌ
π‘in period π‘, the inventory level
πΌπ‘+1
in period π‘ + 1 is given as follows:
πΌπ‘+1
= πΌπ‘+ (π¦π‘β πΌπ‘) β§ π β π
πΈ,π‘(ππ ,π‘
, ππ‘) , (3)
where π β [0, π¦π‘β πΌπ‘] in period π‘ and π β§ π = min[π, π]. Let
ππ‘(πΌ, π₯) be the optimal profit function in period π‘ when the
inventory level is πΌ and the demand from period π‘ β 1 is π₯.Then, the optimality equation is given by
ππ‘ (πΌ, π₯) = max
π¦β₯πΌ,ππ β€ππ β€ππ ,πβ€πβ€π
ππ‘(πΌ; π¦, π
π , π) , (4)
whereππ‘(πΌ; π¦, π
π , π)
= ππ πΈ [ππ (π)] + π
πΈπΈ [ππΈ(ππ , π)]
β ππΈ [(πΌ + π) β§ π¦ β πΌ]
β πΈ [π» ((πΌ + π) β§ π¦ β ππΈ(ππ , π))]
+ πΌπΈ [ππ‘+1
((πΌ + π) β§ π¦ β ππΈ(ππ , π) , π
π (π))] .
(5)
From Lemma 7, we can expect a dynamic programmingmodel with single state as an input variable which is equiva-lent to (4). Moreover, we can see that the optimal solution tothe equivalent dynamic programmingmodel with single statecan be translated into the optimal solution to (4).
Lemma7. Let πΌ andπ¦ be defined as πΌβπ₯ andπ¦βπ₯, respectively.Then, (π¦β
π‘+ π₯, π
β
π ,π‘, πβ
π‘) is the set of optimal solutions to (4) if
and only if (π¦βπ‘, πβ
π ,π‘, πβ
π‘) is the set of optimal solutions to
Gπ‘(πΌ) = max
π¦β₯πΌ,ππ β€ππ β€ππ ,πβ€πβ€π
Ξ π‘(πΌ : π¦, π
π , π) , (6)
where
Ξ π‘(πΌ : π¦, π
π , π)
= ππ πΈ [ππ (π)] + π
πΈπΈ [ππΈ(ππ , π)]
β ππΈ [(πΌ + π) β§ π¦ β πΌ]
β πΈ [π» ((πΌ + π) β§ π¦ β ππΈ(ππ , π))]
+ πΌπΈ [Gπ‘+1
((πΌ + π) β§ π¦ β ππΈ(ππ , π) β π
π (π))] .
(7)
Proof. It is enough to show that, for all period π‘,
ππ‘(πΌ; π¦, π
π , π) = Ξ
π‘(πΌ : π¦, π
π , π) βπΌ, π₯, π¦, π
π , π. (8)
Since
(πΌ + π) β§ π¦ = (πΌ β π₯ + π) β§ (π¦ β π₯)
=
{
{
{
πΌ β π₯ + π, if πΌ β π₯ + π β€ π¦ β π₯
π¦ β π₯, otherwise
= (πΌ + π) β§ π¦ β π₯,
(9)
for all period π‘, the first four terms of ππ‘(πΌ; π¦, π
π , π) andΞ
π‘(πΌ :
π¦, ππ , π) are equal to each other as follows:
ππ πΈ [ππ (π)] + π
πΈπΈ [ππΈ(ππ , π)]
β ππΈ [(πΌ + π) β§ π¦ β πΌ]
β πΈ [π» ((πΌ + π) β§ π¦ β ππΈ(ππ , π))]
= ππ πΈ [ππ (π)] + π
πΈπΈ [ππΈ(ππ , π)]
β ππΈ [(πΌ + π) β§ π¦ β πΌ]
β πΈ [π» ((πΌ + π) β§ π¦ β ππΈ(ππ , π) β π₯)] .
(10)
4 Mathematical Problems in Engineering
In period π, trivially ππ(πΌ, π₯) = G
π(πΌ) = 0. In period π β 1,
ππβ1
(πΌ; π¦, ππ , π)
= ππ πΈ [ππ ,πβ1
(π)] + ππΈπΈ [ππΈ(ππ ,πβ1
, π)]
β ππΈ [(πΌ + π) β§ π¦ β πΌ]
β πΈ [π» ((πΌ + π) β§ π¦ β ππΈ,πβ1(ππ , π) β π₯)]
= ππ πΈ [ππ ,πβ1
(π)] + ππΈπΈ [ππΈ,πβ1
(ππ , π)]
β ππΈ [(πΌ + π) β§ π¦ β πΌ]
β πΈ [π» ((πΌ + π) β§ π¦ β ππΈ,πβ1
(ππ , π))]
= Ξ πβ1
(πΌ : π¦, ππ , π) .
(11)
So, ππβ1
(πΌ, π₯) = Gπβ1
(πΌ)with πΌ = πΌβπ₯. By induction, supposethat, for any πΌ, π₯ with πΌ = πΌ β π₯,
ππ (
πΌ, π₯) = Gπ (πΌ) , π = π, π β 1, . . . , π‘ + 1. (12)
Then
ππ‘(πΌ; π¦, π
π , π)
= ππ πΈ [ππ (π)] + π
πΈπΈ [ππΈ(ππ , π)]
β ππΈ [(πΌ + π) β§ π¦ β πΌ]
β πΈ [π» ((πΌ + π) β§ π¦ β ππΈ(ππ , π) β π₯)]
+ πΌπΈ [ππ‘+1
((πΌ + π) β§ π¦ β ππΈ(ππ , π) β π₯, π
π (π))]
= ππ πΈ [ππ (π)] + π
πΈπΈ [ππΈ(ππ , π)]
β ππΈ [(πΌ + π) β§ π¦ β πΌ]
β πΈ [π» ((πΌ + π) β§ π¦ β ππΈ(ππ , π))]
+ πΌπΈ [Gπ‘+1
((πΌ + π) β§ π¦ β ππΈ(ππ , π) β π
π (π))] .
(13)
The second equality holds since (πΌ+π)β§π¦βππΈ(ππ , π)βπ
π (π) =
(πΌ+π)β§π¦βππΈ(ππ , π)βπ₯βπ
π (π). Therefore, the result holds.
4.1. Optimal Inventory Control Policy. Now, we will show thefollowing optimal inventory control policy: if the inventorylevel at each period π‘ before ordering is less than a predeter-mined level, then we need to order such that the inventorylevel is increased up to the predetermined level. Otherwise,no ordering will be made.
Lemma 8. Suppose that Gπ‘+1
(β ) is concave function and, forgiven π
π and π, let π¦β(π
π , π) be the maximizer of
ππ ππ (π) + π
πΈππΈ(ππ , π) β π (π¦ β πΌ)
β π» (π¦ β ππΈ(ππ , π))
+ πΌGπ‘+1
(π¦ β ππΈ(ππ , π) β π
π (π)) ,
(14)
where ππΈ= ππ + π. Then, π¦β(π
π , π) is also the maximizer of
ππ πΈ [ππ (π)] + π
πΈπΈ [ππΈ(ππ , π)]
β ππΈ [(πΌ + π) β§ π¦ β πΌ]
β πΈ [π» ((πΌ + π) β§ π¦ β ππΈ(ππ , π))]
+ πΌπΈ [Gπ‘+1
((πΌ + π) β§ π¦ β ππΈ(ππ , π) β π
π (π))]
(15)
for any πΌ and, moreover,
ππ πΈ [ππ (π)] + π
πΈπΈ [ππΈ(ππ , π)] β ππΈ [(πΌ + π)
β§ (π¦β(ππ , π) β¨ πΌ) β πΌ] β πΈ [π» ((πΌ + π)
β§ (π¦β(ππ , π) β¨ πΌ) β π
πΈ(ππ , π))]
+ πΌπΈ [Gπ‘+1
((πΌ + π) β§ (π¦β(ππ , π) β¨ πΌ)
β ππΈ(ππ , π) β π
π (π))]
(16)
is jointly concave in (πΌ, ππ , π), where π β¨ π = max[π, π].
Proof. Givenππ andπ,π
π ππ (π)+π
πΈππΈ(ππ , π) is just constant
and thus wewill not consider them for a while. Sinceπ»(β ) andGπ‘+1
(β ) are concave,
ππ ππ (π) + π
πΈππΈ(ππ , π) β ππ¦ β π» (π¦ β π
πΈ(ππ , π))
+ πΌGπ‘+1
(π¦ β ππΈ(ππ , π) β π
π (π))
(17)
is jointly concave in (π¦, ππ , π). Now, take any πΏ > 0 such that
π¦+πΏ < π¦β(ππ , π) and we have π¦β§ (πΌ+π) β€ (π¦+πΏ)β§ (πΌ+π) <
π¦β(ππ , π). So,
β π (π¦ β§ (πΌ + π) β πΌ) β π»((πΌ + π) β§ π¦ β ππΈ(ππ , π))
+ πΌ (Gπ‘+1
((πΌ + π) β§ π¦ β ππΈ(ππ , π) β π
π (π)))
β€ βπ ((π¦ + πΏ) β§ (πΌ + π) β πΌ) β π»((πΌ + π) β§ (π¦
+ πΏ) β ππΈ(ππ , π)) + πΌ (G
π‘+1((πΌ + π) β§ (π¦ + πΏ)
β ππΈ(ππ , π) β π
π (π)))
(18)
for any π. Thus,
πΈ [βπ (π¦ β§ (πΌ + π) β πΌ) β π»((πΌ + π) β§ π¦
β ππΈ(ππ , π)) + πΌG
π‘+1((πΌ + π) β§ π¦ β π
πΈ(ππ , π)
β ππ (π))] β€ πΈ [βπ ((π¦ + πΏ) β§ (πΌ + π) β πΌ)
β π»((πΌ + π) β§ (π¦ + πΏ) β ππΈ(ππ , π))
+ πΌGπ‘+1
((πΌ + π) β§ (π¦ + πΏ) β ππΈ(ππ , π)
β ππ (π))] .
(19)
Mathematical Problems in Engineering 5
Therefore,
β ππΈ [(πΌ + π) β§ π¦ β πΌ]
β πΈ [π» ((πΌ + π) β§ π¦ β ππΈ(ππ , π))]
+ πΌπΈ [Gπ‘+1
((πΌ + π) β§ π¦ β ππΈ(ππ , π) β π
π (π))]
(20)
is increasing in π¦ when π¦ β€ π¦β(ππ , π). By the similar
argument,
β ππΈ [(πΌ + π) β§ π¦ β πΌ]
β πΈ [π» ((πΌ + π) β§ π¦ β ππΈ(ππ , π))]
+ πΌπΈ [Gπ‘+1
((πΌ + π) β§ π¦ β ππΈ(ππ , π) β π
π (π))]
(21)
is decreasing in π¦ when π¦ β₯ π¦β(ππ , π). The first result holds.
Since, for given ππ and π, π¦β(π
π , π) is the maximizer of
β π ((πΌ + π) β πΌ) β π»((πΌ + π) β§ π¦ β ππΈ(ππ , π))
+ πΌGπ‘+1
((πΌ + π) β§ π¦ β ππΈ(ππ , π) β π
π (π))
(22)
for any π¦ β R, for given ππ and π, (π¦β(π
π , π) β¨ πΌ) β§ (πΌ + π) is
the maximizer of
β π ((πΌ + π) β πΌ) β π»((πΌ + π) β§ π¦ β ππΈ(ππ , π))
+ πΌGπ‘+1
((πΌ + π) β§ π¦ β ππΈ(ππ , π) β π
π (π))
(23)
for π¦ β [πΌ, πΌ + π]. Now, take (πΌ1, ππ ,1
, π1), (πΌ2, ππ ,2
, π2), and
πΎ β [0, 1]. Then
πΎ (βπ ((πΌ1+ π) β§ (π¦
β(ππ ,1
, π1) β¨ πΌ1) β πΌ1) β π»((πΌ
1
+ π) β§ (π¦β(ππ ,1
, π1) β¨ πΌ1) β ππΈ(ππ ,1
, π1))
+ πΌ (Gπ‘+1
((πΌ1+ π) β§ (π¦
β(ππ ,1
, π1) β¨ πΌ1)
β ππΈ(ππ ,1
, π1) β ππ (ππ ,1
, π1)))) + (1 β πΎ)
β (βπ ((πΌ2+ π) β§ (π¦
β(ππ ,2
, π2) β¨ πΌ2) β πΌ2)
β π»((πΌ2+ π) β§ (π¦
β(ππ ,2
, π2) β¨ πΌ2)
β ππΈ(ππ ,2
, π2)) + πΌG
π‘+1((πΌ2+ π)
β§ (π¦β(ππ ,2
, π2) β¨ πΌ2) β ππΈ(ππ ,2
, π2)
β ππ (ππ ,2
, π2))) β€ βπ (πΎ [(πΌ
1+ π)
β§ (π¦β(ππ ,1
, π1) β¨ πΌ1)] + (1 β πΎ) [(πΌ
2+ π)
β§ (π¦β(ππ ,2
, π2) β¨ πΌ2)] β πΌ
πΎ) β π»(πΎ [(πΌ
1+ π)
β§ (π¦β(ππ ,1
, π1) β¨ πΌ1)] + (1 β πΎ) [(πΌ
2+ π)
β§ (π¦β(ππ ,2
, π2) β¨ πΌ2)] β π
πΈ(ππ ,πΎ
, ππΈ,πΎ
))
+ πΌGπ‘+1
(πΎ [(πΌ1+ π) β§ (π¦
β(ππ ,1
, π1) β¨ πΌ1)] + (1
β πΎ) [(πΌ2+ π) β§ (π¦
β(ππ ,2
, π2) β¨ πΌ2)]
β πΎ [ππΈ(ππ ,1
, π1) β ππ (ππ ,1
, π1)] β (1 β πΎ)
β [ππΈ(ππ ,2
, π2) β ππ (ππ ,2
, π2)]) = βπ (πΎ [(πΌ
1+ π)
β§ (π¦β(ππ ,1
, π1) β¨ πΌ1)] + (1 β πΎ) [(πΌ
2+ π)
β§ (π¦β(ππ ,2
, π2) β¨ πΌ2)] β πΌ
πΎ) β π»(πΎ [(πΌ
1+ π)
β§ (π¦β(ππ ,1
, π1) β¨ πΌ1)] + (1 β πΎ) [(πΌ
2+ π)
β§ (π¦β(ππ ,2
, π2) β¨ πΌ2)] β π
πΈ(ππ ,πΎ
, ππΈ,πΎ
))
+ πΌGπ‘+1
(πΎ [(πΌ1+ π) β§ (π¦
β(ππ ,1
, π1) β¨ πΌ1)] + (1
β πΎ) [(πΌ2+ π) β§ (π¦
β(ππ ,2
, π2) β¨ πΌ2)] β π
πΈ(ππ ,πΎ
,
ππΎ) β ππ (ππ ,πΎ
, ππΎ)) β€ βπ ((πΌ
πΎ+ π)
β§ (π¦β(ππ ,πΎ
, ππΎ) β¨ πΌπΎ) β πΌπΎ) β π»((πΌ
πΎ+ π)
β§ (π¦β(ππ ,πΎ
, ππΎ) β¨ πΌπΎ) β ππΈ(ππ ,πΎ
, ππΎ))
+ πΌGπ‘+1
((πΌπΎ+ π) β§ (π¦
β(ππ ,πΎ
, ππΎ) β¨ πΌπΎ)
β ππΈ(ππ ,πΎ
, ππΎ) β ππ (ππ ,πΎ
, ππΎ)) ,
(24)
where πΌπΎ
= πΎπΌ1+ (1 β πΎ)πΌ
2, ππ ,πΎ
= πΎππ ,1
+ (1 β πΎ)ππ ,2
, andππΎ
= πΎπ1+ (1 β πΎ)π
2. The first inequality holds due to the
concavity ofπ»π‘andG
π‘+1, the second equality holds due to the
linearity of demand functions, and the last inequality holdssince, for given π
π ,πΎand π
πΎ, π¦ = (πΌ
πΎ+ π) β§ π¦
β(ππ ,πΎ
, ππΎ) β¨ πΌπΎ
is the maximizer of
β π (π¦ β πΌ) β π» (π¦ β ππΈ(ππ , π))
+ πΌGπ‘+1
(π¦ β ππΈ(ππ , π) β π
π (π))
(25)
for π¦ β [πΌπΎ, πΌπΎ+ π]. So, we need to verify that
πΎ [(πΌ1+ π) β§ (π¦
β(ππ ,1
, π1) β¨ πΌ1)]
+ (1 β πΎ) [(πΌ2+ π) β§ (π¦
β(ππ ,2
, π2) β¨ πΌ2)]
(26)
is within [πΌπΎ, πΌπΎ+ π].
πΎ [(πΌ1+ π) β§ (π¦
β(ππ ,1
, π1) β¨ πΌ1)]
+ (1 β πΎ) [(πΌ2+ π) β§ (π¦
β(ππ ,2
, π2) β¨ πΌ2)]
β₯ πΎ [(πΌ1+ π) β§ πΌ
1] + (1 β πΎ) [(πΌ
2+ π) β§ πΌ
2]
= πΎπΌ1+ (1 β πΎ) πΌ
2= πΌπΎ,
6 Mathematical Problems in Engineering
Table 1: Parameters for numerical analysis.
π ππ
ππΈ
Marginal cost Holding cost Backlog cost Salvage price Range for regular price (ππ ) Range for extra charge (π)
250 5 6 30 0.3 28 20 25β65 0β30
πΎ [(πΌ1+ π) β§ (π¦
β(ππ ,1
, π1) β¨ πΌ1)]
+ (1 β πΎ) [(πΌ2+ π) β§ (π¦
β(ππ ,2
, π2) β¨ πΌ2)]
β€ πΎ (πΌ1+ π) + (1 β πΎ) (πΌ
2+ π) = πΌ
πΎ+ π.
(27)
By taking the expectation, we can obtain the second result.
Proposition 9. For all π‘ = 1, 2, . . . , π, Gπ‘(β ) is concave
function.
Proof. Since we consider the πβperiod problem, Gπ+1
(πΌ) =
0 and Gπ+1
(πΌ) is concave in πΌ. By induction, suppose thatGπ‘+1
(β ) is concave. Let π¦β(ππ , π) be defined as in Lemma 8.
By Assumption 4 and Lemma 8,
Ξ π‘(πΌ : π¦β(ππ , π) β¨ πΌ, π
π , π) = π
π πΈ [ππ (π)]
+ ππΈπΈ [ππΈ(ππ , π)] β ππΈ [(πΌ + π) β§ (π¦
β(ππ , π)
β¨ πΌ) β πΌ] β πΈ [π» ((πΌ + π) β§ (π¦β(ππ , π) β¨ πΌ)
β ππΈ(ππ , π))] + πΌπΈ [G
π‘+1((πΌ + π)
β§ (π¦β(ππ , π) β¨ πΌ) β π
πΈ(ππ , π) β π
π (π))]
(28)
is jointly concave in (πΌ, ππ , π). So,
Gπ‘(πΌ) = max
ππ β€ππ β€ππ ,πβ€πβ€π
Ξ π‘(πΌ : π¦β(ππ , π) β¨ πΌ, π
π , π) (29)
is concave in πΌ. Therefore, for all π‘ = 1, 2, . . . , π, Gπ‘(β ) is
concave.
Proposition 10. Suppose that, in period π‘, the inventory levelis πΌπ‘and the advanced demand from period π‘ β 1 which is for
regular service is π₯π‘β1
.Then, in period π‘, there is some finite pair(π¦β
π‘, πβ
π‘, πβ
π ,π‘) such that if πΌ
π‘β π₯π‘β1
β€ π¦β
π‘, it is optimal to place
positive number (π¦βπ‘βπΌπ‘+π₯π‘β1
) of order and to charge the pricesπβ
π ,π‘+πβ
π‘for the express service and πβ
π ,π‘for the regular service.
Otherwise, no ordering will take place.
Proof. As seen in Lemma 8, the optimal solution pair(π¦β
π‘, πβ
π ,π‘, πβ
π‘) does not depend on πΌ
π‘and π₯
π‘β1but π¦βπ‘depends
on (πβπ , ππ‘) which is π¦β
π‘= π¦β
π‘(πβ
π ,π‘, πβ
π‘). So, if πΌ
π‘β π₯π‘β1
β€ π¦β
π‘,
it is optimal to place π¦βπ‘β πΌπ‘+ π₯π‘β1
of orders. Otherwise, theoptimal order up to level is πΌ
π‘β π₯π‘β1
which is not to order.Now, we need to verify that there exists finite solution pair ineach period π‘. In each period π‘, the feasible sets for π
π ,π‘and π
π‘
are bounded so that the optimal pricing πβπ ,π‘
and πβπ‘are finite.
π¦β
π‘= π¦β
π‘(πβ
π ,π‘, πβ
π‘) is solution to
πβ
π ,π‘ππ (πβ
π‘) + (π
β
π ,π‘+ πβ
π‘) ππΈ(πβ
π ,π‘, πβ
π‘) β π (π¦ β πΌ)
β π»(π¦ β ππΈ(πβ
π ,π‘, πβ
π‘))
+ πΌGπ‘+1
(π¦ β ππΈ(πβ
π ,π‘, πβ
π‘) β ππ (πβ
π‘)) .
(30)
By Assumption 6, in each period π‘,
limπ¦βΒ±β
β ππ¦ β π»(π¦ β ππΈ(πβ
π ,π‘, πβ
π‘)) = ββ. (31)
Thus, as π¦ β Β±β
(πβ
π ,π‘ππ (πβ
π‘) + (π
β
π ,π‘+ πβ
π‘) ππΈ(πβ
π ,π‘, πβ
π‘) β π (π¦ β πΌ)
β π»(π¦ β ππΈ(πβ
π ,π‘, πβ
π‘))) + πΌG
π‘+1(π¦
β ππΈ(πβ
π ,π‘, πβ
π‘) β ππ (πβ
π‘))
(32)
will go to β, and thus the solution π¦βπ‘
= π¦β
π‘(πβ
π ,π‘, πβ
π‘) should
be finite. Therefore, the result holds.
5. Numerical Analysis
In this section, we provide a report on a numerical analysiscarried on to obtain insights into the structure of optimalpolicies and their sensitivity and quantitative comparisonwith the traditional policy (single-pricing policy). Among themain questions, we focus on
(1) the benefits of a multiple pricing strategy comparedto a one pricing strategy in settings with continuousinventory replenishment opportunities,
(2) the sensitivity of the optimal base stock and listprices with respect to the degree of variability and theseasonality in the demands,
(3) the comparison of profit with the traditional single-pricing policy.
Our numerical study is based on data in Table 1. As men-tioned in Assumption 3, the demand function is a linearfunction of the regular price and extra charge. That is,
ππΈ(ππ, π) = π Γ πΎ β π
π ππ β ππΈπ + π. (33)
The random term π is assumed to be normally distributedwith mean 0. However, it is truncated to avoid the negativevalue of demand such that the minimum of π is β1 Γ (π βππ ππ βππΈπ). Also, to capture the degree of demand variability,
we have used [c.v. Γ (π β ππ ππ β ππΈπ)]2 as the variance of π,
Mathematical Problems in Engineering 7
Time
1.510.5
02 3 4 51
50
100
150
200
250
300
Opt
imal
bas
e sto
ck
(a) Nonseasonal case
0
50
100
150
200
250
300
350
400
Opt
imal
bas
e sto
ck
Time2 3 4 51
1.510.5
(b) Seasonal case
Figure 1: Optimal base stock levels for varying demand uncertainty.
30
35
40
45
50
55
List
regu
lar p
rice
Time2 3 4 51
1.510.5
(a) Nonseasonal case
30
35
40
45
50
55
60
65
70
75
80
54321
List
regu
lar p
rice
Time
1.510.5
(b) Seasonal case
Figure 2: Price for regular service for varying demand uncertainty.
where c.v. is the coefficient for the variability in demand. πΎ isthe randomly generated number for the demand seasonality.
Figure 1 shows the base stock for both nonseasonal andseasonal demand cases. As you can see in this figure, as thedemand variation increases, the base stock tends to increase.This might be from the fact that more demand fluctuationcan increase the possibility of inventory shortage. So, in orderto decrease the shortage cost, the base stock would increase.Figures 2 and 3 show the threshold price predetermined for
regular service and express service in each time for eachdemand variation (c.v.).
The inventory controlling strategy by multiple servicelevels can provide the retailer or producer with the one-period advanced information regarding demand since somecustomers, who select regular service, are willing to delayshipment for their order. This one-period advanced infor-mation can be useful in managing the inventory in thenext period in the sense that inventory can be controlled
8 Mathematical Problems in Engineering
Time2 3 4 51
1.510.5
30
35
40
45
50
55Li
st ex
pres
s pric
e
(a) Nonseasonal case
30
35
40
45
50
55
60
65
70
75
80
54321
List
expr
ess p
rice
Time
1.510.5
(b) Seasonal case
Figure 3: Price for express service for varying demand uncertainty.
0
5
10
15
20
25
1.510.5(a) Nonseasonal case
0
5
10
15
20
25
1.510.5(b) Seasonal case
Figure 4: Profit comparison between inventory control by multiple service levels and traditional inventory control by single-service level.
more efficiently and moreover there is chance to reduce thecost which is induced from holding too much inventory orshortage inventory.
Figure 4 shows the benefits of an inventory controlthrough multiple service levels compared to traditionalinventory control by single-service level. In both nonseasonaland seasonal case, you can see that our model under unreli-able supplying capacity is more profitable than the traditionalsingle-service and single-pricing model. Moreover, at higherlevel of demand uncertainty (c.v.), the profit increase fromour model for seasonal demand case is higher than thenonseasonal case. Thus, we can see that the benefit frominventory control strategy by multiple service levels canbe relatively large in the environment where the inventorysystem experiences higher demand uncertainty and season-ality: inventory control by multiple service levels under theunreliable supplying capacity, which provides the system
with one-period advanced information regarding demand,efficiently captures the demanduncertainty in order to reducethe cost compared to traditional controlling model by single-service level, thus giving more profit. From this numericalanalysis, we can see that the proposed model provides betterstrategy for the inventory controlling problem even under theunreliable supplying capacity.
6. Conclusion
We study dynamic pricing and inventory replenishmentproblems under unreliable supplying condition.This researchwas initiated by the following practical intuition; that is,if customers are willing to expedite the service for theirorder, then they are willing to pay extra charge. In thispaper, we have verified this intuition through constructinga reasonable demand model (Assumption 3) and by the
Mathematical Problems in Engineering 9
mathematical dynamic programming model where productcan be provided to the customers with multiple (two) pricescorresponding to the service level, either express service orregular service. It is shown in this paper that there exist apair of threshold levels for inventory ordering up to leveland prices corresponding to service levels in the followingstrategy: if the inventory level before ordering in each periodwhich is subtracted by the regular demand advanced fromthe previous period is less than the predetermined threshold,then it is optimal to make ordering decision and increaseinventory level up to the predetermined threshold level andoffer threshold price for each service level. Otherwise, noordering is optimal.
We can extend our model in the following possibleresearch directions. First, we can extend the result to the caseof infinite horizon. Considering an infinite horizon modelwith stationary parameters with some practical assumptions,some optimality result similar to one from the finite horizonmodel might be extended. Second, we assume the depen-dency of demand in a period only on the price in the sameand current period. However, the demand function mightbe extended to depend not only on the current period butalso on the current and the historical prices, which are theprices in past periods. Even though it would be complex toanalyze, it would be interesting to incorporate these moregeneral demand settings into the model and examine thecorresponding optimal policies. Third, our assumption inwhich the backlogging is allowed can be modified such thatthe backlogging is not allowed but lost-sales are assumed.Finally, it would be interesting and important to incorporatea fixed ordering cost into our model. Analysis of suchextensions can be more complicated to analyze but would beworth further exploration.
Competing Interests
The authors declare that there are no competing interestsregarding the publication of this paper.
Acknowledgments
This research was supported by 2014 Hongik UniversityResearch Fund and also partially supported by the NationalResearch Foundation of Korea Grant funded by the KoreanGovernment (NRF-2015R1C1A1A02037096). These researchfunds do not lead to any competing interests regarding thepublication of this manuscript.
References
[1] W. Elmaghraby and P. Keskinocak, βDynamic pricing in thepresence of inventory considerations: research overview, cur-rent practices, and future directions,βManagement Science, vol.49, no. 10, pp. 1287β1309, 2003.
[2] Q. Feng, βIntegrating dynamic pricing and replenishment deci-sions under supply capacity uncertainty,βManagement Science,vol. 56, no. 12, pp. 2154β2172, 2010.
[3] S. Karlin and C. R. Carr, βPrices and optimal inventory policy,βin Studies in Applied Probability and Management Science, vol.4, pp. 159β172, 1962.
[4] N. C. Petruzzi and M. Dada, βPricing and the newsvendorproblem: a reviewwith extensions,βOperations Research, vol. 47,no. 2, pp. 183β194, 1999.
[5] V. Agrawal and S. Seshadri, βImpact of uncertainty and riskaversion on price and order quantity in the newsvendor prob-lem,β Manufacturing and Service Operations Management, vol.2, no. 4, pp. 410β423, 2000.
[6] A. Federgruen and A. Heching, βCombined pricing and inven-tory control under uncertainty,βOperations Research, vol. 47, no.3, pp. 454β475, 1999.
[7] J. Lee, βInventory control by different service levels,β AppliedMathematical Modelling, vol. 35, no. 1, pp. 497β505, 2011.
[8] X. Chen and D. Simchi-Levi, βCoordinating inventory controland pricing strategies with random demand and fixed orderingcost: the finite horizon case,βOperations Research, vol. 52, no. 6,pp. 887β896, 2004.
[9] X. Chen and D. Simchi-Levi, βCoordinating inventory controland pricing strategies with random demand and fixed orderingcost: the infinite horizon case,β Mathematics of OperationsResearch, vol. 29, no. 3, pp. 698β723, 2004.
[10] J. Lee, βDynamic pricing inventory control under fixed cost andlost sales,β Applied Mathematical Modelling, vol. 38, no. 2, pp.712β721, 2014.
[11] M. GuΜray GuΜler, T. BilgicΜ§, and R. GuΜlluΜ, βJoint pricing andinventory control for additive demand models with referenceeffects,β Annals of Operations Research, vol. 226, no. 1, pp. 255β276, 2015.
[12] S. Li, J. Zhang, and W. Tang, βJoint dynamic pricing andinventory control policy for a stochastic inventory systemwith perishable products,β International Journal of ProductionResearch, vol. 53, no. 10, pp. 2937β2950, 2015.
[13] Y. Chen and C. Shi, βJoint pricing and inventory managementwith strategic customers,β 2016.
[14] L. Feng, J. Zhang, andW. Tang, βOptimal inventory control andpricing of perishable items without shortages,β IEEE Transac-tions on Automation Science and Engineering, vol. 13, no. 2, pp.918β931, 2016.
[15] C. A. Yano and H. L. Lee, βLot sizing with random yields: areview,β Operations Research, vol. 43, no. 2, pp. 311β334, 1995.
[16] Y. Wang and Y. Gerchak, βPeriodic review production modelswith variable capacity, random yield, and uncertain demand,βManagement Science, vol. 42, no. 1, pp. 130β137, 1996.
[17] Q. Li and S. Zheng, βJoint inventory replenishment and pricingcontrol for systems with uncertain yield and demand,β Opera-tions Research, vol. 54, no. 4, pp. 696β705, 2006.
[18] X. Chao, H. Chen, and S. Zheng, βJoint replenishment andpricing decisions in inventory systems with stochasticallydependent supply capacity,β European Journal of OperationalResearch, vol. 191, no. 1, pp. 142β155, 2008.
[19] P. Ignaciuk, βDiscrete-time control of production-inventorysystems with deteriorating stock and unreliable supplies,β IEEETransactions on Systems, Man, and Cybernetics: Systems, vol. 45,no. 2, pp. 338β348, 2015.
[20] L. Shu, F. Wu, J. Ni, and L. K. Chu, βOn the risk-averseprocurement strategy under unreliable supply,β Computers &Industrial Engineering, vol. 84, pp. 113β121, 2015.
[21] B. Hu and D. Kostamis, βManaging supply disruptions whensourcing from reliable and unreliable suppliers,β Production andOperations Management, vol. 24, no. 5, pp. 808β820, 2015.
[22] K. Park and K. Lee, βDistribution-robust single-period inven-tory control problem with multiple unreliable suppliers,β ORSpectrum, 2016.
10 Mathematical Problems in Engineering
[23] F. Chen, βMarket segmentation, advanced demand information,and supply chain performance,βManufacturing & Service Oper-ations Management, vol. 3, no. 1, pp. 53β67, 2001.
[24] L. Li, βA stochastic theory of the firm,β Mathematics of Opera-tions Research, vol. 13, no. 3, pp. 447β466, 1988.
Submit your manuscripts athttp://www.hindawi.com
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
MathematicsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Mathematical Problems in Engineering
Hindawi Publishing Corporationhttp://www.hindawi.com
Differential EquationsInternational Journal of
Volume 2014
Applied MathematicsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Probability and StatisticsHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
OptimizationJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
CombinatoricsHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
International Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Operations ResearchAdvances in
Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Function Spaces
Abstract and Applied AnalysisHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
International Journal of Mathematics and Mathematical Sciences
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Algebra
Discrete Dynamics in Nature and Society
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Decision SciencesAdvances in
Discrete MathematicsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com
Volume 2014 Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Stochastic AnalysisInternational Journal of
Recommended