Home Research Teaching Education Personal
RESEARCH INTERESTS
with applications in
PUBLICATIONS
Currently Funded ProjectsTubitak 1001 Grant # 111M533: Coordinated Pricing and Inventory Management of Perishable Products Papers under ReviewKaya, O., Urek, B., Facility Location and Pricing in Closed Loop Supply Chains, submitted for publication Tuncel, Ö, Aksen, D., Kaya, O., Salman, S., An Adaptive Large Neighborhood Search Algorithm for a Selective and Periodic Inventory Routing Problem, submitted for publication Working PapersKaya, O., Ozkok, D., A Network Design Problem for Blood Banks with Inventory, Location and Routing Considerations, in preparation Kaya,O., Polat, A.L., Coordinated Pricing and Inventory Decisions for Perishable Products, in preparation Kaya, O., Ghahroodi, S.R., Dynamic Pricing and Inventory Control for Perishable Products under Uncertain and Time Dependent Demand Publications in Conference Proceedings
CONFERENCES AND INVITED PRESENTATIONS
PROFESSIONAL ACTIVITIES AND AFFILIATIONS
My research mainly focuses on developing effective
models and solution techniques in operations management. My main
research interests include revenue management in various areas,
effective management of perishable products, supply chain coordination
and contracting including
closed loop supply chains and remanufacturing operations, healthcare
management, and risk averse decision making. I am particularly
interested in identifying operational problems with high environmental
and social impact, and modeling and analyzing these problems with
appropriate operations management techniques such as dynamic
programming, game theory, integer programming, queueing theory and
stochastic modeling techniques. I also work on advancing operational
tools that can be used by other academicians or industry practitioners.
First, I present one of my current works in joint
pricing and inventory management of perishable products. I work on this
subject as a part of Tubitak 1001 project for which I am the sole PI.
Effective management of perishable products is an important issue for
many companies and management of these products is especially difficult
for the managers because of the perishability risks of these products in
a very short time. Different
than durable products, as the perishable products age, the demand for
these products start to decline and in a short time these products can
become completely obsolete. Thus, not only the amount of these products
but also their status or age effect the inventory and pricing decisions
of these products making the problem much harder to solve. There are
many studies in literature considering the inventory and pricing
decisions for perishable products, however most of these studies
consider the inventory and pricing decisions separate from each other
and coordinated inventory and pricing decisions are not studied as much.
In addition, in most of the studies in literature, single product
systems are considered and it is assumed that the products are sold
according to firstcomefirstserved order and demand is
timeindependent. However, in reality, due to customer choices, the
actual demand depends on the age or status of the products at hand. When
the products are new or fresh, the demand rate is high, however as the
products age, they will be demanded less and the customers can move to
other types of products. Since the customers prefer newer products, the
firstcomefirstserved assumption may not be valid and other types of
sale orders, such as lastcomefirstserved, needs to be considered. In
addition, in systems with multiple types of products, substitution of
customer demand across products can affect the system significantly and
lead to much different results than the single product systems. In this
project, we aim to fill the gaps in literature by considering the
coordinated inventory and pricing decisions in systems where the demand
is a function of the status of the products at hand, considering
different orders of sales. We already prepared a paper in this area and
it is currently under review for publication. We also continue working
on this subject and plan to submit two more papers in the following
year.
Kaya (2013) is
another study of mine that is in the area of revenue management. I focus
on the dynamic pricing of durable products
considering different types of customers in the market and demand
interactions over time. In this study,
the
customers are grouped into different classes depending on their purchase
probabilities and the customer classes evolve over time depending on the
demand realizations at every period. To decide on the optimal prices at
every period, we model this problem using an infinite horizon stochastic
dynamic program (SDP) and we develop several approximation algorithms to
solve this SDP since the size of the state space of the SDP makes the
optimal solution almost impossible to find. We present the efficiencies
of the heuristics and provide managerial insights through a
computational study in which we compare the revenues obtained with each
heuristic with an upper bound value that we find on the optimal revenues.
In another study that is
currently under review for publication, we consider risk considerations
and risk sensitivity of the managers and focus on the
dynamic pricing of inventories under exponential utility maximization.
In
this paper, we consider the dynamic pricing problem of inventories with
a riskaverse approach. Risk aversion manifests itself in our problem as
a utility maximization formulation. In the classical dynamic pricing
problem, the objective of the seller is purely to maximize expected
total profit obtained by selling its inventories. In our problem, the
objective of the seller is to maximize the expected utility. We use an
exponential utility function where the risk tolerance constant reflects
the risk aversion degree of the seller. For this formulation, we
establish some structural properties of the optimal solution. A
numerical study investigating the influence of risk aversion on pricing
decisions is also presented.
My second research
focus is in the area of remanufacturing and closed loop supply chains.
In a recent study, we consider a biodiesel production company that
collects waste vegetable oil from source points that generate waste in
large amounts. The company uses the collected waste as raw material for
biodiesel production. The manager of this company needs to decide which
of the present source points to include in the collection program, which
of them to visit on each day, which periodic routing schedule to repeat
over an infinite horizon and how many vehicles to operate such that the
total collection, inventory and purchasing costs are minimized while the
production requirements and operational constraints are met. For this
selective and periodic inventory routing problem, we propose two
different formulations, compare them and apply the better performing one
on a realworld problem with 36 scenarios. We generate lower bounds
using a partial linear relaxation model, and observe that the solutions
obtained through our model are within 3.28% of optimality on the
average. Several insights regarding the customer selection, routing and
purchasing decisions are acquired with sensitivity analysis. Please see
Aksen et. al. (2012) for our studies in this area. We continue working
on this area and submitted another paper that focuses on developing
effective heuristics for this problem.
We also study about the production planning in
remanufacturing systems under uncertainty. We focus on the capacity
planning, inventory and production planning decisions under uncertainty
using stochastic and robust optimization techniques.
There is growing interest over closed loop supply
chain management due to the cost and the legislation issues. In this
paper, we are interested in disassembly, remanufacturing and
refurbishing operations of closed loop supply chains. We employ the
stochastic and robust optimization techniques in this study and develop
a generic mathematical model using a mixed integer model. However, due
to the large scale of the model, the solution of the problem becomes
quite complex. In addition, there are two uncertain factors in our
model, which are the distribution of demand and returns. We use a twostage
model; in the first stage we decide on the tactical decisions such as
the capacities of the disassembly and refurbishing sites and in the
second stage we decide on the operational decisions such as the
production quantities. Considering the randomness in the system, we use
robust optimization and stochastic optimization approaches to decide on
the capacity and production decisions and we analyze the expected costs
and the deviations from the optimum values. We analyze the mean and
variance of the costs under different scenarios.
Please see Kaya et. al. (2013a) for our studies
about this subject.
In another study related to remanufacturing, we
consider a manufacturer producing original products using virgin
materials and remanufactured products using returns from the market
where the amount of returns depend on the incentive offered by the
manufacturer. We determine the optimal value of this incentive and the
optimal production quantities in a stochastic demand setting with
partial substitution. We analyze 3 different models in centralized and
decentralized settings where the collection process of the returns is
managed by a collection agency in the decentralized setting. We also
analyze contracts to coordinate the decentralized systems and determine
the optimal contract parameters. Finally, we present our computational
study to observe the effect of different parameters on the system
performance. Please see Kaya (2010) about my paper on this subject.
I also have another paper in the area of closed loop supply chain
management and it is currently under review for publication. We address
the inventorylocation and pricingproblem in a closedloop supply chain
in which the collection of the used products and the distribution of the
newly produced ones are done simultaneously. We focus on the price and
incentive determinations for the sale and collection activities as well
as the determination of the best facility locations to open. We develop
a mixed integer nonlinear facility location allocation model with
inventory considerations to decide both optimal location of these
collection and distribution centers and the optimal price and incentive
values for the sale of new and collection of used products, respectively,
in order to maximize the total profit. To solve this NPhard problem, we
propose three different hybridized heuristic approaches which correspond
to applications of simulated annealing, tabu search and genetic
algorithms, all of which are hybridized with variable neighborhood
search (VNS) algorithm. Different neighborhood structures are embedded
in these heuristics to obtain better results.
I also worked on
the classical supply chain management area in the previous years. In the
papers Kaya et. al. (2013b and 2013c), we focus on the
coordinated production
and shipment problem in a supply chain considering both deterministic
and stochastic demand cases.
In Kaya et.
al. (2013b),
we consider a model in which the customers are willing to wait at the
expense of a waiting cost. Accordingly, the retailer does not hold
inventory but accumulates the customer orders and satisfies them at a
later time. The supplier produces the items, holds the inventory and
ships the products to the retailer to satisfy the external demand. We
investigate both a coordinated production/transportation model and a
decentralized model. In the decentralized model, the retailer manages
his own system and sends orders to the supplier, while the supplier
determines her own production process and the amount to produce in an
inventory replenishment cycle according to the order quantity of the
retailer. However, in the coordinated model, the supplier makes all the
decisions, so that she determines the length of the replenishment and
transportation cycles as well as the shipment quantities to the retailer.
We determine the structure of the optimal replenishment and
transportation cycles in both coordinated and decentralized models and
the corresponding costs. Our computational results compare the optimal
costs under the coordinated and decentralized models. We also
numerically investigate the effects of several parameters on the optimal
solutions.
In a related study to the above paper, in
Kaya et. al.
(2013c),
we consider a similar model in a stochastic environment. In this
stochastic model, the supplier controls the production, holds inventory
and ships the products to the retailer to satisfy the external demand.
We model this system as a Markov decision process and show that the
optimal production and transportation decisions are complex and nonmonotonic.
Therefore, we analyze twowidelyused shipment policies in the industry as well, namely timebased
and quantitybased policies. Using the optimal shipment frequencies and
shipment quantities according to these two policies, we numerically
compare the performances of thesepolicies with respect to the optimal
policy. We show that the quantitybased policy is superior to the timebased
policy and gives very close results to the optimal solution.
In addition, I study on supply chain coordination and contracting
in Kaya (2011) in order to increase the efficiency of the system through
various contracts.
I analyze the effort and pricing decisions in a two facility supply chain
in which one of the parties can exert costly effort to increase demand.
I consider an outsourcing model in which the supplier makes the effort
decision and an inhouse production model in which the manufacturer
decides on the effort level and we compare these two models with each
other. I analyze and compare several contracts for decentralized supply
chain models and I aim to find which contracts are best to use in
different cases. I find the optimal contract parameters in each case and
perform extensive computational testing to compare the efficiencies of
these contracts. I also analyze the effect of the powers of the parties
in the system and the effect of system parameters on the performances of
the contracts and on the optimal values of the model variables such as
price, effort and demand.
As a followup to this study, we consider a similar model in our second
paper, however, in this second paper, both the manufacturer and the
supplier do not have to follow a maketoorder policy and can stock some
level of inventory in order to increase system performance. In this
second paper, we find the optimal levels of inventory that needs to be
stocked at both facilities in addition to developing an algorithm for
scheduling and duedate quotation in this system. In addition to our
analytical results for the optimal inventory levels, we perform
extensive computational testing of our heuristics to show their closetooptimal
performances.
Finally, in our third paper, we consider a more general model using a
supply chain network composed of several centrally managed production
facilities as well as external suppliers. We design effective heuristics
for inventory positioning, order sequencing, and short and reliable duedate
quotation for this supply chain. We perform extensive computational
testing to assess the effectiveness of our algorithms, and we explore
the impact of supply chain topology on inventory costs and effective duedate
quotation.
Please see the papers Kaminsky and Kaya (2008a,
2008b and 2009) about our studies in this area.

