Transactions on Operational Research

DATABASE

TOR HOME Article Index Author Index

ARTICLE ABSTRACTS

Authors	Ao Li, Levent Tunçel
Title	Some Applications of Symmetric Cone Programming in Financial Mathematics
Issue	NOVEMBER 2006 VOLUME: 17 NUMBER: l, 2
Abstract
We review a few of the relavitely recent developments in cone programming that seem to have important applications in financial planning. In particular, we go over the semidefinite programming representation of a polynomial inequality as given by Nesterov. We mention some relevant references which show the power of cone programming in portfolio optimization. Then we turn to the recent work of Lobo et al. which showed how to use Second Order Cone Programs to model portfolio optimization problems with transaction costs. We extend their model to a multi-period decision making situation and we allow cash infusions into the portfolio every period. We conclude with some computational experiments using real data.
Keywords	Convex optimization, semidefinite programming, second order cone, financial mathematics, portfolio optimization

Authors	Birnur Özbaş, İlhan Or
Title	Simulation of Maritime Transit Traffic in the İstanbul Channel
Issue	NOVEMBER 2006 VOLUME: 17 NUMBER: l, 2
Abstract
This study involves a functional simulation model for the maritime transit traffic in the İstanbul Channel that investigates the Channel Traffic Rules and Regulations, vessel types, cargo characteristics, meteorological and geographical conditions, pilotage and tugboat services. The simulation model mainly focuses on the transit traffic in the Channel. It assumes two primary transit lanes and one additional lane for overtaking. The entities of the model, which are comprised of five types of vessel, and their attributes are generated randomly according to the historical data. The arrived vessels check if the Channel is available and safe for the passage against the Rules and Regulations. This study provides a platform to analyze the effects of factors such as Maritime Traffic Rules, number of tugboats and pilots, vessel type, traffic density and meteorological conditions on the maritime traffic in the Channels. Results obtained and reported in the study clearly indicate some interesting relationships among these factors. As the effect of six factors on 8 responses are examined, the most significant factor is determined as number of pilots and tugboats in the service, while the second effective factor is the arrival rate of vessels.
Keywords	Simulation, İstanbul Channel, maritime transportation

Authors	Ömer Akgöbek
Title	Veri Madenciliğinde Otomatik Kural Üretebilen Bir Uzman Sınıflandırma Sisteminin Geliştirilmesi
Issue	NOVEMBER 2006 VOLUME: 17 NUMBER: l, 2
Abstract
The biggest problem for an expert system is knowledge acquisition and its evaluation. Acquiring knowledge from an expert is a time-consuming work requiring many meetings, attention to details and methodical studies. Although experts use their expert knowledge easily in daily work, they can not perform the same success in summarizing their knowledge and making them available for an expert system. The analysis of this knowledge and the creation of knowledge base from this knowledge for an expert system require another expertise. In this work, we have developed a system which can produce the rules according to the algorithms of REX (REX-1, REX-2, REX-3) family and generate a result based on the values produced by using these rules. With the help of this system, users can create their own database, enter their knowledge and produce rules from this knowledge without requiring another expert.
Keywords	Data mining, expert system, knowledge acquisition, rule induction

Authors	Beate Brocmüller, Oktay Günlük, Laurence A. Wolsey
Title	Designing Private Line Networks
Issue	NOVEMBER 2004 VOLUME: 16 NUMBER: l, 2
Abstract
We study a capacitated network design problem arising in the design of private line networks. Given a complete graph, a subset of its node set (the “hub" set), and point-to-point traffic demands, the objective is to install capacity on the edges (using several batch sizes and nonlinear costs), and route traffic in the resulting capacitated network, so that 1) all the demand between a pair of nodes is routed along a single path, and 2) the demand is either sent directly from source to sink, or via a number of hub nodes. We first formulate an initial integer program, and various approximations to it. Valid inequalities are then derived for a special knapsack problem involving both integer and 0-1 variables arising from the capacity constraints on an edge. These and related inequalities are then used to strengthen the formulations. Computational results using these inequalities within a general purpose branch-and-cut system are presented.
Keywords

Authors	Payam Hanefizadeh, Abbas Seifi
Title	A Unified Model for robust Optimization of Linear Programs with Uncertain Parameters
Issue	NOVEMBER 2004 VOLUME: 16 NUMBER: l, 2
Abstract
This paper presents a general model for unification of the robust counterparts of uncertain linear programs (LP). We extend the robust optimization methodology for LP, introduced by Ben-Tal and Nemirovski, to a unified model in which the uncertainty region is approximated by an appropriate norm body. We derive the robust counterpart of an LP whose parameters may fall in any of the or matrix-norm bodies. An application to portfolio selection shows how an investor’s risk can be measured by a proper norm. Then, various solutions are shown for a practical example. The numerical results indicate that choosing norm leads to higher utility for an investor when the uncertainty region is large.
Keywords	Robust optimization, linear programming, data uncertainty, portfolio selection.

Authors	Hugo J. Lara, Clovis J. Gonzaga, Levent Tunçel
Title	On the Limiting Properties of the Affine-Scaling Directions
Issue	NOVEMBER 2004 VOLUME: 16 NUMBER: l, 2
Abstract
We study the limiting properties of the affine-scaling directions for linear programming problems. The worst-case angle between the affine scaling directions and the objective function vector provides an interesting measure that has been very helpful in convergence analyses and in understanding the behaviour of various interior-point algorithms. We establish new relations between this measure and some other complexity measures which are used in the complexity analyses of algorithms for linear programming. We also provide a new characterization of the smallest large variable complexity measure of Ye.
Keywords	Affine-Scaling Direction , Linear Programming, Complexity Measures, Interior-Point Methods

Authors	Alifetah Shahbazov, Vedat Sağlam
Title	Asymptotic Analysis and Optimization of the Reliability of Two Unit Systems With Quick
Issue	NOVEMBER 2004 VOLUME: 15 NUMBER: l, 2
Abstract
A redundant system consisting of two dissimilar units and a single repair facility is considered. If the probability that the repair time is larger than the failure time tends to zero, then it is proved that the limiting distribution of the normalised system lifetime is exponential. The maximisation of the mean lifetime of the system with exponential repair times under the restriction that total service rate is fixed is investigated. Under the assumption that mean failure time of units is given it is proved that the mean of the system lifetime is maximised when the failure time of units is constant with probability one. Optimal starting order of units which maximises the mean of the system lifetime is analysed for special cases.
Keywords	Failure time, Repair time, Working unit, Standby unit, System lifetime, Markov renewal process, Laplace-Stieltjes transform, Stochastic ordering.

Authors	Fulya Altıparmak, Berna Dengiz
Title	Yapay Sinir Ağları ile İletişim Ağlarının Güvenilirlik Tahmini
Issue	NOVEMBER 2004 VOLUME: 15 NUMBER: l, 2
Abstract
Bir bilgisayar iletişim ağının tasarımında dikkate alınması gereken en önemli kriterlerden birisi iletişim ağının güvenilirliğidir. Güvenilirlik, bu tür tasarım problemlerinde amaç fonksiyonu veya kısıt olarak dikkate alınır. Bir iletişim ağının çalışıyor olmasını tanımlamak için deterministik veya olasılıklı güvenilirlik kriterleri kullanılır. Bu çalışmada, olasılıklı kriterler sınıfında yer alan tüm-terminal güvenilirliği dikkate alınmıştır. Tüm-terminal güvenilirliği, ağdaki her düğümün diğer düğümler ile iletişim kurabilme olasılığıdır. Ancak, güvenilirliğin tam değerinin hesaplanması NP-zor problemdir, yani hesaplama zamanı ağın boyutuna bağlı olarak üstel artış gösterir. Bu nedenle, bu çalışmada ağ güvenilirliğinin tahmini için yapay sinir ağları (YSA) önerilmekte ve YSA’nın eğitiminde kullanılmak üzere rassal ve faktöriyel olmak üzere iki farklı deney tasarımı ile bir ağ içinde hat güvenilirliklerinin aynı ve farklı olduğu durumlar dikkate alınarak üretilen veri kümelerinin YSA’nın performansı üzerindeki etkileri incelenmektedir.
Keywords	Ağ Güvenilirliği, Ağ Tasarımı, Yapay Sinir Ağları, Deney Tasarımı

Authors	Hakan Ümit, Bahar Y. Kara
Title	Büyük Ölçekli Bir Tehlikeli Madde Taşımacılığının Coğrafi Bilgi Sistemleri Tabanlı Analizi: Shell Türkiye Örneği
Issue	NOVEMBER 2004 VOLUME: 15 NUMBER: l, 2
Abstract
Tehlikeli madde taşımacılığı yıllardır resmi otoritelerin, çevreci grupların, bilim çevrelerinin ve son yıllarda da bazı taşıyıcı firmaların diönemli ölçüde dikkatini çeken bir konu olmuştur. Tehlikeli madde taşımacılığı literatüründe, karayolları üzerindeki tehlikeli madde taşımacılığı ile ilgili iki önemli karar verici grubun iki ayrı problemi yer almaktadır. Birinci grup yani, resmi otoritelerin problemi karayolları üzerinde tehlikeli madde taşıyanların çevreye karşı yarattığı ‘risk’i enazlamaktır. İkinci grup yani, taşıyıcı firmaların problemi ise hem taşımacılıktan kaynaklanan ‘risk’i, hem de kendi taşıma maliyetlerini enazlayan güzergahları seçmektir. Çoğunlukla birbiriyle çelişen bu iki amacın (taşıma ‘risk’ ve maliyetleri), her iki grubun önceliklerini mümkün olduğunca tatmin edebilmesi gerekmektedir. Bu çalışmada, tehlikeli madde üreten bir firmanın tehlikeli madde taşımacılığı problemine çözümler önerilmektedir.
Keywords	Tehlikeli Madde Taşımacılığı, Coğrafi Bilgi Sistemleri

Authors	Çiğdem Alabaş, Berna Dengiz
Title	Tabu Arama Algoritması ile Yapay Sinir Ağlarının Eğitimi
Issue	NOVEMBER 2002 VOLUME: 13, 14 NUMBER: l, 2
Abstract
Yapay sinir ağlarının (YSA) ilham kaynağı biyolojik beynin gücü, esnekliği ve duyarlılığıdır. YSA, beynin temel biyolojik bileşenlerinden sinir hücreleri, synapsis ve dendritelerin matematiksel modeli olup, basit matematiksel elemanlardan oluşmaktadır. YSA'da öğrenme işlemi gözetimli (supervised) ve gözetimsiz (unsupervised) olmak üzere iki şekilde gerçekleşir. Gözetimli öğrenme işleminde, her girdi seti için bir çıktı seti gereklidir ve her ikisi birlikte öğrenme setini oluşturur. Genellikle belirli sayıdaki bu öğrenme çiftlerinin YSA'ya tanıtılması ile öğrenme gerçekleştirilir. Öğrenme sürecinde; YSA'ya bir girdi seti verilir ve çıktısı hesaplanır. Hesaplanan bu çıktı ile mevcut çıktı seti değeri arasındaki sapmayı enküçükleyen bir algoritma uyarınca istenilen sapma düzeyine ulaşılana dek YSA ağırlıkları değiştirilir. Böylece YSA eğitilmiş, dolayısıyla ağırlıklar en iyi değerlerini almış olur. Gözetimli öğrenmede genellikle Geri Yayılma (Back Propagation) algoritması kullanılır. Geri Yayılma algoritması dışbükey olmayan hata fonksiyonunu enküçükleyen ağırlık değerlerini hesaplamak için kullanıldığından yakınsama hızı yavaştır ve yerel eniyi çözümlerde durabilmektedir. Bu nedenle bu çalışmada, YSA'nın ağırlıklarının belirlenmesinde genel eniyilemeyi sağlamak amacıyla bir tabu arama (TA) algoritması sunulmaktadır. Rassal hareket mekanizması ve yayılma stratejileri kullanılarak geliştirilen TA algoritması iki farklı sınıflandırma problemine uygulanarak hem yakınsama hızı, hem de yerel eniyilerden kurtulma ölçütlerine göre değerlendirilmiştir.
Keywords	Yapay sinir ağları, Tabu arama, Gözetimli öğrenme

Authors	U. G. Nuriyev, Pınar Dündar
Title	Çok Boyutlu Yükleme Probleminin Çözümü İçin Kendi Kendini Düzenleyen Bir Algoritma
Issue	NOVEMBER 2002 VOLUME: 13, 14 NUMBER: l, 2
Abstract
Bu çalışmada Boole değişkenli çok boyutlu Yükleme(Knapsack) probleminin çözümü için heuristik bir algoritma tasarlanmıştır. Önerilen sert kısıtlama yöntemi, geçerli bilgiye göre çözüm arama sürecinde değişebilen bir esnek tarama şeması tanımlama olanağı sağlamaktadır. Ayrıca, verilen problemin özelliklerine uyan bir taramayı da mümkün kılmaktadır. Yapılan hesaplama denemeleri önerilen algoritmanın verimliliğinin yüksek olduğunu göstermektedir. MS Windows ortamında Delphi 4.0 kullanılarak önerilen algoritmaları içeren KnapsacK paket programı hazırlanmıştır. Makalede KnapsacK paket programı ile yapılan performans testlerinin sonuçları da verilmiştir.
Keywords	Boole değişkenli çok boyutlu Yükleme problemi, heuristic algoritma, sertkısıtlama yöntemi, kendi kendini düzenleyen algoritma, paket program

Authors	Ching-Wen Chen, Paul H. Randolph
Title	Emprical Bayes Analysis for Monitoring A Production Process with Geometric Poisson Properties in the Quality Characteristics
Issue	NOVEMBER 2002 VOLUME: 13, 14 NUMBER: l, 2
Abstract
For count data of the number of defects per unit, the common assumption is that the Poisson distribution holds. However, the Poisson is not the only underlying distribution for the count of defects. In some situations, the distribution of defects are more appropriately modeled by a compound Poisson distribution. In this paper the underlying distribution is assumed to be the geometric Poisson distribution. A Shewhart-like charting technique incorporated with the empirical Bayes analysis is developed to monitor a production process in which the distribution of defects can be modeled by the geometric Poisson distribution. The empirical Bayes approach utilizes all relevant data since this gains knowledge about the process parameters over time. Using simulated data, the performance of the compound Poisson control charts and classical Poisson-based control chart are compared. The test results show an underestimation of Type I error and thus the number of false alarms generated, if the underlying defect distribution is incorrectly assumed
Keywords

Authors	Emin Kahya
Title	Uzun Dönem Kapasite Artış Planlaması Probleminin Fayda Teorisi Yaklaşımıyla Modellenmesi ve Çözümü
Issue	NOVEMBER 2000 VOLUME: 12 NUMBER: l, 2
Abstract
Üretim ve hizmet sistemlerinde, talebin zamanla artması halinde, kapasitenin ne zaman ve hangi miktarda arttırılacağı problemleri, uzun dönem kapasite artış planlaması problemi olarak bilinir. Talebin belirli ve zamanla dogrusal olarak arttığı ve planlama uzayının sonsuz olduğu uzun dönem kapasite artış planlaması problemlerinde, genellikle, planlama uzayı boyunca, kapasite artışlarına ilişkin tüm maliyetlerin peşin değeri toplamının en küçüklenmesi amaçlanmaktadır. Bu çalışmada, fayda teorisinin temel özellikleri kullanılarak, toplam maliyetin en küçüklenmesi yerine, toplam maliyetin karar vericiye olan faydasının en büyüklenmesini amaçlayan bir kapasite artış modeli sunulmuştur. Maliyetin karar vericiye olan faydasının üstel fayda fonksiyonu ile belirlenebilmesi halinde, kapasite artışı fayda fonksiyonu gelişebilmiş, modelin analitik çözümü ve amaç fonksiyonunun konkavlığı araştırılmıştır. Model, gerek diğer fayda fonksiyonu tipleri ve gerekse maliyet ve talep parametreleri için irdelenmiştir. Önerilen yaklaşım, bir hastahanenin ortopedi servisinin kapasite artış planlaması için uygulanmış ve en iyi çözüm GINO paket programıyla bulunmuştur. Ayrıca, modelde yer alan en önemli parametreler için duyarlılık analizleri yapılarak dikkat çekici sonuçlar yorumlanmıştır.
Keywords	Kapasite artış planlaması, Fayda teorisi, Üstel fayda fonksiyonu, Sayısal çözüm, GINO Paket Programı

Authors	İ. Kuban Altınel, Necati Aras
Title	Yapay Sinir Ağları ve Uzaklıkların Kestirimi
Issue	NOVEMBER 2000 VOLUME: 12 NUMBER: l, 2
Abstract
İki yerleşim merkezi arasındaki gerçek uzaklık merkezleri bağlayan en kısa yolun uzunluğu olarak tanımlanır. Bir bölgedeki herhangi iki nokta arasındaki gerçek uzaklığı ölçmek ve saklamak her zaman olanaklı olmayabilir. Bu durumda gerçek uzaklığın kestirilmesi söz konusudur. Uzaklık kestirimi için yagın yaklaşım aralarındaki uzaklığın hesaplanacağı noktalara ait elde edilmesi kolay bilginin parametrik bir fonksiyonu olan kestiriciler kullanmaktır. Daha değişik bir yaklaşım ise uzaklıkları yapay sinir ağları yardımıyla kestirmektir. Bunlardan çok katmanlı algaçlar kullanan yöntemler, parametresiz olmaları nedeniyle önsel herhangi bir model varsaymazlar ve bir öğrenme kümesi üzerindeveriyi doğrudan kullanırlar. Her ne kadar algaç kullanan parametresiz kestiriciler parametreli uzaklık fonksiyonlarına göre daha başarılı olsalar da noktaların herhangi bir kestirim girişiminden önce vektör nicemleme ile kümelenmesi sonuçları daha da iyileştirmektedir. Bu çalışmada uzaklık kestirimleri için yapay sinir ağlarını kullanan yöntemler açıklanmakta ve Türkiye’den örneklenen yerleşim merkezleri üzerinde elde edilen sonuçlar vektör nicemleme ile yapılan kümelendirme sonrası uygulanan parametrik yöntemlerin verdiği sonuçla kıyaslanmaktadır.
Keywords	Sinir ağları, Taşımacılık, Yapay Zeka, Yer Seçimi, Vektör Nicemleme

Authors	Fulya Altıparmak, Berna Dengiz, Alice E. Smith
Title	An Evolutionary Approach for Reliability Optimized In Fixed Topology Computer Networks
Issue	NOVEMBER 2000 VOLUME: 12 NUMBER: l, 2
Abstract
This paper presents a meta-heuristic approech using a genetic algorithm (GA) to optimize the overall  reliability of computer communication networks with fixed topology subject to a budget constraint. İn most network design problem, such as those faced in hospitals, universities, research centers and commercial business, network topology is fixed because of geographical and physical  contraints and presence of an existing system. When the topology is known, the problem of choosing types of links and computer systems among alternatives, having different reliability and cost, to optimize system reliability is a NP-hard combinatorial problem. The heuristic is shown to be effective and computationally efficient compared to hillclimbing and random search on a suite of diverse test problems.
Keywords	Genetic Algorithms, Hillclimbing, Random Search, Network Reliability, Network Design

Authors	Kemal Altınkemer
Title	Operations Research and Economics in Applications of Management Information Systems
Issue	NOVEMBER 1999 VOLUME: 11 NUMBER: l, 2
Abstract
The science of collecting, storing, analyzing, mining, processing and disseminating information for useful and effective decision making is called Management Information Systems (MIS). MIS needs databases, datamarts and data warehouses to store data. Consequently it needs data mining to analyze and observe unexpected relationships between data attributes. It also has to assist business intelligence to provide the required information to the right decision maker at the right time. This is accomplished by the infrastructure for disseminating. It includes how to design and manage computer networks which encompasses INTERNET, INTERNET2, Next Generation Internet. In decision making sometimes you need new techniques such as outsourcing and business process reengineering. Virtual reality, visualization, war gaming, synthetic economic environment for scenario analysis, and Decision Support Systems (DSS), neural networks, Artificial Intelligence (AI) to make decisions help agents have the information at the right time for the right person. OR has extensively been used to design tools such as databases, computer networks. In the uncertain environment we live in, Economics and OR together help to solve problems, such as allocating scarce resources i.e. bandwidth and pricing the information goods over the Internet.
Keywords	MIS, Internet, Communication Infrastructure, Network Design, Data Mining, Outsourcing, Business Process Reengineering, OR, Economics

Authors	İlhan Or, Köksal Atik
Title	Comparison of Quadratic Assignment Problem Solution Methods for the Automated Assembly of Printed Circuit Boards
Issue	NOVEMBER 1999 VOLUME: 11 NUMBER: l, 2
Abstract
In this study, "the feeder carriage configuraion problem in electronic component placement machines", which arises in automated assembly of Printed Circuit Borards, is considered. This problem can be formulated as a Quadratic Assignment Problem (QAP), for which a fast and good solution procedure is quite desirable. Furthermore, under some placement machine architectures, the feeder configuration and the placement sequencing problems need to be addressed together, in which case the repeated solution of a QAP is needed in iterative manner, so a fast and good solution procedure becomes very important. The QAP is an NP-hard problem frequently arising in many manufacturing and service industries. There are many heuristic solution procedures with comparable, good performance records on randomly generated data. However, the performance of most heuristic procedures and the quality of the near optimal solutions they generate depend on problem parameters, while in many implementations the problem data carries a certain structure. In this study, some characteristics of the problem parameters is "the PCB asembly feeder configuration environment" (flow dominance, sparsity, etc.) are identified. The performance of various heuristic solution procedures are compared against each other by using test problems and having such characteristics.
Keywords	Quadratic Assignment Problem, Taboo Search, Simulated Annealing, Genetic Algorithms, Operations Sequencing

Authors	Fikri Karaesmen, Surendra M. Gupta
Title	Service Control In A Finite Buffered Queue With Holding and Setup Costs
Issue	NOVEMBER 1998 VOLUME: 9, 10 NUMBER: l, 2
Abstract
We consider a finite buffered queue where the queue length is controlled by shutting down and restarting the server. In particular, we analyze the problem with the inclusion of holding costs for customers (or items) whereas previous research concentrated on the case without holding costs. To study the effect of holding costs, we first establish some stochastic comparisons that permit us to compare different operating policies. These comparisons lead to sufficient conditions that can simplify the structure of the policy. We also present new results on the queue length distribution for the system. Our methods enable us to obtain the queue length distribution in closed form for phase type service distributions. As a consequence, we provide extensive numerical examples over a range of the problem parameters which uncover some intriguing properties ofthe optimal revenue depending on the service time distribution.
Keywords	Queueing, Stochastic Processes, Control

Authors	Berna Dengiz, Fulya Altıparmak, Akif A. Bulgak
Title	Optimization of Stochastic Systems Using GeneticAlgorithms
Issue	NOVEMBER 1998 VOLUME: 9, 10 NUMBER: l, 2
Abstract
Genetic Algorithms (GAs) are probabilistic search algorithms based on the mechanics of natural selection in genetics. In this study, we propose the use of GAs in simulation optimization in an attempt to extend their domain of applicability to the optimization of stochastic systems such as inventory control system and asynchronous automatic assembly system. Two preliminary application of GAs in simulation optimization are presented. The first application is the optimization of reorder point (s) and target level (S) for a periodic review inventory control system. The second example is the optimization of buffer size of an asynchronous automatic assembly system (AAS). Preliminary results indicate that GAs appear to be promising in simulation optimization.
Keywords	Genetic Algorithms, Inventory Control System, Asynchronous Automatic Assembly System

Authors	Achille Messac, Surendra M. Gupta, Burak Akbulut
Title	Linear Physical Programming: A New Approach toMultiple Objective Optimization
Issue	NOVEMBER 1996 VOLUME: 8 NUMBER: 2
Abstract
Optimization problems can be partitioned into two categories hereby called blind optimization and physical optimization. In blind optimization the analyst has no knowledge of the physical meaning of the problem at hand, or of the nature of its anticipated solution. In physical optimization the decision maker has substantive knowledge and often clear objectives, regarding aspects of the problem at hand, that can be articulated in physically meaningful terms (e.g. a Return On Investment of 10% is desirable; 8%, tolerable; and 5%, unacceptable). Nearly all operational research or engineering design problems belong to the latter category. This paper explores a new optimization philosophy, linear physical programming, for operational research applications by addressing the distinct issues related to multiple objective optimization: (i) the problem formulation, (ii) the nature of the obtainable solutions, and (iii) thealgorithm.
Keywords	Physical programming, multi-objective optimization, decision support, goal programming, fuzzy optimization, Multiple Criteria Decision Making(MCDM).

Authors	Tülin Yazgaç, Linet Özdamar
Title	Sipariş Üzerine Üretimde Ürün Ailelerine Bağlı BirÜretim Planlama Yaklaşımı
Issue	NOVEMBER 1996 VOLUME: 8 NUMBER: 2
Abstract
Sipariş üzerine üretim yapan firmalarda, müşteriye teslim tarihi verme olgusunun zorluğu pratikte bilinmektedir. Literatürde, bu konuyla ilgili olarak, önsüre yönetimini sağlayacak olan girdi/çıktı kontrolü gibi yöntemlerle karşılaşılmaktadır. Bu çalışmada ise, sipariş, üzerine üretim sistemleri, ürün sınıflandırılması yapılarak ürün ailesi bazında ele alınmakta ve haftalik üretim planları, kapasite sınırlarının da dikkate alındığı bir doğrusal programlama modeli yardımıyla elde edilmektedir. Geliştirilen modelin ürün ailesi bazında olması, kullanılan değişken ve kısıt sayılarını minimuma indirgemektedir.Teslim tarihleri ise modelin çözümü sonucunda doğal olarak ortaya çıkmaktadır. Ancak, planlamacı, verilebilecek maksimum teslim sürelerini saptama esnekliğine de sahiptir. Modelin çözümü, fazla mesai ve ardısmarlama maliyetlerinin toplamı açısından eniyilenmiş bir Ana Uretim Programi oluşturmaktadır. Ürün ailelerinin geçişleri arasındaki hazirlık süreleri Ana Üretim Programına dahil edildiğinde bazı kapasite aşımlarına yol açmaktadır. Bunun engellenmesi amacıyla, normal kapasite sınırlarının revize edildiği iteratif bir algoritma önerilmiştir. Bu algoritmanın amacı, haftalararası yükü dağıtarak hazirlık sürelerine yer açmaktir.Önerilen model ile iteratif algoritma, bir mobilya fabrikasında üretimin planlanması amacıyla uygulanmış, ve elde edilen sonuçlar gerçekleşen üretim ile karşılaştırılmıştır. Bu karşılaştırma sonucunda, toplam maliyetin önemli ölçüde azaldığı ve haftalararası yükün olabildiğince dengelendiği görülmektedir.
Keywords

Authors	M. Selim Aktürk, İhsan Durusoy
Title	Joint Economic Lot-Sizing Approach to the Just-In-Time Purchasing Problem
Issue	NOVEMBER 1996 VOLUME: 8 NUMBER: 2
Abstract
In a just-in-time (JIT) environment, purchaser demands frequent deliveries of small lots from the suppliers. However, this issue is settled between the purchaser and supplier depending on the existing balance of power. The result of such decisions may end with ordering policies not compatible with JIT logic and place some disadvantages to either one of them or both. Furthermore, these policies may not consider the effect of transportation cost on the optimal ordering and shipment size quantities. A general joint economic lot size (JELS) model is developed under deterministic conditions focusing on the shipment size and its impact on the joint total cost, which also includes the transportation cost. Then a computational analysis is performed between individual lot size models and the proposed JELS model by utilizing a full factorial design. Finally, the effects of the experimental factors on the joint total cost are investigated by using an analysis of variance test.
Keywords

Authors	Pascal Lang, Reha Yolalan
Title	On Finite Multiplier Bounds In Data EnvelopmentAnalysis
Issue	MAY 1996 VOLUME: 8 NUMBER:1
Abstract
The role of the non-archimedean bounds on multipliers iri Data Envelopment Analysis has been the object of discussions in the literature. In this note, we examine some implications of using finite bounds as a proxy and sketch a modified procedure.
Keywords	Data Envelopment Analysis, Efficiency Measurement.

Authors	İlhan Or, E. Demirkol
Title	Optimization Issues in Automated Production of Printed Circuit Boards: Operations Sequencing and Feeder Configuration Problems
Issue	MAY 1996 VOLUME: 8 NUMBER:1
Abstract
Automatic electronic component insertion machines are integral parts of Computer Integrated Manufacturing Systems in electronics industry. During the automated manufacturing of a printed circuit board (PCB), these machines insert electronic components, which are supplied from a set of sequential feeders, into predefined locations on the PCB. The sequencing of the insertion operation and the assignment of the component types to feeders directly influence production rate and overall system productivity. In this research, a heuristic procedure is defined and implemented, where these two interrelated problems are taken up iteratively, first one leading to a four dimensional asymmetric, non-Eucledian traveling salesman problem and the second one leading to a quadratic assignment problem. Through this approach, near optimal, but very practical and effective insertion sequences and feeder assignments can be generated for electronic components insertion machines. Result obtained for a series of random testproblems are quite encouraging and will be presented, along with some important insights they have led to, about the developed solution procedure. An implementation to a real case, in a TV set production facility, will also be presented and, with this context, issues related to the validity of some critical model assumptions will be discussed.
Keywords	Operations Sequencing, Production Sequencing, Factory Automation, Computer Integrated Manugacturing Systems, Traveling SalesmanProblem, Quadratic Assignment Problem

Authors	Levent Kandiller
Title	En İyi İmalat Hücresini Belirleme Problemi
Issue	MAY 1996 VOLUME: 8 NUMBER:1
Abstract
Hücre tipi imalat sistemlerinin tasarımında, en iyi hücreyi belirleme problemi tek bir imalat hücresine konsantre olduğumuzda karşımıza çıkar. Makalede, imalat sistemleri hiperçizge ile modellenip, üzerinde en iyi hücreyi belirleme probleminin formülasyonu verilmiştir. Bu problemin, kısıtlarının belirlediği politopun yüzey (facet) tanımlayan eşitsizliklerinin yardımıyla, amaç fonksiyonunun yarattığı özel halde, polinom zamanda çözülebileceği ispatlanmıştır. Yapılan deneysel analizin sonuçları özetlenmiştir.
Keywords	Hücre Tipi İmalat Sistemleri, Polihedral Kombinatoriks

Authors	Gülser Köksal, William A. Smith, Jr.
Title	A Multi-Stage Application of Quality Function Deployment in Textiles Dyeing and Finishing
Issue	NOVEMBER 1995 VOLUME: 7 NUMBER: 2
Abstract
This paper presents a (Quality Function Deployment (QFD) application in a textiles dyeing and finishing case. The case has been formulated as a learning exercise for two textiles companies who have participated in this study. The paper presents a QFD application framework and results of the study for the stages of identifying and prioritizing customer requirements, translating those concurrently into product and process specifications, modeling relationships among product and process parameters, and developing process control plans.
Keywords	Quality Function Deployment, Quality Planning, Quality Assurance, Total Quality Control, Textiles Dyeing and Finishing.

Authors	Emrullah Demirci
Title	Üretim Benzetiminde Çıktıların Değerlendirilmesi ve Bir Uygulama
Issue	NOVEMBER 1995 VOLUME: 7 NUMBER: 2
Abstract
Üretim sistemleriyle ilgili sorunların ve uygun çözümlerin araştırılması benzetim tekniği ile yapılabilmektedir. Özellikle üretim sistemlerinin çalışma stratejilerinin benzetim modeli üzerinde sınanmaları, benzetim çıktılarının değerlendirilmesi ve sonuçların istatistiksel olarak değerlendirilerek çeşitli gösterimlerle verilmesi bu çalışmanın amacını oluşturmaktadır. Bu çalışmada geleneksel üretim sistemlerinden atölye tipi üretim sistemi özelliği gösteren her hangi bir atölyenin benzetim sonuçlarının değerlendirilmesi ve gösterimi yapılmıştır. Atölye tipi üretim sistemi için geliştirilen benzetim modeli gerçek sistem verileriyle etkileşimli olarak çalıştırılmış ve benzetim çıktıları istatistiksel olarak değerlendirilerek sonuçlar çizelge, grafik gibi şekillerle verilmiştir.
Keywords	Atölye Tipi Üretim Sistemi, Benzetim, Benzetim Çıktısı, İstatistiksel Analiz, Uygulama

Authors	Mustafa Ç. Pınar, Stavros A. Zenios
Title	An Entropic Approximation of the l1 Penalty Function
Issue	NOVEMBER 1995 VOLUME: 7 NUMBER: 2
Abstract
We propose an alternative to the well-known quadratic smoothing technique for the l1 penalty function. An entropic function is introduced and is shown to preserve the approximate exactness properties of the quadratic smoothing function. We develop a simple algorithm based on these ideas and give preliminary computational results using nonlinear minimax problems. Numerical results indicate that the performance of the entropic penalty function is comparable to the quadratic smoothing function. Furthermore, it is demonstrated that the penalty algorithm can be competitive with state-of the-art software for nonlinear programming.
Keywords	Convex Constrained Optimization, Exact Penalty Functions, Smoothing, Approximate Exactness, Entropy Functions, Minimax Optimization.

Authors	Azer Önel, Ömer Saatçioğlu
Title	Trend Analysis and Regression Modelling for the Assessment of Research  Performance at a Turkish University
Issue	MAY 1995 VOLUME: 7 NUMBER: 1
Abstract
Higher education institutions in several countries are going through the process of assessing their performance for various reasons. In this paper we discuss trend analysis in detecting changes in research performance and investigate into factors which are likely to affect performance. Multiple regression analyses are conducted to determine factors which are likely to influence research output. We demonstrate their applications with actual data on the operations of a Turkish university. International journal articles per professor is used as a surrogate measure of research output. We also discuss issues related to performance indicators and outline problems associated with data collection.
Keywords	Assessment of Research Performance, Performance Indicators, Trend Analysis, Regression Modelling

Authors	Esra Doğan, Tayfur Altıok
Title	Approximate Analysis of Transfer Lines with Finite Buffers
Issue	MAY 1995 VOLUME: 7 NUMBER: 1
Abstract
We are concerned with performance evaluation of transfer lines, also referred to as synchronous production lines, consisting of unreliable machines connected by conveyors (primary buffers). Upon a machine failure, the part on the machine is discarded. There are finite buffers between the machines to absorb the impact of failures. Parts are placed into a buffer when its downstream machine is down, and parts (if any) are removed from a buffer when its upstream machine is down. Blocking and starvation of the machines exist due to finite capacities of the buffers. We have developed an approximation method to obtain the average buffer contents and the system output rate. Exact results are obtained in the case of a two-machine-line. For longer lines, the approximation method decomposes the system into subsystems of two machines and analyzes them in isolation. The  two-node systems are linked to each other by a fixed-point iteration. The algorithm is shown to be highly efficient and accurate.
Keywords

Authors	Surendra M. Gupta
Title	N-Policy Queueing System with Finite Population
Issue	MAY 1995 VOLUME: 7 NUMBER: 1
Abstract
In this paper, the machine interference problem is considered under N-policy service with a startup time (or setup time). The steady state distribution of the number of machines in the system is sought. Closed form solution for such a system is not attainable. However, we observe that it is possible to calculate the steady state probabilities recursively and we derive closed form expressions for them. Also derived are the expressions for the measures of effectiveness of these types of models. Several special cases are considered. It is shown that one of the special cases of the model can easily be extended to incorporate various vacation models. Closed form expressions for those are provided. An efficient and generalized algorithm is presented to find the steady state distribution of the number of machines in the system as well as the measures of effectiveness. An example and some sensitivity results are also presented.
Keywords	Finite source Queues, N-Policy problem, Vacation models,       Performance-analysis, Machine interference problem.

Authors	Jeffrey E. Teich, Hannele Wallenius, Jyrki Wallenius
Title	Advances in Negotiation Science
Issue	NOVEMBER 1994 VOLUME: 6 NUMBER:2
Abstract
We review the state-of-the-art of negotiation science. Initially, the problem characteristics of negotiations are defined. Literature on behavioral aspects is then reviewed. We continue with a description of the management science negotiation and modeling approach. Reflecting our own research in the field, we emphasize the negotiation analysis modeling aspects, including a description of a number of support systems to aid negotiation processes.
Keywords

Authors	Barış Tan, Sencer Yeralan
Title	A Unified Boundary-Interior Model for a Two-Station Continuous Materials Flow Production Line with Station Breakdown
Issue	NOVEMBER 1994 VOLUME: 6 NUMBER:2
Abstract
A new methodology to solve unreliable continuous materials flow production line problems is presented. A model of an unreliable two-station production line with an interior-station buffer storage and continuous material flow is discussed. The steady-state probabilities are determined by solving a set of differential equations. The model uses a limit argument to describe the behaviour of the system during the periods of blocking and starvation. Closed-form expressions are developed for the system availability and expected level of the buffer. Limiting behavior of the production rate and the expected level of the buffer are examined.
Keywords	Production lines, buffer, differential equations, steady-state probablities, stochastic processes, production.

Authors	İ. Kuban Altınel, Necati Aras
Title	Türkiye İçinde Karayolu Uzaklıklarının Yaklaşık Hesabı
Issue	NOVEMBER 1994 VOLUME: 6 NUMBER:2
Abstract
Bir bölge içinde iki nokta arasındaki gerçek yol uzaklığını, bütün nokta çiftleri için ölçmek, her zaman olanaklı olmayabilir. Böyle bir