Research Outputs

Permanent URI for this communityhttps://hdl.handle.net/20.500.14288/2

Browse

Filters

Advanced Search

Filter by

Settings

Department: search.filters.department.Department of Industrial EngineeringQuartile: search.filters.quartile.Q4

Search Results

Now showing 1 - 10 of 22
  • Placeholder
    PublicationMetadata only
    A kernel-based multilayer perceptron framework to identify pathways related to cancer stages
    (Springer International Publishing Ag, 2023) Mokhtaridoost, Milad; Department of Industrial Engineering; Soleimanpoor, Marzieh; Gönen, Mehmet; Department of Industrial Engineering; Graduate School of Sciences and Engineering; College of Engineering
    Standard machine learning algorithms have limited knowledge extraction capability in discriminating cancer stages based on genomic characterizations, due to the strongly correlated nature of high-dimensional genomic data. Moreover, activation of pathways plays a crucial role in the growth and progression of cancer from early-stage to latestage. That is why we implemented a kernel-based neural network framework that integrates pathways and gene expression data using multiple kernels and discriminates early- and late-stages of cancers. Our goal is to identify the relevant molecular mechanisms of the biological processes which might be driving cancer progression. As the input of developed multilayer perceptron (MLP), we constructed kernel matrices on multiple views of expression profiles of primary tumors extracted from pathways. We used Hallmark and Pathway Interaction Database (PID) datasets to restrict the search area to interpretable solutions. We applied our algorithm to 12 cancer cohorts from the Cancer Genome Atlas (TCGA), including more than 5100 primary tumors. The results showed that our algorithm could extract meaningful and disease-specific mechanisms of cancers. We tested the predictive performance of our MLP algorithm and compared it against three existing classification algorithms, namely, random forests, support vector machines, and multiple kernel learning. Our MLP method obtained better or comparable predictive performance against these algorithms.
  • Placeholder
    PublicationMetadata only
    A tandem queueing model with coupled processors
    (Elsevier, 2003) Resing, Jacques; Department of Industrial Engineering; Örmeci, Lerzan; Faculty Member; Department of Industrial Engineering; College of Engineering; 32863
    We consider a tandem queue with coupled processors and analyze the two-dimensional Markov process representing the numbers of jobs in the two stations. A functional equation for the generating function of the stationary distribution of this two-dimensional process is derived and solved through the theory of Riemann-Hilbert boundary value problems.
  • Placeholder
    PublicationMetadata only
    Admission and termination control of a two class loss system
    (Taylor & Francis Inc, 2011) Ulukuş, Mehmet Yasin; Güllü, Refik; Department of Industrial Engineering; Örmeci, Lerzan; Faculty Member; Department of Industrial Engineering; College of Engineering; 32863
    We consider dynamic admission and termination control policies in a Markovian loss system with two classes, each with a fixed reward, a termination cost, an arrival and service rate. The system may admit or reject an arriving job or admit it by terminating a job in the system to maximize its total expected discounted reward. We prove that (1) when there is an idle server, it is never optimal to terminate a job, (2) there exists an optimal threshold policy for both admission and termination decisions. Furthermore, we identify the conditions which ensure that a class is "preferred" or "strongly preferred."
  • Placeholder
    PublicationMetadata only
    Admission policies for a two class loss system with general interarrival times
    (Taylor & Francis Inc, 2006) van der Wal, J.; Department of Industrial Engineering; Örmeci, Lerzan; Faculty Member; Department of Industrial Engineering; College of Engineering; 32863
    This paper considers the problem of dynamic admission control in a loss queueing system with two classes of jobs. The jobs require an exponential amount of service time with different means and bring different revenues, whereas the arrivals occur according to a general distribution. We establish the existence of optimal acceptance thresholds for both job classes and show that under certain conditions there exists a preferred class. We also provide an example to demonstrate that for a Markov modulated Poisson arrival process there may be states in which both classes are rejected.
  • Placeholder
    PublicationMetadata only
    Characterizing the performance of process flexibility structures
    (Elsevier, 2007) N/A; Department of Business Administration; Department of Industrial Engineering; Karaesmen, Zeynep Akşin; Karaesmen, Fikri; Faculty Member; Faculty Member; Department of Business Administration; Department of Industrial Engineering; College of Administrative Sciences and Economics; College of Engineering; 4534; 3579
    The objective is to identify preferred flexibility structures in service or manufacturing systems, when demand is random and capacity is finite. Considering a network flow type model as the basis of the analysis, general structural properties of flexibility design pertaining to the marginal values of flexibility and capacity are identified.
  • Placeholder
    PublicationMetadata only
    Constant risk aversion in stochastic contests with exponential completion times
    (Wiley, 2019) Rothblum, Uriel G.; Department of Industrial Engineering; Canbolat, Pelin Gülşah; Faculty Member; Department of Industrial Engineering; College of Engineering; 108242
    This article analyzes a class of stochastic contests among multiple players under risk-averse exponential utility. In these contests, players compete over the completion of a task by simultaneously deciding on their investment, which determines how fast they complete the task. The completion time of the task for each player is assumed to be an exponentially distributed random variable with rate linear in the player's investment and the completion times of different players are assumed to be stochastically independent. The player that completes the task first earns a prize whereas the remaining players earn nothing. The article establishes a one-to-one correspondence between the Nash equilibrium of this contest with respect to risk-averse exponential utilities and the nonnegative solution of a nonlinear equation. Using the properties of the latter, it proves the existence and the uniqueness of the Nash equilibrium, and provides an efficient method to compute it. It exploits the resulting representation of the equilibrium investments to determine the effects of risk aversion and the differences between the outcome of the Nash equilibrium and that of a centralized version.(c) 2016 Wiley Periodicals, Inc. Naval Research Logistics 66:4-14, 2019
  • Placeholder
    PublicationMetadata only
    Dynamic pricing and scheduling in a multi-class single-server queueing system
    (Springer, 2011) Cil, Eren Basar; Department of Industrial Engineering; Department of Industrial Engineering; Karaesmen, Fikri; Örmeci, Lerzan; Faculty Member; Faculty Member; Department of Industrial Engineering; College of Engineering; College of Engineering; 3579; 32863
    This paper investigates an optimal sequencing and dynamic pricing problem for a two-class queueing system. Using a Markov Decision Process based model, we obtain structural characterizations of optimal policies. In particular, it is shown that the optimal pricing policy depends on the entire queue length vector but some monotonicity results prevail as the composition of this vector changes. A numerical study finds that static pricing policies may have significant suboptimality but simple dynamic pricing policies perform well in most situations.
  • Placeholder
    PublicationMetadata only
    Effects of system parameters on the optimal policy structure in a class of queueing control problems
    (Springer, 2009) Cil, Eren Basar; Department of Industrial Engineering; Department of Industrial Engineering; Örmeci, Lerzan; Karaesmen, Fikri; Faculty Member; Faculty Member; Department of Industrial Engineering; College of Engineering; College of Engineering; 32863; 3579
    This paper studies a class of queueing control problems involving commonly used control mechanisms such as admission control and pricing. It is well established that in a number of these problems, there is an optimal policy that can be described by a few parameters. From a design point of view, it is useful to understand how such an optimal policy varies with changes in system parameters. We present a general framework to investigate the policy implications of the changes in system parameters by using event-based dynamic programming. In this framework, the control model is represented by a number of common operators, and the effect of system parameters on the structured optimal policy is analyzed for each individual operator. Whenever a queueing control problem can be modeled by these operators, the effects of system parameters on the optimal policy follow from this analysis.
  • Placeholder
    PublicationMetadata only
    Inferring transferable intermolecular potential models
    (2008) Üçyigitler, Sinan; Çamurdan, Mehmet C.; Elliott, J. Richard; Department of Industrial Engineering; Türkay, Metin; Faculty Member; Department of Industrial Engineering; College of Engineering; 24956
    Discontinuous molecular dynamics is combined with thermodynamic perturbation theory to provide an efficient basis for characterising molecular interactions based on vapour pressure and liquid density data. Several prospective potential models are discretised to permit treatment by Barker–Henderson perturbation theory. The potentials are characterised by 11 wells ranging over radial distances from the site diameter to three times that diameter. Considered potential models include the Lennard-Jones (LJ), square-well, Yukawa (Yuk) and multi-line potentials, and their combinations. The optimal model is found to be a combination of square-well and Yuk potentials, with the switch position and Yuk decay set to universal values. This model provides average vapour pressure deviations of less than 10% for a database of 86 aliphatic, aromatic and naphthenic compounds. The LJ potential provides the least competitive accuracy. Considering statistical information criteria facilitates the identification of the optimal model.
  • Placeholder
    PublicationMetadata only
    MTTF and availability of semi-Markov missions with non-identical generally distributed component lifetimes
    (Taylor & Francis) Cekyay, Bora; Department of Industrial Engineering; Özekici, Süleyman; Faculty Member; Department of Industrial Engineering; College of Engineering; 32631
    We analyze mean time to failure and availability of systems that perform semi-Markov missions. The mission process is the minimal semi-Markov process associated with a Markov renewal process. Therefore, the successive phases of the mission follow a Markov chain, and the phase durations are generally distributed. The lifetimes of the non-identical components in the system are assumed to be generally distributed and are modeled using intrinsic aging concepts. Moreover, the lifetime parameters of the components and the configuration of the system change depending on the phases of the mission. We characterize the mean time to failure through solving a Poisson equation, and we analyze the system availability assuming that repair duration has a general distribution which is dependent on the phase of the mission during which the failure has occurred and on the deterioration level of the system.