Industrial & Enterprise Systems Engineering Calendar
http://illinois.edu/calendar/list/3604
Includes events and seminars for ISE and MSFEEnd of Summer Session 2
http://illinois.edu/calendar/detail/3604/29250961
Academichttp://illinois.edu/calendar/detail/3604/29250961Thu, 07 Aug 2014 08:30:00 CDTFinal Exams Begin
http://illinois.edu/calendar/detail/3604/29250963
Academichttp://illinois.edu/calendar/detail/3604/29250963Fri, 08 Aug 2014 08:30:00 CDTXi Chen Final Exam-NEW MODEL-BASED METHODS FOR NON-DIFFERENTIABLE OPTIMIZATION
http://illinois.edu/calendar/detail/3604/31981329
Seminarhttp://illinois.edu/calendar/detail/3604/31981329Fri, 15 Aug 2014 13:30:00 CDTAbstractModel-based optimization methods are effective for solving optimization problems with little structure, such as convexity and differentiability. The algorithms iteratively find candidate solutions by generating samples from a parameterized probabilistic model on the solution space, and update the parameter of the probabilistic model based on the objective function evaluations. This dissertation explores new model-based optimization methods, and mainly consists of three topics.The first topic of the dissertation proposes two new model-based algorithms for discrete optimization under the framework of gradient-based adaptive stochastic search (GASS), where the parameter of the probabilistic model is updated based on a direct gradient method. We propose two discrete optimization algorithms: discrete gradient-based adaptive stochastic search (discrete-GASS) and annealing gradient-based adaptive stochastic search (annealing-GASS). The first algorithm, discrete-GASS, converts the discrete optimization problem to a continuous problem on the parameter space of a family of independent discrete distributions, and applies a gradient-based method to find the optimal parameter such that the corresponding distribution has the best capability to generate optimal solution(s) to the original discrete problem. The second algorithm, annealing-GASS, uses Boltzmann distribution as the parameterized probabilistic model, and derives a gradient-based temperature schedule, which changes adaptively with respect to the current performance of the algorithm, for updating the Boltzmann distribution. We prove the convergence of the two proposed methods, and conduct numerical experiments to compare these two methods as well as some other existing methods.The second topic of the dissertation proposes a framework of population model-based optimization (PMO) in order to better capture the multi-modality of the objective functions than the traditional model-based methods which use only a single model at every iteration. This PMO framework uses a population of models at every iteration with an adaptive mechanism to propagate the population over iterations. The adaptive mechanism is derived from estimating the optimal parameter of the probabilistic model in a Bayesian manner, and thus provides a proper way to determine the diversity in the population of the models. We provide theoretical justification on the convergence of this framework by showing that the posterior distribution of the parameter asymptotically converges to a degenerate distribution concentrating on the optimal parameter. Under this framework, we develop two practical algorithms by incorporating sequential Monte Carlo methods, and carry out numerical experiments to illustrate their performance.The last topic of the dissertation considers simulation optimization, where the objective function cannot be evaluated exactly and must be estimated by stochastic simulation. The objective functions are usually expensive to evaluate and the problems are difficult to solve. The idea of model-based methods for deterministic optimization is extended to stochastic optimization. We present two simulation optimization algorithms: gradient-based adaptive stochastic search for simulation optimization (GASSO) and simulation optimization with approximate Bayesian computation (SO-ABC), and carry out numerical experiments of these two algorithms. The first algorithm, GASSO, is an extension of the deterministic version of GASS. It reformulates the original simulation optimization problem into another optimization problem on the parameter space of the probabilistic model and uses a direct gradient method on the parameter space to update the probabilistic model. The second algorithm, SO-ABC, views the simulation optimization problem as an estimation problem, and uses the approximate Bayesian computation (ABC) technique to estimate the optimal solution.Alberth Alvarado Ortiz-Final Exam-CENTRALIZED AND DISTRIBUTED RESOURCE ALLOCATION WITH APPLICATIONS TO SIGNAL PROCESSING IN COMMUNICATIONS
http://illinois.edu/calendar/detail/3604/31981331
Final Examhttp://illinois.edu/calendar/detail/3604/31981331Wed, 20 Aug 2014 10:30:00 CDTAbstractModel-based optimization methods are effective for solving optimization problems with little structure, such as convexity and differentiability. The algorithms iteratively find candidate solutions by generating samples from a parameterized probabilistic model on the solution space, and update the parameter of the probabilistic model based on the objective function evaluations. This dissertation explores new model-based optimization methods, and mainly consists of three topics.The first topic of the dissertation proposes two new model-based algorithms for discrete optimization under the framework of gradient-based adaptive stochastic search (GASS), where the parameter of the probabilistic model is updated based on a direct gradient method. We propose two discrete optimization algorithms: discrete gradient-based adaptive stochastic search (discrete-GASS) and annealing gradient-based adaptive stochastic search (annealing-GASS). The first algorithm, discrete-GASS, converts the discrete optimization problem to a continuous problem on the parameter space of a family of independent discrete distributions, and applies a gradient-based method to find the optimal parameter such that the corresponding distribution has the best capability to generate optimal solution(s) to the original discrete problem. The second algorithm, annealing-GASS, uses Boltzmann distribution as the parameterized probabilistic model, and derives a gradient-based temperature schedule, which changes adaptively with respect to the current performance of the algorithm, for updating the Boltzmann distribution. We prove the convergence of the two proposed methods, and conduct numerical experiments to compare these two methods as well as some other existing methods.The second topic of the dissertation proposes a framework of population model-based optimization (PMO) in order to better capture the multi-modality of the objective functions than the traditional model-based methods which use only a single model at every iteration. This PMO framework uses a population of models at every iteration with an adaptive mechanism to propagate the population over iterations. The adaptive mechanism is derived from estimating the optimal parameter of the probabilistic model in a Bayesian manner, and thus provides a proper way to determine the diversity in the population of the models. We provide theoretical justification on the convergence of this framework by showing that the posterior distribution of the parameter asymptotically converges to a degenerate distribution concentrating on the optimal parameter. Under this framework, we develop two practical algorithms by incorporating sequential Monte Carlo methods, and carry out numerical experiments to illustrate their performance.The last topic of the dissertation considers simulation optimization, where the objective function cannot be evaluated exactly and must be estimated by stochastic simulation. The objective functions are usually expensive to evaluate and the problems are difficult to solve. The idea of model-based methods for deterministic optimization is extended to stochastic optimization. We present two simulation optimization algorithms: gradient-based adaptive stochastic search for simulation optimization (GASSO) and simulation optimization with approximate Bayesian computation (SO-ABC), and carry out numerical experiments of these two algorithms. The first algorithm, GASSO, is an extension of the deterministic version of GASS. It reformulates the original simulation optimization problem into another optimization problem on the parameter space of the probabilistic model and uses a direct gradient method on the parameter space to update the probabilistic model. The second algorithm, SO-ABC, views the simulation optimization problem as an estimation problem, and uses the approximate Bayesian computation (ABC) technique to estimate the optimal solution.First Day of Instruction, Fall Semester
http://illinois.edu/calendar/detail/3604/29250964
Academichttp://illinois.edu/calendar/detail/3604/29250964Mon, 25 Aug 2014 08:30:00 CDTGraduate Student Information and Welcome Session
http://illinois.edu/calendar/detail/3604/31989350
Seminarhttp://illinois.edu/calendar/detail/3604/31989350Thu, 28 Aug 2014 16:00:00 CDTAll ISE Graduate Students, Faculty, and Staff Welcome!! Food and Beverages Following the Presentation in303 Transportation Bldg.ISE Seminar featuring Yuan Zhong, Columbia University
http://illinois.edu/calendar/detail/3604/31992394
GE/IE 590 Seminarhttp://illinois.edu/calendar/detail/3604/31992394Thu, 04 Sep 2014 16:00:00 CDTISE Seminar featuring Hamsa Balakrishnan, MIT
http://illinois.edu/calendar/detail/3604/31992783
GE/IE 590 Seminarhttp://illinois.edu/calendar/detail/3604/31992783Thu, 11 Sep 2014 16:00:00 CDTISE Seminar featuring Dr. Don Wagner, US Navy
http://illinois.edu/calendar/detail/3604/31992792
GE/IE 590 Seminarhttp://illinois.edu/calendar/detail/3604/31992792Thu, 18 Sep 2014 11:00:00 CDTISE Seminar featuring Dr. Don Wagner, US Navy
http://illinois.edu/calendar/detail/3604/31992793
GE/IE 590 Seminarhttp://illinois.edu/calendar/detail/3604/31992793Thu, 18 Sep 2014 16:00:00 CDTISE Seminar featuring Alan Scheller-Wolf, Carnegie Mellon University
http://illinois.edu/calendar/detail/3604/31992942
GE/IE 590 Seminarhttp://illinois.edu/calendar/detail/3604/31992942Thu, 25 Sep 2014 16:00:00 CDTNational Security Agency Summer Program Information Session
http://illinois.edu/calendar/detail/3604/31992974
Seminarhttp://illinois.edu/calendar/detail/3604/31992974Fri, 26 Sep 2014 09:00:00 CDTSPORT (Summer Program for Operations Research Technology). Paid internship working current operations research problems at the NSA. Applicants must be U.S. Citizens and graduate students. Students in all engineering disciplines, mathematics, statistics, and computer science are welcome to attend!Program Highlights:-- Paid internship (12 weeks, May-August 2015)-- Applications accepted September 1 - November 15, 2014 (subject to change)-- Opportunity to apply operations research, mathematics, statistics, computer science, and/or engineering skills-- Real NSA mission problems-- Paid annual and sick leave, housing available, most travel costs coveredEngineering Career Services Resume Workshop
http://illinois.edu/calendar/detail/3604/31992975
Otherhttp://illinois.edu/calendar/detail/3604/31992975Thu, 09 Oct 2014 16:00:00 CDTISE Seminar featuring Phil Smith, The Ohio State University
http://illinois.edu/calendar/detail/3604/31992977
GE/IE 590 Seminarhttp://illinois.edu/calendar/detail/3604/31992977Thu, 13 Nov 2014 16:00:00 CSTThanksgiving Break Begins
http://illinois.edu/calendar/detail/3604/29250965
Academichttp://illinois.edu/calendar/detail/3604/29250965Sat, 22 Nov 2014 08:30:00 CSTInstruction resumes
http://illinois.edu/calendar/detail/3604/29250949
Academichttp://illinois.edu/calendar/detail/3604/29250949Mon, 01 Dec 2014 08:30:00 CSTLast Day of Instruction, Fall Semester
http://illinois.edu/calendar/detail/3604/29251031
Academichttp://illinois.edu/calendar/detail/3604/29251031Wed, 10 Dec 2014 08:30:00 CSTIT No Change Period in effect
http://illinois.edu/calendar/detail/3604/31225358
Academichttp://illinois.edu/calendar/detail/3604/31225358Thu, 11 Dec 2014 08:30:00 CSTThe CITES No Change period is designed to minimize the number of technical disruptions that might occur around final exams.
Beginning on Reading Day and continuing a few days past the end of finals, CITES tries to make no technical changes to hardware or software that might disrupt the availability of the campus network, data storage, web sites or any of the software-based services that CITES provides to campus. Changes or updates will only be made by CITES for emergency purposes during the No Change period.Final Exams Begin
http://illinois.edu/calendar/detail/3604/29251032
Academichttp://illinois.edu/calendar/detail/3604/29251032Fri, 12 Dec 2014 08:30:00 CSTFirst Day of Instruction, Spring Semester
http://illinois.edu/calendar/detail/3604/29251098
Academichttp://illinois.edu/calendar/detail/3604/29251098Tue, 20 Jan 2015 08:30:00 CST