Operations Research, Systems Engineering and Industrial Engineering Commons^™

The Delta Test: The 1-Nn Estimator As A Feature Selection Criterion, Emil Eirola, Amaury Lendasse, Francesco Corona, Michel Verleysen

Engineering Management and Systems Engineering Faculty Research & Creative Works

Feature Selection is Essential in Many Machines Learning Problem, But It is Often Not Clear on Which Grounds Variables Should Be Included or Excluded. This Paper Shows that the Mean Squared Leave-One-Out Error of the First Nearest-Neighbour Estimator is Effective as a Cost Function When Selecting Input Variables for Regression Tasks. a Theoretical Analysis of the Estimator's Properties is Presented to Support its Use for Feature Selection. an Experimental Comparison to Alternative Selection Criteria (Including Mutual Information, Least Angle Regression, and the Rrelieff Algorithm) Demonstrates Reliable Performance on Several Regression Tasks.

Variable Selection For Regression Problems Using Gaussian Mixture Models To Estimate Mutual Information, Emil Eirola, Amaury Lendasse, Juha Karhunen

Engineering Management and Systems Engineering Faculty Research & Creative Works

Variable Selection is a Crucial Part of Building Regression Models and is Preferably Done as a Filtering Method Independently from the Model Training. Mutual Information is a Popular Relevance Criterion for This, But It is Not Trivial to Estimate Accurately from a Limited Amount of Data. in This Paper, a Method is Presented Where a Gaussian Mixture Model is Used to Estimate the Joint Density of the Input and Output Variables, and Subsequently Used to Select the Most Relevant Variables by Maximizing the Mutual Information Which Can Be Estimated using the Model.

Rssi And Lqi Data Clustering Techniques To Determine The Number Of Nodes In Wireless Sensor Networks, Yanwen Wang, Ivan G. Guardiola, Xiaoling Wu

Engineering Management and Systems Engineering Faculty Research & Creative Works

With the rapid proliferation of wireless sensor networks, different network topologies are likely to exist in the same geographical region, each of which is able to perform its own functions individually. However, these networks are prone to cause interference to neighbor networks, such as data duplication or interception. How to detect, determine, and locate the unknown wireless topologies in a given geographical area has become a significant issue in the wireless industry. This problem is especially acute in military use, such as spy-nodes detection and communication orientation systems. In this paper, three different clustering methods are applied to classify the …

How To Rein In The Volatile Actor: A New Bounded Perspective, Abhijit Gosavi

Engineering Management and Systems Engineering Faculty Research & Creative Works

Actor-critic algorithms are amongst the most well-studied reinforcement learning algorithms that can be used to solve Markov decision processes (MDPs) via simulation. Unfortunately, the parameters of the so-called "actor" in the classical actor-critic algorithm exhibit great volatility - getting unbounded in practice, whence they have to be artificially constrained to obtain solutions in practice. The algorithm is often used in conjunction with Boltzmann action selection, where one may have to use a temperature to get the algorithm to work, but the convergence of the algorithm has only been proved when the temperature equals 1. We propose a new actor-critic algorithm …

Carbon Constrained Integrated Inventory Control And Truckload Transportation With Heterogeneous Freight Trucks, Dinçer Konur

Engineering Management and Systems Engineering Faculty Research & Creative Works

This paper analyzes an integrated inventory control and transportation problem with environmental considerations. Particularly, explicit transportation modeling is included with inventory control decisions to capture per truck costs and per truck capacities. Furthermore, a carbon cap constraint on the total emissions is formulated by considering emission characteristics of various trucks that can be used for inbound transportation. Due to complexity of the resulting optimization problem, a heuristic search method is proposed based on the properties of the problem. Numerical studies illustrate the efficiency of the proposed method. Furthermore, numerical examples are presented to show that both costs and emissions can …

Nonlinear Modeling Using Neural Networks For Trading The Soybean Complex, Phoebe S. Wiles, David Lee Enke

Engineering Management and Systems Engineering Faculty Research & Creative Works

Recently, there has been a spike in the prices and popularity of commodities. On a macroeconomic level, developing countries are increasing production; while on a microeconomic level, speculative traders are becoming more involved in the market. Agricultural products have a diverse array of factors that can affect the price (i.e. political, government, population, weather, supply and demand). Commodity prices can suffer from extreme volatility in the short term, changing as much as 50% in one year. This research uses the soybean crush spread as a model. The soybean complex adds an interesting component as the underlying soybean product can be …

Computational Complexity Measures For Many-Objective Optimization Problems, David M. Curry, Cihan H. Dagli

Engineering Management and Systems Engineering Faculty Research & Creative Works

Multi-objective Optimization Problems (MOPs) are commonly encountered in the study and design of complex systems. Pareto dominance is the most common relationship used to compare solutions in MOPs, however as the number of objectives grows beyond three, Pareto dominance alone is no longer satisfactory. These problems are termed "Many-Objective Optimization Problems (MaOPs)". While most MaOP algorithms are modifications of common MOP algorithms, determining the impact on their computational complexity is difficult. This paper defines computational complexity measures for these algorithms and applies these measures to a Multi-Objective Evolutionary Algorithm (MOEA) and its MaOP counterpart.

Study Of The Use Of A Genetic Algorithm To Improve Networked System-Of-Systems Resilience, Charles O. Adler, Cihan H. Dagli

Engineering Management and Systems Engineering Faculty Research & Creative Works

Large scale failures or degradation resulting from smaller initial failures or disruptions in networked system-of-systems are an issue in multiple areas - for example cascading failures in electrical power grids or large disruptions in national air traffic due to local or regional weather conditions. The system architecture can have a significant impact on system-of-systems susceptibility to large scale failures. The study presented in this paper uses a simple interdependent networked system-of-systems failure model, integrated into a unique objective function that addresses both the overall level of failure and the rate of failure progression, and a genetic algorithm to demonstrate an …

Ieee Cibcb 2014-Computational Intelligence In Bioinformatics And Computational Biology [Conference Reports], Steven Corns

Engineering Management and Systems Engineering Faculty Research & Creative Works

The 2014 Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB) was held at the Hilton Hawaiian Village, Honolulu, from May 21-25. The papers presented at the conference covered the full range of computational intelligence techniques and applications. James Hughes from Brock University received the best student paper award from General Chair Steven Corns for his paper titled, 'Recentering and Restarting Genetic Algorithms Variations for DNA Fragment Assembly.' The best overall paper was awarded to Jennifer Hallinan, Owen Gilfellon, Goksel Misirli and Anil Wipat, all of Newcastle University, for their work 'Tuning Receiver Characteristics in Bacterial Quorum Communication: An …

Volatility Forecasting Using A Hybrid Gjr-Garch Neural Network Model, Soheil Almasi Monfared, David Lee Enke

Engineering Management and Systems Engineering Faculty Research & Creative Works

Volatility forecasting in the financial markets, along with the development of financial models, is important in the areas of risk management and asset pricing, among others. Previous testing has shown that asymmetric GARCH models outperform other GARCH family models with regard to volatility prediction. Utilizing this information, three popular Neural Network models (Feed-Forward with Back Propagation, Generalized Regression, and Radial Basis Function) are implemented to help improve the performance of the GJR (1,1) method for estimating volatility over the next forty-four trading days. During training and testing, four different economic cycles have been considered between 1997-2011 to represent real and …

A Study Of The Effect Of Basic Network Characteristics On System-Of-System Failure Propagation, Charles O. Adler, Cihan H. Dagli

Engineering Management and Systems Engineering Faculty Research & Creative Works

Real life examples of large system level failures in complex networked system are fairly common, with national and regional level power grids providing numerous cases where local failures have resulted in broad systemic failures. As large networked systems-of-systems are increasingly common, their susceptibility to large scale failures is of significant interest. This paper presents the results of a study using a simple model to investigate the impacts of basic architecture characteristics on the spread of failures through a system-of-system after an initial failure occurs. The study reported here uses a non-symmetric inter-grid only failure model to investigate the sensitivity of …

On The Flexibility Of Systems In System Of Systems Architecting, Dincer Konur, Hadi Farhangi, Cihan H. Dagli

Engineering Management and Systems Engineering Faculty Research & Creative Works

System of Systems (SoS) architecting requires analyzing a set of individual but interconnected systems simultaneously in order to build a communicating SoS, which can provide the capabilities needed. In general, the systems can provide a set of capabilities and the SoS architect needs to decide which systems to include in the SoS so that each capability is provided by at least one system. In this case, the systems are inflexible, i.e., a selected system will contribute to the SoS with all the capabilities it can provide. On the other hand, if SoS architect can incentivize systems to contribute specific capabilities …

Fuzzy Optimization Of Acknowledged System Of Systems Meta-Architectures For Agent Based Modeling Of Development, Louis Pape, Siddhartha Agarwal, Kristin Giammarco, Cihan Dagli

Engineering Management and Systems Engineering Faculty Research & Creative Works

Acknowledged systems of systems (SoS) lie on a continuum between authoritarian central control and anarchy. The constituent systems are independent, with a life and purpose of their own. The systems require not only technical interfaces, but also social interactions, to create the SoS. A fuzzy optimization process may be used to select a desirable SoS configuration, but it may be unachievable due to the inability to persuade the systems to cooperate in the plan. Modeling the systems' internal decision processes could help understand how to design better SoS architectures. This research used generic, modular modeling processes to examine two proposed …

Conquering Complexity: Challenges And Opportunities: Preface, Cihan H. Dagli

Engineering Management and Systems Engineering Faculty Research & Creative Works

No abstract provided.

Part Viii: Health Care Analytics: Preface, Cilan H. Dagli

Engineering Management and Systems Engineering Faculty Research & Creative Works

No abstract provided.

A Hybrid Neuro-Fuzzy Model To Forecast Inflation, David Lee Enke, Nijat Mehdiyev

Engineering Management and Systems Engineering Faculty Research & Creative Works

One of the key issues in constructing monetary policy is accurate prediction of the inflation level. The complex behavior and non-linear nature of the financial markets makes it hard to forecast the inflation rate precisely. This paper introduces a hybrid model that attempts to forecast the inflation rate with a combination of a subtractive clustering technique and a fuzzy inference neural network to overcome the shortcomings of the individual methodologies. Selected macroeconomic factors were used to predict the historical CPI data from the US Markets. The results of the proposed hybrid model are measured in RMSE.

A Hybrid Genetic Algorithm And Particle Swarm Optimization With Type-2 Fuzzy Sets For Generating Systems Of Systems Architectures, Siddhartha Agarwal, Louis E. Pape, Cihan H. Dagli

Engineering Management and Systems Engineering Faculty Research & Creative Works

Both modeling and simulating a system of systems (SoS) are difficult due not only to a changing environment but also the unique behavior that is linked to different participating systems. Generating architectures for a SoS is a multi-objective optimization problem with large number of variables and constraints. The paper presents several of computational intelligence techniques that can generate SoS architectures, such as genetic algorithms (GA), and particle swarm optimization (PSO) combined with Type 2 Fuzzy logic nets. The Maritime search and rescue (SAR) were used as a SoS domain scenario to both implement and demonstrate the architecting methodology. SAR utilizes …

Joint Decisions On Inventory Replenishment And Emission Reduction Investment Under Different Emission Regulations, Ayşegül Toptal, Haşim Özlü, Dinçer Konur

Engineering Management and Systems Engineering Faculty Research & Creative Works

Carbon emission regulation policies have emerged as mechanisms to control firms' carbon emissions. To meet regulatory requirements, firms can make changes in their production planning decisions or invest in green technologies. In this study, we analyze a retailer's joint decisions on inventory replenishment and carbon emission reduction investment under three carbon emission regulation policies. Particularly, we extend the economic order quantity model to consider carbon emissions reduction investment availability under carbon cap, tax and cap-and-trade policies. We analytically show that carbon emission reduction investment opportunities, additional to reducing emissions as per regulations, further reduce carbon emissions while reducing costs. We …

Long-Term Time Series Prediction Using Op-Elm, Alexander Grigorievskiy, Yoan Miche, Anne Mari Ventelä, Eric Séverin, Amaury Lendasse

Engineering Management and Systems Engineering Faculty Research & Creative Works

In This Paper, an Optimally Pruned Extreme Learning Machine (Op-Elm) is Applied to the Problem of Long-Term Time Series Prediction. Three Known Strategies for the Long-Term Time Series Prediction I.e. Recursive, Direct and Dirrec Are Considered in Combination with Op-Elm and Compared with a Baseline Linear Least Squares Model and Least-Squares Support Vector Machines (Ls-Svm). among These Three Strategies Dirrec is the Most Time Consuming and its Usage with Nonlinear Models Like Ls-Svm, Where Several Hyperparameters Need to Be Adjusted, Leads to Relatively Heavy Computations. It is Shown that Op-Elm, Being Also a Nonlinear Model, Allows Reasonable Computational Time for …

Rmse-Elm: Recursive Model Based Selective Ensemble Of Extreme Learning Machines For Robustness Improvement, Bo Han, Bo He, Mengmeng Ma, Tingting Sun, Tianhong Yan, Amaury Lendasse

Engineering Management and Systems Engineering Faculty Research & Creative Works

For Blended Data, the Robustness of Extreme Learning Machine (Elm) is So Weak Because the Coefficients (Weights and Biases) of Hidden Nodes Are Set Randomly and the Noisy Data Exert a Negative Effect. to Solve This Problem, a New Framework Called "RMSE-ELM" is Proposed in This Paper. It is a Two-Layer Recursive Model. in the First Layer, the Framework Trains Lots of Elms in Different Ensemble Groups Concurrently and Then Employs Selective Ensemble Approach to Pick Out an Optimal Set of Elms in Each Group, Which Can Be Merged into a Large Group of Elms Called Candidate Pool. in the …

Fast Feature Selection In A Gpu Cluster Using The Delta Test, Alberto Guillén, M. Isabel García Arenas, Mark Van Heeswijk, Dusan Sovilj, Amaury Lendasse, Luis Javier Herrera, Héctor Pomares, Ignacio Rojas

Engineering Management and Systems Engineering Faculty Research & Creative Works

Feature or Variable Selection Still Remains an Unsolved Problem, Due to the Infeasible Evaluation of All the Solution Space. Several Algorithms based on Heuristics Have Been Proposed So Far with Successful Results. However, These Algorithms Were Not Designed for Considering Very Large Datasets, Making their Execution Impossible, Due to the Memory and Time Limitations. This Paper Presents an Implementation of a Genetic Algorithm that Has Been Parallelized using the Classical Island Approach, But Also Considering Graphic Processing Units to Speed Up the Computation of the Fitness Function. Special Attention Has Been Paid to the Population Evaluation, as Well as to …

Sustainability Analysis In Integrated Inventory Control And Transportation Systems, Brian Joseph Schaefer

Doctoral Dissertations

"Due to the importance of costs as well as environmental effects of logistical activities throughout supply chains, such as inventory holding, freight transportation, and warehousing activities, this dissertation models and analyzes four integrated inventory control and transportation problems that account for economic and environmental aspects of a supply chain agents related decisions.

The first model presents an integrated inventory control and transportation problem in a single item deterministic demand setting. A supply chain agents inventory control and transportation mode selection problem is solved under carbon cap, carbon cap and trade, carbon cap and offset, and carbon tax regulations. The second …

Workplace Sleepiness: Causes, Consequences And Countermeasures, Neda Khalafi

Masters Theses

"Sleep deprivation is an epidemic problem in our society that adversely affects the quality and safety of our daily lives. At home and on the job, inadequate sleep is associated with considerable social, financial, and health-related costs, due to instability of waking functions. Research has shown that, sleep deprivation among workers, whether from sleep disorders, lifestyle or shift work, can hinder the proper performance, and in extreme cases may pose hazard to the workers themselves and their environment. Deficits in daytime performance can have profound effects in the processing industry where workers are often required to perform monotonous tasks that …

Data Classification And Forecasting Using The Mahalanobis-Taguchi Method, Adebolaji A. Jobi-Taiwo

Masters Theses

"Classification and forecasting are useful concepts in the field of condition monitoring. Condition monitoring refers to the analysis and monitoring of system characteristics to understand and identify deviations from normal operating conditions. This can be performed for prediction, diagnosis, or prognosis or a combination of any these purposes. Fault identification and diagnosis are usually achieved through data classification, while forecasting methods are usually used to accomplish the prediction objective. Data gathered from monitoring systems often consists of multiple multivariate time series and is fed into a model for data analysis using various techniques. One of the data analysis techniques used …

The Use Of Statistics And Its Analysis In Resolving Transportation Related Problems, Venkat Sravan Kothapalli

Masters Theses

"Statistical methods are traditionally used to summarize or describe a collection of data. This can be termed as descriptive statistics. This research focuses on application of certain statistical approaches, tools to solve transportation related problems. This research focuses on usage of statistical analysis in determine some of the significant factors and assumptions which should be taken into consideration while developing safety and maintenance related measures by transportation agencies. This research mainly focuses on; testing null hypothesis and frequency analysis. The research therefore is presented in two different case studies: 1) Statistical analysis of the work zone crash historical raw data …

Quantifying Economic Benefits For Rail Infrastructure Projects, Akhilesh Ojha

Masters Theses

"Investment in rail infrastructure is necessary to maintain existing service and to cater for future growth in freight and passenger services. Many communities have realized the importance of investment in rail infrastructure projects and set up goals and visions to achieve economic development through investing in such projects. Due to limited funds available, communities have to select a single or very few projects from a variety of projects. It is very critical that right projects must be selected at the right time for a community to realize economic development. The limited methods for quantifying the economic benefits to the stakeholders …

Use On Multinomial Logistic Regression In Work Zone Crash Analysis For Missouri Work Zones, Paul Robin

Masters Theses

"This study focuses on the use of statistical data analysis procedures in identifying factors which affect the severity of crashes in work zones. Work zones are unsafe for the traffic passing through as well as the workers. Multinomial Logistic Regression has been used to analyse Missouri work zone crash data to identify significant factors which affect the severity of crashes. This particular type of regression analysis was used due to the mixed nature of data. Multinomial regression was used to compare crashes with severity Property Damage Only against crashes with Minor Injuries and Disabling Injuries/ Fatal. The factors considered were …

Hedge Fund Replication Using A Strategy Specific Modeling Approach, Sujit Subhash

Masters Theses

"Institutional investors and wealthy individuals have in the past allocated a significant portion of their portfolios to hedge funds with the expectation of unconditional and uncorrelated returns to the market. However, the financial crisis of 2008 has heightened investor sensitivity to the high fees, illiquidity, and lockup periods typically associated with hedge funds. Hedge fund indexes showing excellent returns and low volatility contain funds that are closed to new investments, while the performance of investable funds have been shown to be inferior to their non-investable counterparts. The lack of transparency and extreme variation in the performance of hedge funds make …

Economic And Environmental Comparison Of Different Ordering Policies For An Integrated Inventory Control And Supplier Selection Problem, Sepideh Almasi Monfared

Masters Theses

"This study analyzes an integrated inventory control and supplier selection problem in stochastic demand environment under carbon emissions regulations. In particular, a continuous review inventory model with multiple suppliers is investigated under carbon taxing and carbon trading regulations. We analyze and compare the optimal supplier selection and order splitting decisions with single sourcing and two alternative delivery structures for multi-sourcing, namely, sequential ordering and sequential delivery. For each of the three ordering policies, a solution method is proposed and these policies are compared in terms of their economic as well as environmental performances. A numerical study is conducted to demonstrate …

Timekeeping Issues In Ultra-Quality Metering Systems, David D. Haynes, Steven M. Corns

Engineering Management and Systems Engineering Faculty Research & Creative Works

The improvements made to smart metering systems in terms of performance and accuracy have in many cases created a gap in the requirements for timekeeping performance. When clock errors accumulate in different metrology systems, data collected from one system cannot be compared to data collected from another system. This paper recommends the use of a common time base for all data gathering systems, and that systems clocks are corrected with numerous minor corrections when possible, rather than major corrections. © 2013 IEEE.