CEACSE (Center of Excellence in Applied Computational Science and Engineering)
The mission of the Center of Excellence in Applied Computational Science and Engineering (CEACSE) is to establish and expand a cohesive multidisciplinary effort in applied computational science and engineering that is leveraged across UTC and produce sustained growth in research funding, excellence in integrated education and research, and to increase national and international stature and competitiveness in Tennessee.
These funds are awarded annually on a competitive basis. The primary goal of this program is to enable development of new capabilities and extramural projects in the area of Computational Sciences. Awards support CEACSE strategic priority areas of Urban Science, Energy & Environment, Defense/Aerospace, and Biomedical research.
The overall purpose of the Center of Excellence in Applied Computational Science is to establish a cohesive and expanding base of multidisciplinary research in applied computational science and engineering to produce sustained growth in research funding, excellence in integrated research and education, and increases in national and international stature and economic competitiveness for Tennessee.
Computational simulation is critically important for the analysis and design of future high technology products and systems in a competitive global marketplace. The future security and economic well being of our country will depend in part on an adequate supply of scientists and engineers who are highly skilled in the use of computers to solve important engineering problems using modeling and simulation.
This evolution is transforming the use of high technology by introducing computational simulation and design software that supplements experiments and testing to produce competitive advantages in critical areas such as price, time-to-market, life-cycle costs, and overhead. Although these benefits to industry are driving the changes in engineering practice, science education in the U.S. has not responded adequately to the challenge of providing graduates who are adequately prepared.
In view of the extensive use of computational methodologies in design by industry, there is a significant role for innovative programs of integrated research and graduate education (i.e., graduate research in an applications environment) that is distinct from traditional university research activity.
The use of computers to solve complex, large-scale, practical problems is a trend that will accelerate in years to come.
UTC has recognized that these prospects now offer a dramatic window of opportunity to provide the leadership in computational applications driven research and education needed for future competitiveness in the high-technology sector of the global economy. UTC has also positioned itself through past research and faculty additions to provide this leadership for Tennessee.
This year a total of $684,332 was awarded to nine lead principal investigators and twelve collaborating investigators across eight different departments.
2018-2019 CEACSE Awardees
Project Title: “Urban Electric Vehicle Charging Markets: Computational Modeling and Optimal Design”
Dr. Vahid Rasouli Disfani, Lead PI in collaboration with Dr. Mina Sartipi, and Dr. M. Ahmadi
Abstract: Maximizing utilization of electric vehicle supply equipment (EVSE)—or electric vehicle (EV) charging stations—is still a challenge for cities like Chattanooga despite the emergence of EV station locators like PlugShare and ChargeHub. The missing key element in this market is the lack of data from the demand side of EVs, which often leaves EVs desperate for charging not connected while EVSEs are available nearby. This project computationally models and designs an infrastructure that simultaneously gathers demand (EV) data—including desired destinations, connection period, and energy demand—as well as EVSE availability data to optimally match them to maximize social welfare.
Project Title: “3D Drone Delivery Transportation Problem”
Dr. Ignatius Formunung, Lead PI in collaboration with Dr. Mbakisya A. Onyango, Dr. Arash Ghasemi, and Dr. Joseph Owino
Abstract: In this work, we consider the realistic model of the three-dimensional motion of a self-controlled drone in a densely populated urban environment. The objective is to deliver packages from multiple points to the multiple destinations using a connected group of drones. The cityscape is first modeled accurately using a tetrahedral grid generated around the GIS data. This grid is then used to determine the connectivity of the destination points. A recent algorithm developed by our team will be utilized to find the minimum routing for each drone. These routes are then corrected by incorporating the wind forces obtained using a computational fluid dynamic (CFD) solver. The idea is to weight the graph such that the drones travel in the wakes of the buildings to have minimum fuel (energy) consumption. We utilize our CFD solver to achieve this goal. Also, we use a 6DOF model for drone and aerodynamic forces obtained by the CFD solver. A simple PID controller is used in each drone to augment the path. The results have vital applications in military data collection by flying spy drones, optimized package delivery using drones, and smart and futuristic cities (where cars can fly!).
Project Title: “Estimating the Youden Index under the Multivariate ROC Curve in the Presence of Missing Values of Mass Diseased and Healthy Biomarker Data”
Dr. Sumith Gunasekera, Lead PI in collaboration with Dr. L. Weerasena, Dr. H. Qin, and Mr. Aruna Saram
Abstract: In the context of Binary classification, Receiver Operating Characteristic curves have played an important role in classifying individuals/objects into one of the two predefined classes/populations. These procedures explain how to estimate the Youden index that measures the accuracy of a diagnostic test. However, problem arises when data contains missing values. The proposed research demonstrates how the Youden index for the diseased and healthy subjects can be extended to multi-biomarkers in the higher-dimensional space by analytic and extensive computational continuation of the mass missing data of multi-biomarkers from breast cancer and by intensive and extensive computations of the simulated mass data with the aid of generalized variable method. This computational-extensive mass-data-based procedure is novel and reduces the high number of unnecessary breast biopsies by helping physicians in their decision to perform a breast biopsy on a suspicious lesion seen in a mammogram or to perform a short-term follow-up examination instead. This goal is accomplished by the comparison of classical and generalized variable procedures for the multivariate Youden Index for the multi-biomarkers with missing data, where missing data are cleaned or tackled with the aid of imputation using parallel programing procedures in machine learning.
Project Title: “STC3: A Smart Trust-based Connected Autonomous Collaborative Communities”
Dr. Farah Kandah, Lead PI
Abstract: Connected autonomous vehicles (CAV) are among the key components contributing to Smart City initiatives. Besides communication protocols, securing the network and establishing trust between network entities are among the main challenges that need to be addressed in the field. Securing the network against outsiders’ attacks—trying to bypass the authentication scheme—as well as insiders’ attacks—trying to pollute the network with forged information—are essences to be addressed. Thus, there is both a critical and urgent need to design, prototype, validate, and demonstrate an integrated, real-time system that is better able to ensure the safety of the system by identifying, reporting, and isolating suspicious activities that require immediate attention. In the absence of such information, comprehensive prevention of trust attacks will be impossible, threatening human lives and inhibiting the further development and expansion of the connected and autonomous vehicle industry. The PIs at the University of Tennessee at Chattanooga (UTC) are uniquely qualified to address the proposed research. Prior work by the team has produced significant early findings that enabled the PIs to design and prototype an effective system. Specific strengths in software-defined networking (SDN) and mmWave enables the team to introduce those concepts as key to improving the proposed trust approach.
Project Title: “Using Computational Tools to Understand the Fundamental Rules of Life”
Dr. Hope Klug, Lead PI in collaboration with Dr. Jennifer Boyd, Dr. Azad Hossain and Dr. Hong Qin
Abstract: A fundamental goal in biology is to understand the diversity of life in relation to interactions among organisms and their environment. Most biological studies thus far have involved the analysis of relatively small data sets. To understand diversity on a large scale, we need to shift our focus to the analysis of large datasets. To address the question of why we see striking variation in living organisms, we will use big data and cutting-edge computational tools to: 1) enhance our understanding of biological robustness by examining gene/protein interaction networks; 2) explore the factors that make some species rare and other species common; and 3) investigate how abiotic and biotic factors drive the evolution of individual-level traits. In all cases, we will evaluate species network configurations using environmental fluctuations across spatial and temporal scales.
Project Title: “Modeling Online Social Network Dynamics and Predicting Information Diffusion with Fractional Differential Equations”
Dr. Lingju Kong, Lead PI in collaboration with Dr. John R. Graef, and Dr. Andrew Ledoan
Abstract: The use of social media has been spreading at an accelerated rate in the last decade. Today, there are many social media platforms such as blogs and social network sites. While the dynamics of online social networks have been studied using several models formulated via classical derivatives, these models are local, fail to capture the memory of the system, and have some other deficiencies. The aim of the proposed project is to improve on these studies by utilizing the theory of fractional calculus. Two new dynamic mathematical models based on fractional calculus will be proposed to serve as effective tools for analyzing the mechanisms of online social networks. More precisely, the investigators will first use fractional ordinary differential equations to construct a model to better understand the adoption and abandonment of a social network. Next, they will employ a fractional partial differential equation to model the spatial and temporal characteristics of information diffusion. These models will be compared with real datasets from selected networks. Various model properties such as existence, uniqueness, and stability of solutions will be investigated. Moreover, extensive numerical simulations will be performed to facilitate the analysis and refinement of these models.
Project Title: “Ionizing Radiation Effects Spectroscopy for Secure Space and Defense Communications”
Dr. T. Daniel Loveless, Lead PI in collaboration with Dr. Donald R. Reising
Abstract: Process-induced variability and device-level reliability have been identified as bottlenecks to system reliability, introducing a stochastic nature chip functionality. This disruption necessitates (1) new techniques for measurement of stochastic time-dependent defects; (2) a framework for understanding the dominant device-level reliability failure mechanisms in emerging and disruptive technologies for higher-fidelity predictions of lifetime; and (3) a fundamental understanding of the interplay of variability, operational constraints, and device performace for development of future electronics infrastructure with clear applications in Internet-of-Things and Space and Defense systems. These goals will be accomplished through the integration of computational modeling techniques and experimental measurements. We will (1) perform time-dependent defect measurements on advanced FinFET devices; (2) develop stochastic-based models that describe the reliability failure mechanisms and compact models of the time-dependent defects for integration into device and circuit simulators; and (3) provide a novel tool, Ionizing Radiation Effects Spectroscopy (IRES), for measuring the impact of such effects in operational communications systems in situ. This work will offer a fundamentally new approach to evaluating system-level reliability vulnerabilities and has the potential for transforming the way industry assesses electronic device, component, and system reliability.
Project Title: “Investigating the Flow of Nanodrugs through Bio-Inspired Hydrogel Channels”
Dr. Soubantika Palchoudhury, Lead PI in collaboration with Dr. Abdollah (Abi) Arabshahi
Abstract: Nanodrugs are highly attractive for next-generation medicine because they can be selectively targeted to diseased sites, provide diagnostic capability, and show better solubility compared to conventional therapeutics. However, their transport properties and accumulation within the body are largely unknown, due to experimental challenges in imaging the nanodrugs in complex medium. Recently, we developed a combined experimental and computational fluid dynamic approach at UTC to predict the velocity of a new Pt-iron oxide nanodrug through channels of different shapes. In this project, we aim to answer the fundamental question about transport behavior of the nanodrug through custom-designed channels made of materials that closely mimic bronchial airway. The channels will be experimentally developed through two novel approaches: 3D bioprinting and growing different hydrogels within the channel walls. We will develop a computational fluid dynamic model to predict the flow of nanodrug through these bio-inspired channels for the first time in-house at UTC. The proposed project will have two major outcomes. The computational fluid dynamic model will be a significant breakthrough in drug development and delivery, and using bio-inspired engineering to develop the flow path for nanodrugs will be a key experimental achievement. The project will be used to develop external proposals and publications.
Project Title: “Analyzing Bioimage Big Data with Deep Learning Neural Networks”
Dr. Hong Qin, Lead PI in collaboration with Dr. Joey Shaw, Dr. Yu Liang and Dr. Craig Tanis
Abstract: Our goal is to develop state-of-the-art deep convolutional neural networks models (CNNs) to transform two fields of biological research: cellular aging and plant species identification. For cellular aging, we plan to first develop supervised machine learning methods to cluster and label microscopic image for dividing yeast cells. We will then use these labeled images to train CNNs to automatically infer cell division events. For plant species identification, we plan to develop two CNN models and apply them sequentially: the first model will identify plant object regions from herbarium sheets, and the second model will use these objects to classify plant samples into meaningful clusters. Our proposed research will significantly advance the current bioimage big data analytics in these two fields.
Project Title: “Improving Post-Stroke Management Efficiency and Patient Outcomes through Analytics”
Dr. Mina Sartipi, Lead PI in collaboration with Dr. Nancy Fell
Abstract: For this CEACSE research project, our multidisciplinary team of academic researchers from the Computer Science and Engineering and Physical Therapy Departments will work together to develop a data-driven precision healthcare ecosystem for the management of stroke, the leading cause of long-term disability in the United States. This problem also aligns with the recently launched “big data to knowledge” initiative by NIH. Large-scale multi-modal heterogeneous data and big data analytics are the body and soul of the proposed research, respectively. Data preprocessing, predictive modeling, and prescriptive analytics will be explored and exploited to close the loop of big data analytics for precision healthcare. The computationally intensive concepts, models, algorithms, and functions will be designed and developed to transfer rich data to knowledge—and further to personalized decision support. The proposed inter-professional research will benefit both academic and healthcare communities.
Project Title: “Urban Resilience in the Post-Evacuation Age: Combining CFD and ABM for Megacities”
Dr. Kidambi Sreenivas, Lead PI in collaboration with Dr. Abdollah Arabshahi and Dr. Ethan Hereth
Abstract: The overarching goal of the proposed project is to reconstitute the capability (at the SimCenter) to carry out city-scale simulations such that evacuation planning can be carried out. This work will be carried out in collaboration with Dr. Epstein from NYU. The simulations will be carried out using technology developed at the SimCenter, while the agent-based models (ABM) will use agents developed by Dr. Epstein. Upon successful completion, results from this project will be used for an article that is to appear in Science. This approach of coupling computational fluid dynamics (CFD) and ABM has applications beyond the proposed project and can be used, for example, to track the spread of pandemics, etc.
Project Title: “Waterborne Infections and Pathogen Dynamics: Modeling, Experimentation, and Large-Scale Computation”
Dr. Jin Wang, Lead PI in collaboration with Dr. David Giles and Dr. Bradley Harris
Abstract: Waterborne infectious diseases remain a significant public heath burden worldwide. In particular, cholera, a severe intestinal infection caused by virulent strains of the bacterium Vibrio cholerae, has expanded in Africa and South Asia and re-emerged in the Americas in recent years as a serious health threat, with an estimated 2-4 million of cases per year reported by the World Health Organization. Effective outbreak response and control strategies for waterborne diseases rely on a deep understanding of the pathogen dynamics in reference to the epidemiologic triad of agent, host, and environment. The proposed research aims to establish a new mathematical and computational framework to investigate the pathogen dynamics related to waterborne infections, with a focus on cholera, and to make new discoveries regarding disease transmission and pathogen evolution. The project will combine mathematical models, biological experiments, and advanced numerical methods, with an emphasis on large-scale computation for model implementation and realistic application. The project belongs to the Health/Biomedical priority area.
Project Title: “Modeling Fate and Transport of Engineered Nanomaterial in Surface Water Systems”
Dr. Weidong Wu, Lead PI in collaboration with Dr. Jejal Reddy Bathi and Dr. Robert Webster
Abstract: Unique properties of engineered nanomaterials (ENM) have resulted in their increased production. However, it is unclear how these emerging ENM will move and react once released to the environment. One approach for addressing possible exposure of ENM in surface waters is by using numerical, mechanistic fate and transport models. There are no reliable fate models currently available that have the ability to simulate ENM behavior in the environment. Our proposed research will explore capabilities of the Environmental Fluid Dynamic Code (EFDC) model, originally developed by the U.S. Environmental Protection Agency (EPA) for simulating hydrodynamics of surface waters, for simulating ENM. We will examine the model algorithms to address the processes governing ENM in aqueous media. Since the literature pertaining to type and quantity of ENM in surface water environment is limited, as the first phase of the proposed research, a systematic evaluation of available literature to identify expected ENM and their physical, chemical, and biological properties that are important in pollutants fate assessment will be conducted. Second and third phases of the proposed research will include development of a calibrated EFDC model for a river hydraulics and ENM fate simulation under varied scenarios of changed river flows and pollutant loads.