DEVELOPMENT OF AN OPTIMIZATION MODEL FOR BIOFUEL FACILITY SIZE AND LOCATION AND A SIMULATION MODEL FOR DESIGN OF A BIOFUEL SUPPLY CHAIN

A range of societal issues have been caused by fossil fuel consumption in the transportation sector in the United States (U.S.), including health related air pollution, climate change, the dependence on imported oil, and other oil related national security concerns. Biofuels production from various lignocellulosic biomass types such as wood, forest residues, and agriculture residues have the potential to replace a substantial portion of the total fossil fuel consumption. This research focuses on locating biofuel facilities and designing the biofuel supply chain to minimize the overall cost. For this purpose an integrated methodology was proposed by combining the GIS technology with simulation and optimization modeling methods. The GIS based methodology was used as a precursor for selecting biofuel facility locations by employing a series of decision factors. The resulted candidate sites for biofuel production served as inputs for simulation and optimization modeling. As a precursor to simulation or optimization modeling, the GIS-based methodology was used to preselect potential biofuel facility locations for biofuel production from forest biomass. Candidate locations were selected based on a set of evaluation criteria, including: county boundaries, a railroad transportation network, a state/federal road transportation network, water body (rivers, lakes, etc.) dispersion, city and village dispersion, a population census, biomass production, and no co-location with co-fired power plants. The simulation and optimization models were built around key supply activities including biomass harvesting/forwarding, transportation and storage. The built onsite storage served for spring breakup period where road restrictions were in place and truck transportation on certain roads was limited. Both models were evaluated using multiple performance indicators, including cost (consisting of the delivered feedstock cost, and inventory holding cost), energy consumption, and GHG emissions. The impact of energy consumption and GHG emissions were expressed in monetary terms to keep consistent with cost. Compared with the optimization model, the simulation model represents a more dynamic look at a 20-year operation by considering the impacts associated with building inventory at the biorefinery to address the limited availability of biomass feedstock during the spring breakup period. The number of trucks required per day was estimated and the inventory level all year around was tracked. Through the exchange of information across different procedures (harvesting, transportation, and biomass feedstock processing procedures), a smooth flow of biomass from harvesting areas to a biofuel facility was implemented. The optimization model was developed to address issues related to locating multiple biofuel facilities simultaneously. The size of the potential biofuel facility is set up with an upper bound of 50 MGY and a lower bound of 30 MGY. The optimization model is a static, Mathematical Programming Language (MPL)-based application which allows for sensitivity analysis by changing inputs to evaluate different scenarios. It was found that annual biofuel demand and biomass availability impacts the optimal results of biofuel facility locations and sizes.

AGENT BASED, DISTRIBUTED CONTROL STRATEGIES AND OPTIMIZATION OF PLUG-IN ELECTRIC VEHICLES ON SMART/MICROGRID ARCHITECTURES

This thesis will present strategies for the use of plug-in electric vehicles on smart and microgrids. MATLAB is used as the design tool for all models and simulations. First, a scenario will be explored using the dispatchable loads of electric vehicles to stabilize a microgrid with a high penetration of renewable power generation. Grid components for a microgrid with 50% photovoltaic solar production will be sized through an optimization routine to maintain storage system, load, and vehicle states over a 24-hour period. The findings of this portion are that the dispatchable loads can be used to guard against unpredictable losses in renewable generation output. Second, the use of distributed control strategies for the charging of electric vehicles utilizing an agent-based approach on a smart grid will be studied. The vehicles are regarded as additional loads to a primary forecasted load and use information transfer with the grid to make their charging decisions. Three lightweight control strategies and their effects on the power grid will be presented. The findings are that the charging behavior and peak loads on the grid can be reduced through the use of distributed control strategies.

A computer-aided diagnostic system to characterize CT focal liver lesions: design and optimization of a neural network classifier

In this paper, a computer-aided diagnostic (CAD) system for the classification of hepatic lesions from computed tomography (CT) images is presented. Regions of interest (ROIs) taken from nonenhanced CT images of normal liver, hepatic cysts, hemangiomas, and hepatocellular carcinomas have been used as input to the system. The proposed system consists of two modules: the feature extraction and the classification modules. The feature extraction module calculates the average gray level and 48 texture characteristics, which are derived from the spatial gray-level co-occurrence matrices, obtained from the ROIs. The classifier module consists of three sequentially placed feed-forward neural networks (NNs). The first NN classifies into normal or pathological liver regions. The pathological liver regions are characterized by the second NN as cyst or "other disease." The third NN classifies "other disease" into hemangioma or hepatocellular carcinoma. Three feature selection techniques have been applied to each individual NN: the sequential forward selection, the sequential floating forward selection, and a genetic algorithm for feature selection. The comparative study of the above dimensionality reduction methods shows that genetic algorithms result in lower dimension feature vectors and improved classification performance.

Characterization and optimization of antigen-specific T cell responses during ex vivo expansion of melanoma tumor-infiltrating lymphocytes

Treatment of metastatic melanoma with tumor reactive T cells (adoptive T cell therapy, ACT) is a promising approach associated with a high clinical response rate. However, further optimization of this treatment modality is required to increase the clinical response after this therapy. ACT in melanoma involves an initial phase (pre-REP) of tumor-infiltrating lymphocyte (TIL) expansion ex vivo from tumor isolates followed by a second phase, “rapid expansion protocol” (REP) generating the billions of cells used as the TIL infusion product. The main question addressed in this thesis was how the currently used REP affected the responsiveness of the CD8+ T cells to defined melanoma antigens. We hypothesized that the REP drives the TIL to further differentiate and become hyporesponsive to antigen restimulation, therefore, proper cytokine treatment or other ways to expand TIL is required to improve upon this outcome. We evaluated the response of CD8+ TIL to melanoma antigen restimulation using MART-1 peptide-pulsed mature DC in vitro. Post-REP TILs were mostly hypo-responsive with poor proliferation and higher apoptosis. Phenotypic analysis revealed that the expression of CD28 was significantly reduced in post-REP TILs. By sorting experiment and microarray analysis, we confirmed that the few CD28+ post-REP TILs had superior survival capacity and proliferated after restimulation. We then went on to investigate methods to maintain CD28 expression during the REP and improve TIL responsiveness. Firstly, IL-15 and IL-21 were found to synergize in maintaining TIL CD28 expression and antigenic responsiveness during REP. Secondly, we found IL-15 was superior as compared to IL-2 in supporting the long-term expansion of antigen-specific CD8+ TIL after restimulation. These results suggest that current expansion protocols used for adoptive T-cell therapy in melanoma yield largely hyporesponsive products containing CD8+ T cells unable to respond in vivo to re-stimulation with antigen. A modification of our current approaches by using IL-15+IL-21 as supporting cytokines in the REP, or/and administration of IL-15 instead of IL-2 after TIL infusion, may enhance the anti-tumor efficacy and long-term persistence of infused T cells in vivo.

Preparation and optimization of calcium fluoride particles for dental applications.

Fluorides are used in dental care due to their beneficial effect in tooth enamel de-/remineralization cycles. To achieve a desired constant supply of soluble fluorides in the oral cavity, different approaches have been followed. Here we present results on the preparation of CaF2 particles and their characterization with respect to a potential application as enamel associated fluoride releasing reservoirs. CaF2 particles were synthesized by precipitation from soluble NaF and CaCl2 salt solutions of defined concentrations and their morphology analyzed by scanning electron microscopy. CaF2 particles with defined sizes and shapes could be synthesized by adjusting the concentrations of the precursor salt solutions. Such particles interacted with enamel surfaces when applied at fluoride concentrations correlating to typical dental care products. Fluoride release from the synthesized CaF2 particles was observed to be largely influenced by the concentration of phosphate in the solution. Physiological solutions with phosphate concentration similar to saliva (3.5 mM) reduced the fluoride release from pure CaF2 particles by a factor of 10-20 × as compared to phosphate free buffer solutions. Fluoride release was even lower in human saliva. The fluoride release could be increased by the addition of phosphate in substoichiometric amounts during CaF2 particle synthesis. The presented results demonstrate that the morphology and fluoride release characteristics of CaF2 particles can be tuned and provide evidence of the suitability of synthetic CaF2 particles as enamel associated fluoride reservoirs.

Regulating a Benzodifuran Single Molecule Redox Switch via Electrochemical Gating and Optimization of Molecule/Electrode Coupling

We report a novel strategy for the regulation of charge transport through single molecule junctions via the combination of external stimuli of electrode potential, internal modulation of molecular structures, and optimization of anchoring groups. We have designed redox-active benzodifuran (BDF) compounds as functional electronic units to fabricate metal–molecule–metal (m–M–m) junction devices by scanning tunneling microscopy (STM) and mechanically controllable break junctions (MCBJ). The conductance of thiol-terminated BDF can be tuned by changing the electrode potentials showing clearly an off/on/off single molecule redox switching effect. To optimize the response, a BDF molecule tailored with carbodithioate (−CS2–) anchoring groups was synthesized. Our studies show that replacement of thiol by carbodithioate not only enhances the junction conductance but also substantially improves the switching effect by enhancing the on/off ratio from 2.5 to 8.

Computer assisted planning and navigation of periacetabular osteotomy with range of motion optimization

Femoroacetabular impingement (FAI) before or after Periacetabular Osteotomy (PAO) is surprisingly frequent and surgeons need to be aware of the risk preoperatively and be able to avoid it intraoperatively. In this paper we present a novel computer assisted planning and navigation system for PAO with impingement analysis and range of motion (ROM) optimization. Our system starts with a fully automatic detection of the acetabular rim, which allows for quantifying the acetabular morphology with parameters such as acetabular version, inclination and femoral head coverage ratio for a computer assisted diagnosis and planning. The planned situation was optimized with impingement simulation by balancing acetabuar coverage with ROM. Intra-operatively navigation was conducted until the optimized planning situation was achieved. Our experimental results demonstrated: 1) The fully automated acetabular rim detection was validated with accuracy 1.1 ± 0.7mm; 2) The optimized PAO planning improved ROM significantly compared to that without ROM optimization; 3) By comparing the pre-operatively planned situation and the intra-operatively achieved situation, sub-degree accuracy was achieved for all directions.

Global Optimization with Sparse and Local Gaussian Process Models

We present a novel surrogate model-based global optimization framework allowing a large number of function evaluations. The method, called SpLEGO, is based on a multi-scale expected improvement (EI) framework relying on both sparse and local Gaussian process (GP) models. First, a bi-objective approach relying on a global sparse GP model is used to determine potential next sampling regions. Local GP models are then constructed within each selected region. The method subsequently employs the standard expected improvement criterion to deal with the exploration-exploitation trade-off within selected local models, leading to a decision on where to perform the next function evaluation(s). The potential of our approach is demonstrated using the so-called Sparse Pseudo-input GP as a global model. The algorithm is tested on four benchmark problems, whose number of starting points ranges from 102 to 104. Our results show that SpLEGO is effective and capable of solving problems with large number of starting points, and it even provides significant advantages when compared with state-of-the-art EI algorithms.

Generation of an hTERT-restricted, CMV-augmented oncolytic adenovirus capable of tumor-specific transgene expression: Development, characterization, and optimization

Current shortcomings in cancer therapy require the generation of new, broadly applicable, potent, targeted treatments. Here, an adenovirus is engineered to replicate specifically in cells with active human telomerase promotion using a modified hTERT promoter, fused to a CMV promoter element. The virus was also modified to contain a visible reporter transgene, GFP. The virus, Ad/hTC-GFP-E1 was characterized in vitro and demonstrated tumor specific activity both by dose and over time course experiments in a variety of cell lines. In vivo, Ad/hTC-GFP-E1 was affected at suppressing tumor growth and providing a survival benefit without causing any measurable toxicity. To increase the host range of the vector, the fiber region was modified to contain an RGD-motif. The vector, AdRGD/hTC-GFP-E1, was recharacterized in vitro, revealing heightened levels of infectivity and toxicity however maintaining a therapeutic window between cancer and normal cell toxicity. AdRGD/hTC-GFP-E1 was administered in vivo by limb perfusion and was observed to be tumor specific both in expression and replication. To further enhance the efficacy of viral vectors in lung delivery, asthma medications were investigated for their abilities to enhance transgene delivery and expression. A combination of bronchodilators, mast cell inhibitors, and mucolytic agents was devised which demonstrated fold increases in expression in immunocompetent mouse lungs as single agents and more homogenous, intense levels of expression when done in combination of all agents. To characterize the methods in which some cancers are resistant or may become resistant to oncolytic treatments, several small molecule inhibitors of metabolic pathways were applied in combination with oncolytic infection in vitro. SP600125 and PD 98059, respective JNK and ERK inhibitors, successfully suppressed oncolytic toxicity, however did not affect infectivity or transgene expression of Ad/hTC-GFP-E1. JNK and ERK inhibition did significantly suppress viral replication, however, as analyzed by lysate transfer and titration assays. In contrast, SB 203580, an inhibitor for p38, did not demonstrate any protective effects with infected cells. Flow cytometric analysis indicated a possible correlation with G1 arrest and suppressed viral production, however more compounds must be investigated to clarify this observation. ^

Design and Optimization of Four-dimensional Cone-beam Computed Tomography in Image-guided Radiation Therapy

The influence of respiratory motion on patient anatomy poses a challenge to accurate radiation therapy, especially in lung cancer treatment. Modern radiation therapy planning uses models of tumor respiratory motion to account for target motion in targeting. The tumor motion model can be verified on a per-treatment session basis with four-dimensional cone-beam computed tomography (4D-CBCT), which acquires an image set of the dynamic target throughout the respiratory cycle during the therapy session. 4D-CBCT is undersampled if the scan time is too short. However, short scan time is desirable in clinical practice to reduce patient setup time. This dissertation presents the design and optimization of 4D-CBCT to reduce the impact of undersampling artifacts with short scan times. This work measures the impact of undersampling artifacts on the accuracy of target motion measurement under different sampling conditions and for various object sizes and motions. The results provide a minimum scan time such that the target tracking error is less than a specified tolerance. This work also presents new image reconstruction algorithms for reducing undersampling artifacts in undersampled datasets by taking advantage of the assumption that the relevant motion of interest is contained within a volume-of-interest (VOI). It is shown that the VOI-based reconstruction provides more accurate image intensity than standard reconstruction. The VOI-based reconstruction produced 43% fewer least-squares error inside the VOI and 84% fewer error throughout the image in a study designed to simulate target motion. The VOI-based reconstruction approach can reduce acquisition time and improve image quality in 4D-CBCT.