{"618869":{"#nid":"618869","#data":{"type":"event","title":"PhD Defense by Andrea Garbo","body":[{"value":"\u003Cp\u003EPh.D. Thesis Defense\u003Cbr \/\u003E\r\nby\u003Cbr \/\u003E\r\nAndrea Garbo\u003Cbr \/\u003E\r\n(Advisor: Dr. Brian J. German)\u003Cbr \/\u003E\r\n2 PM, Friday, March 8th, 2018\u003Cbr \/\u003E\r\nWeber building, CoDE room\u003Cbr \/\u003E\r\nA Sequential Adaptive Sampling Technique Based on Local Linear Model for Computer\u003Cbr \/\u003E\r\nExperiment Applications\u003Cbr \/\u003E\r\nABSTACT:\u003Cbr \/\u003E\r\nThe objective of this dissertation research is to develop a model independent sequential adaptive sampling\u003Cbr \/\u003E\r\ntechnique for surrogate model (SM) applications based on a local linear model. This technique, called Nearest\u003Cbr \/\u003E\r\nNeighbors Adaptive Sampling (NNAS), is conceived to be conceptually simple, computationally robust, and\u003Cbr \/\u003E\r\neasy to apply, all characteristics that are crucial for effective surrogate modeling application during early phases\u003Cbr \/\u003E\r\nof the engineering design process. SMs are now regarded as powerful engineering tools for the approximation of\u003Cbr \/\u003E\r\nexpensive responses \u0026ndash; obtained either from computer simulations or real experiments \u0026ndash; via less computationally\u003Cbr \/\u003E\r\nexpensive mathematical models. The use of SMs is especially valuable during the preliminary design phase\u003Cbr \/\u003E\r\nwhen engineers need fast and accurate tools to assess the performance of different configurations and to define\u003Cbr \/\u003E\r\nthe top-level specifications that will guide the entire design process. Due to the increasing importance of SMs,\u003Cbr \/\u003E\r\nnew strategies are continuously being devised to build more flexible SM formulations, to rigorously select an\u003Cbr \/\u003E\r\nSM technique from a set of candidates, and to efficiently sample the design space to collect the data required to\u003Cbr \/\u003E\r\ntrain an SM.\u003Cbr \/\u003E\r\nThe considerable influence of the sample distribution on SM accuracy motivates efforts to develop\u003Cbr \/\u003E\r\nadvanced strategies to improve the sampling process. In particular, the adoption of sequential adaptive sampling\u003Cbr \/\u003E\r\ntechniques has been empirically shown to reduce the number of samples required to obtain an SM of specified\u003Cbr \/\u003E\r\naccuracy. However, these techniques are typically challenging to implement, limited by assumptions about the\u003Cbr \/\u003E\r\nresponse, and dependent on the SM formulation selected to supervise the sampling process (e.g. cross validation\u003Cbr \/\u003E\r\nand Kriging based strategies), making them impracticable for most engineering design applications . In\u003Cbr \/\u003E\r\nparticular, model dependence \u0026ndash; a common characteristic of most state-of-the-art adaptive sequential sampling\u003Cbr \/\u003E\r\ntechniques \u0026ndash; may decrease the sampling efficiency if the guiding SM is inappropriately chosen.\u003Cbr \/\u003E\r\nThe proposed NNAS technique avoids the limitations of model dependence by introducing a new\u003Cbr \/\u003E\r\nrefinement metric \u0026ndash; the Non Linearity Index (NLI) \u0026ndash; which estimates the local nonlinear characteristics as the\u003Cbr \/\u003E\r\ndifference between the actual response value f(xT,i) and the local function approximation represented by the\u003Cbr \/\u003E\r\nhyperplane obtained via Weighted Least Squares Regression of the closest D+k points in the neighborhood of\u003Cbr \/\u003E\r\nxT,i, where D is the domain dimensionality. The use of local linear models to assess the nonlinear characteristics\u003Cbr \/\u003E\r\nof the response without the need for a global SM is the key characteristic of NNAS that makes this strategy\u003Cbr \/\u003E\r\nmodel independent. Additionally, NNAS introduces a new stochastic Pareto-ranking-based selection criterion to\u003Cbr \/\u003E\r\nsimultaneously maximize the refinement and exploration of the design space search, thereby ensuring a balance\u003Cbr \/\u003E\r\nbetween the two behaviors. The initial NNAS and NLI formulations have also been expanded to include a form\u003Cbr \/\u003E\r\nof directional sampling in which the algorithm identifies both region and direction of sampling.\u003Cbr \/\u003E\r\nNNAS embodies the capabilities of sampling multi-response design spaces, working in batch-mode (i.e.\u003Cbr \/\u003E\r\nadding more than one sample at time), and continuing the sampling process even in the event of a critical error\u003Cbr \/\u003E\r\nin the f evaluation, e.g. the lack of convergence of a computational model at points in the design space. These\u003Cbr \/\u003E\r\ncharacteristics together with its ease of implementation make NNAS a valuable, efficient and robust sampling\u003Cbr \/\u003E\r\nstrategy to use during the early phases of engineering design.\u003Cbr \/\u003E\r\nCommittee:\u003Cbr \/\u003E\r\nDr. Brian J. German, School of Aerospace Engineering, Georgia Institute of Technology\u003Cbr \/\u003E\r\nDr. Graeme J. Kennedy, School of Aerospace Engineering, Georgia Institute of Technology\u003Cbr \/\u003E\r\nDr. E. Glenn Lightsey, School of Aerospace Engineering, Georgia Institute of Technology\u003Cbr \/\u003E\r\nDr Dimitri N. Mavris, School of Aerospace Engineering, Georgia Institute of Technology\u003Cbr \/\u003E\r\nDr. Daniel W. Apley, School of Engineering and Applied Science, Northwestern University\u003C\/p\u003E\r\n","summary":null,"format":"limited_html"}],"field_subtitle":"","field_summary":"","field_summary_sentence":[{"value":"A Sequential Adaptive Sampling Technique Based on Local Linear Model for Computer Experiment Applications"}],"uid":"27707","created_gmt":"2019-03-06 13:58:16","changed_gmt":"2019-03-06 13:58:38","author":"Tatianna Richardson","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2019-03-08T14:00:00-05:00","event_time_end":"2019-03-08T16:00:00-05:00","event_time_end_last":"2019-03-08T16:00:00-05:00","gmt_time_start":"2019-03-08 19:00:00","gmt_time_end":"2019-03-08 21:00:00","gmt_time_end_last":"2019-03-08 21:00:00","rrule":null,"timezone":"America\/New_York"},"extras":[],"groups":[{"id":"221981","name":"Graduate Studies"}],"categories":[],"keywords":[{"id":"100811","name":"Phd Defense"}],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"1788","name":"Other\/Miscellaneous"}],"invited_audience":[{"id":"78761","name":"Faculty\/Staff"},{"id":"78771","name":"Public"},{"id":"174045","name":"Graduate students"},{"id":"78751","name":"Undergraduate students"}],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[],"email":[],"slides":[],"orientation":[],"userdata":""}}}