BEGIN:VCALENDAR CALSCALE:GREGORIAN METHOD:PUBLISH PRODID:-//appliedmath.arizona.edu//NONSGML iCalcreator 2.2// VERSION:2.0 X-WR-CALNAME:Modeling and Computation Seminar X-WR-CALDESC:Modeling and Computation Seminar Calendar, University of Arizo na Program in Applied Mathematics X-WR-TIMEZONE:US/Arizona BEGIN:VTIMEZONE TZID:US/Arizona BEGIN:STANDARD DTSTART:19700101T000000 TZOFFSETFROM:-0700 TZOFFSETTO:-0700 TZNAME:MST END:STANDARD END:VTIMEZONE BEGIN:VEVENT CATEGORIES:Modeling & Computation DESCRIPTION:Otoacoustic emissions (OAEs) are sounds generated and subsequen tly emitted by a healthy ear (detectable using a sensitive microphone) whi ch appear in a wide range of vertebrate species. While the exact generatio n mechanisms remain unclear\, OAEs evoked using an external stimulus exhib it significant group delays across a wide frequency range\, on the order o f 1-2 ms or greater. In mammals such as humans\, these delays are generall y thought to arise due to the presence of cochlear traveling waves. Howeve r\, in classes such as lizards\, such waves are noticeably absent. The pre sent study hypothesizes that these delays are in fact associated with the sharp tuning manifested in the auditory periphery and represent the build- up time of highly tuned coupled oscillators. Preliminary model results for the gecko ear show remarkable agreement with empirical data and predict t hat emission group delays increase with increasing sharpness of tuning (as typically measured via auditory nerve fiber responses). DTSTAMP:20080426T232023Z DTSTART;TZID=US/Arizona:20080424T123000 DURATION:PT1H LOCATION:Math 402 SUMMARY:Christopher Bergevin: Modeling Otoacoustic Emission Group Delays in the Lizard Auditory Periphery UID:20080426T012023CEST-T1rkBUC5PJ@appliedmath.arizona.edu END:VEVENT BEGIN:VEVENT CATEGORIES:Modeling & Computation DESCRIPTION:Particle Filters are sample-based numerical methods for the dis crete-time Filtering Problem. These methods suffer from large operations c ount and troubles defining prediction. This work introduces a particle fil ter method for the discrete-time Filtering Problem with SODEs\, along with a suitable definition of prediction. The method\, to be called Diffusion Kernel Filter\, applies when the dynamics of the SODE develops few moments (i.e.\, is weakly nonlinear) on ``branches of prediction'' between the fi ltering times (which is expected to be the case in several applications in the geosciences) and is arrived at by a parametrization of small fluctuat ions of Wiener-driven paths about deterministic paths and a local use of t his parametrization in the referential Bootstrap Filter.\n\nThe parametriz ation is derived by reformulation of the SODE problem into a Liouville SPD E problem\, application of Duhamel's principle to this problem\, restricti on of the resulting to nonlinear SODE open problems for the flows of ``bra nches of prediction\,'' closure of these. This was inspired by Chorin's `` Optimal prediction with memory\,'' where a similar technique is used to ta ckle the dimension reduction problem for the dynamics of a nonlinear ODE. Results obtained with the early Lorenz equations and a set of equations of point-vortex interactions are presented.\n DTSTAMP:20080426T232023Z DTSTART;TZID=US/Arizona:20080417T123000 DURATION:PT1H LOCATION:Math 402 SUMMARY:Paul Krause: The Diffusion Kernel Filter UID:20080426T012023CEST-jzvloGIfh6@appliedmath.arizona.edu END:VEVENT BEGIN:VEVENT CATEGORIES:Modeling & Computation DESCRIPTION:Red blood cell movement\, deformation\, and partitioning in sma ll diverging microvessel bifurcations are simulated using a two-dimensiona l\, flexible-particle model. For isolate red blood cell movement\, while s imulated red blood cell trajectories tend to follow background fluid strea mlines\, significant deviations from these streamlines can occur because o f red blood cell migration towards vessel centerlines and red blood cell o bstruction of downstream vessels. The net effect of these behaviors is exp lored in symmetric and asymmetric vessels to produce results comparable wi th experiment. In addition\, preliminary results and insights are presente d for multiple red blood cell motion in straight vessels and in bifurcatio ns. DTSTAMP:20080426T232023Z DTSTART;TZID=US/Arizona:20080410T123000 DURATION:PT1H LOCATION:Math 402 SUMMARY:Jared Barber: Two-Dimensional Model of Red Blood Cell Motion in Mic rovessels UID:20080426T012023CEST-FCtVC51PjE@appliedmath.arizona.edu END:VEVENT BEGIN:VEVENT CATEGORIES:Modeling & Computation DESCRIPTION:In this talk\, we first discuss simulation-based shop floor pla nning and control\, where 1) on-line simulation is used to evaluate decisi on alternatives at the planning stage\, 2) the same simulation model (exec uting in the fast mode) used at the planning stage is used as a real-time task generator (real-time simulation) during the control stage\, and 3) th e real-time simulation drives the manufacturing system by sending and rece iving messages to an executor (Finite State Automata). We then discuss how simulation-based shop floor planning and control can be extended to enter prise level activities (top floor). To this end\, we discuss the analogies between the shop floor and top floor in terms of the components required to construct simulation-based planning and control systems such as resourc e models\, coordination models\, physical entities\, and simulation models . Differences between them are also discussed in order to identify new cha llenges that we face for top floor planning and control. A major differenc e is the way a simulation model is constructed so that it can be used for planning\, depending on whether time synchronization among member simulati ons becomes an issue or not. We also discuss the distributed computing pla tform using web services and grid computing technologies\, which allow us to integrate simulation and decision models\, and software and hardware co mponents. Finally\, we discuss other emerging applications for the propose d simulation-based planning and control\, such as emergency evacuation and blood supply network. DTSTAMP:20080426T232023Z DTSTART;TZID=US/Arizona:20080403T123000 DURATION:PT1H LOCATION:Math 402 SUMMARY:Young-Jun Son: Distributed Federation of Multi-paradigm Simulations and Decision Models for Planning and Control: From Shop Floor to Top Floo r UID:20080426T012023CEST-amSt5j0OSS@appliedmath.arizona.edu END:VEVENT BEGIN:VEVENT CATEGORIES:Modeling & Computation DESCRIPTION:The remarkable functional versatility of proteins is made possi ble by the diverse array of three-dimensional folds that they adopt. The c onventional representation of protein structure is a discrete coordinate m odel listing the positions of all atoms in the structure. While this repre sentation is very useful for understanding intricate chemical details\, it is not well suited to addressing more general questions about the nature of protein folds\, their variability\, and the relationships between them. To investigate such questions\, we have developed a continuous representa tion of proteins based on the geometry of space curves. The description of a protein fold in terms of its underlying geometry has proved to be much more efficient than the coordinate representation\, suggesting that sparse experimental data may be sufficient to restrain a curve model where a con ventional coordinate model would be underdetermined. Many proteins are not amenable to high-resolution structural analysis\, and for these challengi ng cases it is important to make the best use of the limited experimental information available. The talk will describe the application of the curve representation to diffraction techniques focusing in particular on low-re solution X-ray crystallography.\n DTSTAMP:20080426T232023Z DTSTART;TZID=US/Arizona:20080327T123000 DURATION:PT1H LOCATION:Math 402 SUMMARY:Andrew Hausrath: Applications of Geometry in Protein Structure Anal ysis UID:20080426T012023CEST-HTbNlaUMDV@appliedmath.arizona.edu END:VEVENT BEGIN:VEVENT CATEGORIES:Modeling & Computation DESCRIPTION:The simulation-based design of nonlinear systems is hampered by several hurdles such as high CPU times\, the difficulty to evaluate gradi ents\, and the acute sensitivity of responses to loading and design uncert ainties. In addition\, the system's responses might be discontinuous due t o the presence of numerous limit and bifurcation points. This considerably limits the blind use of traditional optimization and probabilistic method s. Typical examples of problems with discontinuous behaviors are structura l impacts and nonlinear aeroelasticity with limit cycle oscillations (LCO) .\n\nThis seminar will describe a methodology which facilitates the probab ilistic (optimal) simulation-based design of nonlinear problems. The appro ach\, referred to as explicit design space decomposition\, is based on dat a mining and machine learning techniques. The main feature of this approac h lies in the explicit definition of limit state functions (or constraints ) constructed from a design of experiments (DOE). A method to adaptatively update the limit state function and refine the DOE will be presented. \n\ nSeveral test examples will demonstrate the efficiency of the approach in the case of the reliability-based optimization of nonlinear structures and LCO problems. DTSTAMP:20080426T232023Z DTSTART;TZID=US/Arizona:20080313T123000 DURATION:PT1H LOCATION:Math 402 SUMMARY:Samy Missoum: Probabilistic Optimal Design for Highly Nonlinear Pro blems UID:20080426T012023CEST-3GDStUID1T@appliedmath.arizona.edu END:VEVENT BEGIN:VEVENT CATEGORIES:Modeling & Computation DESCRIPTION:Tightly packed elastic structures can be found in a wide variet y of physical and biological systems. Traditionally mechanical and geometr ical aspects are treated separately due to the complex nature of the obser ved patterns (e.g.\, a piece of crumpled paper). We present a statistical field theory to study the packing of an elastic rod (1D) confined in 2D sp ace. An advantage of this approach is that it puts geometry and mechanics on an equal footing. We show that a self-reorganization of the rod becomes favorable at a critical density. This configurational phase transition (i sotropic-nematic) leads to a more efficient packing. For even higher confi nements we predict the existence of a jamming transition hinting at the gl assy character of this system. DTSTAMP:20080426T232023Z DTSTART;TZID=US/Arizona:20080306T123000 DURATION:PT1H LOCATION:Math 402 SUMMARY:Laurent Boue: A Statistical Physics Approach to Packing Problems UID:20080426T012023CEST-HsKjswU3dh@appliedmath.arizona.edu END:VEVENT BEGIN:VEVENT CATEGORIES:Modeling & Computation DESCRIPTION:We present a novel computational methodology aimed at overcomin g the aforementioned difficulties. At the heart of our approach are integr al equation formulations that exhibit excellent spectral properties. In th e case of scattering from perfectly conducting structures\, and just as th e classical Combined Field Integral Equation (CFIE)\, our equations result from representations of the scattered fields as a combination of magnetic - and electric-dipole distributions on the surface of the scatterer. In co ntrast with the classical equations\, however\, our electric-dipole operat ors involve use of certain types of regularizing operators whose design is based on the pseudodifferential calculus on manifolds. We call the result ing equations Regularized Combined Field Integral Equations (CFIE-R). Unli ke the CFIE\, the CFIE-R are well-conditioned equations\; careful selectio n of coupling parameters\, further\, yields CFIE-R operators with excellen t spectral distributions--with closely clustered eigenvalues--so that smal l numbers of iterations suffice to solve the corresponding equations by me ans of Krylov subspace iterative solvers such as GMRES. We present a high- order Nystrom approach based on use of partitions of unity and high-order integration schemes that produces high-order algorithms for acoustic and e lectromagnetic scattering problems. A variety of numerical results demonst rate that\, for a given accuracy\, the new equations can give rise to orde r-of-magnitude reductions in computational costs over those resulting from previous approaches. DTSTAMP:20080426T232023Z DTSTART;TZID=US/Arizona:20080228T123000 DURATION:PT1H LOCATION:Math 402 SUMMARY:Catalin Turc: Fast\, high-order\, well-conditioned algorithms for t he solution of three-dimensional acoustic and electromagnetic scattering p roblems UID:20080426T012023CEST-ReSc8ah20o@appliedmath.arizona.edu END:VEVENT BEGIN:VEVENT CATEGORIES:Modeling & Computation DESCRIPTION:I will present results from experiments on shaping of thin shee ts via lateral growth. Scaling of lengths scales and energies will be sugg ested and a possible geometrical origin of the appearance of small scale s tructure will be discussed. I will present first results from a study of g rowing leaves and will describe ongoing experiments. DTSTAMP:20080426T232023Z DTSTART;TZID=US/Arizona:20080221T123000 DURATION:PT1H LOCATION:Math 402 SUMMARY:Eran Sharon: Experiments in Non-Euclidean Plates and Leaves UID:20080426T012023CEST-V9WURDVCie@appliedmath.arizona.edu END:VEVENT BEGIN:VEVENT CATEGORIES:Modeling & Computation DESCRIPTION:By direct numerical simulation direct and inverse cascades in s urface gravity waves turbulence were observed. Formation of condensate in the low frequency waves region leads to the distortion of the exponents in Kolmogorov-like spectra predicted by the theory of weak turbulence. Also the influence of the wavenumbers grid discreteness was observed. The param eters of the simulation are typical for the laboratory water tanks experim ents. DTSTAMP:20080426T232023Z DTSTART;TZID=US/Arizona:20080131T123000 DURATION:PT1H LOCATION:Math 402 SUMMARY:Alexander O. Korotkevich: Simultaneous Observation of Direct and In verse Cascade in Surface Gravity Waves Turbulence UID:20080426T012023CEST-FTLoBTjtWT@appliedmath.arizona.edu END:VEVENT BEGIN:VEVENT CATEGORIES:Modeling & Computation DTSTAMP:20080426T232023Z DTSTART;TZID=US/Arizona:20080124T123000 DURATION:PT1H LOCATION:Math 402 SUMMARY:Organizational meeting UID:20080426T012023CEST-8k6Zcoj6v2@appliedmath.arizona.edu END:VEVENT BEGIN:VEVENT CATEGORIES:Modeling & Computation DESCRIPTION:We analyze the interaction of an electromagnetic spike (one cyc le) with a thin layer of ferroelectric medium with two equilibrium states. The model is the set of Maxwell equations coupled to the undamped Landau- Khalatnikov equation\, where we do not assume slowly varying envelopes. Fr om linear scattering theory\, we show that low amplitude pulses can be com pletely reflected by the medium. Large amplitude pulses can switch the fer roelectric. Using numerical simulations and analysis\, we study this switc hing for long and short pulses\, estimate the switching times and provide useful information for experiments. DTSTAMP:20080426T232023Z DTSTART;TZID=US/Arizona:20080103T123000 DURATION:PT1H LOCATION:Math 402 SUMMARY:Jean-Guy Caputo: Electromagnetically induced switching of ferroelec tric thin films UID:20080426T012023CEST-JtVWJBc1ej@appliedmath.arizona.edu END:VEVENT END:VCALENDAR