A Parameterized Mask Model for Lithography Simulation

Size: px
Start display at page:

Download "A Parameterized Mask Model for Lithography Simulation"

Transcription

1 A Parameterize Mask Moel for Lithography Simulation henhai hu Caence Research Labs, Berkeley, CA ABSTRACT e formulate the mask moeling as a parametric moel orer reuction problem base on the finite element iscretization of the Helmholtz equation. By using a new parametric mesh an a machine learning technique calle Kernel Metho, we convert the nonlinearly parameterize FEM matrices into affine forms. This allows the application of a well-unerstoo parametric reuction technique to generate compact mask moel. Since this moel is base on the first principle, it naturally inclues iffraction an couplings, important effects that are poorly hanle by the existing heuristic mask moels. Further more, the new mask moel offers the capability to make a smooth trae-off between accuracy an spee. Categories an Subject Descriptors B.7. Integrate Circuits: Design Aies Simulation General Terms Algorithms,Performance,Design Keywors Lithography, Mask Moel, Parameterize Moel Orer Reuction. INTRODUCTION From lithography simulation point of view, two classes of esigns pose particularly challenging problems. The first class is the memory esign. ith just six or seven transistors, each memory cell has small layout. The carefully esigne cell is uplicate hunre millions to billions of times. Since the manufacturing efects in one cell affect the entire row of cells, the memory esign is highly sensitive to such efects. Therefore, high accuracy lithography simulation is manatory for the memory esign. The secon class is the custom logic esign. ith a finite number of unique stanar cells as the builing blocks, the layout is typically large, in the orer of 1 1mm. Therefore, highly efficient lithography simulation is manatory for the custom logic esign. Since timing an power are the main concerns, the accuracy of lithography simulation has to be goo enough to moel the impact of manufacturing imperfection to the timing an power. Clearly, a goo lithography simulation tool shoul inherently have the capability to make smooth trae-off between accuracy an spee for various esigns. An it shoul be base on rigorous mathematical founation to ensure its robustness. Permission to make igital or har copies of all or part of this work for personal or classroom use is grante without fee provie that copies are not mae or istribute for profit or commercial avantage an that copies bear this notice an the full citation on the first page. To copy otherwise, to republish, to post on servers or to reistribute to lists, requires prior specific permission an/or a fee. DAC 009, July 6-31, 009, San Francisco, California, USA. Copyright 009 ACM ACM /08/ $5.00. There are three main steps in lithography simulation 1: photo mask moeling, aerial image simulation an photo resist simulation. This paper focuses on the photo mask moeling. The goal of this step is to compute the near fiel, the electric fiel at the bottom of the computational omain that contains the mask pattern. The existing approaches can be categorize into two groups: fiel solvers an heuristic moels. The fiel solver approach is to solve the Maxwell s equations using well-known numerical techniques such as finite ifference time omain (FDTD) or finite element metho (FEM) 3. This is the most accurate an robust approach. But experience inicates that even the state-of-the-art fiel solvers are too slow or memoryboune to hanle a meium-size mask pattern. In practice, the heuristic moels are use instea to obtain the approximate solution. A commonly use moel is base on the so-calle Kirchhoff approximation 4: if there is a mask opening, the light shines through it without any change in magnitue an phase; otherwise, light is completely blocke. This approximation neglects the effects such as iffraction, polarization an coupling. Attempts have been mae to obtain a moifie mask moel to improve the accuracy 5, 6. The piecewise constant curve fitting approach in 5, 6 is simple an efficient but not accurate an robust. For example, it has been shown in 7 that such moel can result in wrong wafer imaging preiction. The fiel solvers in, 3 an the heuristic moels in 5, 6 sit at the opposite corners in the accuracy-spee trae-off matrix an it is ifficult to smoothly trae accuracy with spee. In our early work 8, the mask moeling was formulate as a parametric moel orer reuction problem base on the finite element iscretization of the Helmholtz equation. The iscretization in 8 uses a uniform an rectangular mesh. This natually leas to the affine parametric form for the stiff an the mass matrices in FEM an hence the well-unerstoo parametric reuction technique 9, 10 can be irectly applie to generate the compact mask moel. However, the non-uniform an unstructure mesh is inispensable to hanle complicate mask patterns. Unfortunately, as will be shown in section 6, a irect application of the technique in 9 coul result in large mask moels. In this paper, we present a new approach to approximate the nonlinearly parameterize FEM matrices with affine forms. This allows the irect application of the parametric reuction in 9, 10 again. Though we emonstrate this new technique for the Helmholtz equation that governs the photo mask moeling, it shoul be straightforwar to apply the same approach to other partial ifferential equations in the context of the parametric moel orer reuction. 3. PROBLEM FORMULATION For the sake of simplicity, we present the problem formulation base on the D example shown in Fig 1. The extension to 3D cases is straightforwar. Assuming an S-polarization (TE) case, the governing D Helmholtz

2 Glass Figure 1: The structure of the D phase-shift mask. Variables w 1, w an s are the parameters in the reuce mask moel. equation is Air u ω εµu= u+k u = 0 (1) where u(x,y) is the z-component of the total electric fiel, ω is the frequency, ε an µ are respectively the ieletric constant an permeability. Following the stanar FEM proceure 11, we obtain the weak form s v u ω Ω Ω sεµvu lvdtn(u) = Ω Ω lvdtn(u in ), () where Ω an Ω are respectively the computational omain an its bounary, v is the testing function, u in is the known incience plane wave an the DtN( ) operator efines the transparent bounary conition 1. In this paper, we assume the perioic bounary conition at the east an the west sie of the computational omain an transparent or non-reflecting bounary conition on the north an the south sie of the computational omain. Using the stanar FEM piecewise polynomial basis functions 11 to iscretize (), we arrive at the parameterize system equation S( w, s) M( w, s) Bū = r, (3) where vector w an s respectively contain the withs an the spacings in the layout, the stiff matrix S( w, s) correspons to the first term in (), the mass matrix M( w, s) correspons to the secon term in (), the matrix B correspons to the thir term in (), an vector r correspons to the right-han-sie term in (). For mask patterns with fixe topology but ifferent with an spacing values, computing the near fiel involves solving equation (3) for ifferent w an s. This kin of multiple-inquery scenario is precisely what the parametric moel orer reuction approach is esigne for. 4. PARAMETRIC MODEL ORDER REDUC- TION A parametric moel orer reuction technique calle Reuce Basis metho has been evelope by the finite element research community 9. A similar iea has also been inepenently propose in the area of parameterize moel orer reuction for circuit simulation 10. As will be shown later, the Reuce Basis is a very powerful iea on top of which we can buil our new mask moel to achieve the esirable accuracy an spee trae-off. Here we summarize its key steps. Please refer to 9 for more etails. Suppose the parameterize governing equation for the problem at han is A( σ)ū = A 0 + f i ( σ)a i ū = r, (4) i where the scalar function f i ( σ) can be arbitrary an the size of vector ū, r an constant matrix A i is N 1, N 1 an N N, respectively. In practical applications, N can be as large as a few millions for a meium-size 3D structure. The Reuce Basis metho in 9 has two stages: the off-line pre-characterization an the on-line evaluation. Off-line pre-characterization stage. e ranomly generate a set σ k = {σ k 1,σk,...} using the given ranges of σ i an solve (4) for ū k. After a few sampling solves, we collect all solutions into the projection matrix an perform projection P = ū 1,ū,...,ū M (5)  i = P T A i P; ˆr = P T r. (6) Now we arrive at the reuce governing equation Aˆ 0 + f i ( σ)â i û = ˆr, (7) i where the size of matrix  i is M M an M is the number of sampling solves. Similar to the stanar proceure in the Moel Orer Reuction 10, the columns in the projection matrix P are orthogonalize using techniques such as incremental QR ecomposition. This makes the matrix P well conitione. In aition, both theoretical an practical ways to estimate the error of the reuce moel are reaily available 9, 10. Hence the off-line moel generation can be mae incremental. On-line evaluation stage. e substitute the given set σ into (7) an solve for û. The approximate solution to equation (4) is obtaine from u = Pû. (8) The key observation here is that the CPU time of the on-line stage is only relate to M, not to N in (4). An M is typically many orers of magnitue smaller than N, as shown by the extensive experiments in 9, 10. Hence equation (7) is a much more efficient but still accurate reuce moel than the original moel in (4). However, this ramatic efficiency gain critically epens on the fact that matrix A i in (4) is not a function of σ. Otherwise, the projection step in (6) involves calculating A i ( σ) at a particular value σ k. This essentially means that the CPU time use by the reuce moel in (7) is relate to the original problem size N an hence we have gaine no efficiency at all 9, 13. This issue of representing the potentially arbitrary nonlinearity in A( σ) in the form amenable to the projection framework in (6) is one of the main challenges in the nonlinear Moel Orer Reuction 14. The main contribution in this paper is to show how to convert S( w, s) an M( w, s) into the appropriate form similar to that in equation (4) so that we can apply the congruence projection in (6). This is one in two steps: parameterization of mesh an parameterization of the FEM matrices. 5. PARAMETERIE MESH In this section, we show an effective technique to parameterize the unstructure triangular mesh in an affine form of the geometry parameters w an s. This is an important first step towar parameterizing the stiff matrix S( w, s) an the mass matrix M( w, s) in (3). hen the size of a geometric feature changes, say w in Fig 1, the mesh points surrouning the feature will move as well. However, to capture the changes in the resulting electric fiel, not all mesh points in the computational omain have to be move. Only mesh points within a certain istance from the changing geometric feature nee to be move. To measure such a istance, we borrow a well-establishe concept calle istance function from the celebrate Level Set Metho 15.

3 5.1 The Distance Function e use a simple example to explain the basic ieas in the istance function. Let (x 1,y 1 ) an (x,y ) be the lower-left an upperright corner of a rectangle, respectively. The istance from a point (x, y) to such a rectangle is efine as (x,y) = min(min(y y 1,y y),min(x x 1,x x)), (9) where function min(a, b) returns the smaller value of the two variables. Fig. shows the contour plot of the istance function for a square where x 1 = y 1 = 5 an x = y = 10. It shoul be note that the zero level set correspons to the bounary of the square. All points outsie of the square have positive istance an all points insie the square have negative istance. 5. The Blening Function The movement of mesh points ue to geometry changes epen on the istance of the mesh points to the changing geometry. Intuitively, the function that characterizes such movement shoul satisfy the following constrains b(η) = 1, i f η = 0 (10) b η < 0, i f 0 < η < 1 (11) b(η) = 0, i f η 1 (1) b η = 0, i f η = 1 (13) where η is the normalize istance efine as η = (x,y) B, (14) B is a user-efine raius of influence, an (x,y) is the istance function like the one efine in (9). Equation (10) means that the mesh points on the geometry move by the same amount as that of the moving geometry features. Equation (11) means that the movement magnitue ecreases monotonically as the mesh points are away from the geometry. Equation (1) means that the movement magnitue is zero if the mesh point is not insie the influence region. Equation (13) ensures a smooth transition of the movement at the bounary of the influence region. In spirit, function b(η) is similar to the so-calle blening function in 16 use to generate the mesh for the computational omain with moving bounaries. In this paper we have chosen the following function as the blening function b(η) = (1 p( η + 0.5)), η 0,1 (15) where p(η) is a thir-orer polynomial p 3 (η) = 3η η 3. (16) 5.3 The Parametric Form of Mesh Arme with the blening function in (15) base on the istance function like the one efine in (9), we are reay to parameterize the movement of the mesh points ue to the change of geometric features. ithout loss of generality, we use the parameter w in Fig. 1 as an example. Suppose w is change by δw. Further more, again without loss of generality, suppose that the center of the opening oes not move, only the left an the right wall move by δw an δw, respectively. ithin the influence region aroun the left an the right sie wall, the location (x,y) of a mesh point can be expresse as x = x 0 + b R b L δw = x 0 + b δw, (17) y = y 0 (18) Figure : Contour plot of the istance to a square. Figure 3: Triangular mesh for the D mask in Fig. 1. Notice the eformation of the 90-egree triangles ue to the change in w 1 an w. where b L = b( L(x 0,y 0 ) B ), b R = b( R(x 0,y 0 ) ) (19) B (x 0,y 0 ) is the initial location, an L (x 0,y 0 ) an R (x 0,y 0 ) are its istance to the left an the right sie wall, respectively. Using linear superposition, one can easily generalize (17) to the multiple parameter cases x = x 0 + < b, σ > (0) where <, > is the inner prouct, an b i is the blening term for parameter σ i which represents the change of w 1, w or s in Figure 1. Clearly the parametric form in (0) is an affine function of the perturbation in each geometry parameter. It shoul be pointe out that the simple parametric form in (17) an (18) can be easily extene to hanle parameter changes along arbitrary irections. But for the purpose of moeling masks, it is probably sufficient to consier just the cases where the parameters are changing only along the horizontal irection (x-irection for D). This will be the case in the remaining part of this paper. Figure 3 shows the eformation in the triangular mesh for the mask shown in Fig. 1 ue to the change in w 1 an w. 6. PARAMETERIE THE FEM MATRICES For the isoparametric linear triangular elements efine on the triangle element with vertexes (x i,y i ),i = 1,,3, the element stiff

4 matrix an mass matrix are respectively 11 an S e = δ x δt x + δ y δt y 4A T (1) M e = 4AM c () where A is the area of the triangle an can be written as A = 1 et( δ x, δ y ), (3) δ x = x 1 x 3,x x 1 T, (4) δ y = y 3 y 1,y 1 y T, (5) 1 1 = 1 0, (6) 0 1 M c = (7) The Parametric Element FEM Matrices Suppose a triangle element is within the region of the influence efine in section 5, in view of (0) an (18), (4) an (5) become x1 x δ x = 3 x1,0 x = 3,0 + < 13, σ > x x 1 x,0 x 1,0 + < (8) 1, σ > δ y = y3,0 y 1,0 y 1,0 y,0 (9) where (x i,0,y i,0 ) is the nominal position for the noe-i, (k) i j = b (k) i b (k) j an b (k) i is the k-th component of the vector b in (0) for noei. For the sake of clarity, we assume that the triangle is a 90-egree triangle that satisfies x1,0 x 3, y3,0 y =, 1,0 hy =, (30) x,0 x 1,0 hx y 1,0 y, then we have < 13 δ x = h x, σ > 1+ < = h 1 x, σ > δ y = h y 1 0 Y 1+ (31), (3) A = 0.5 h y (1+) (33) where Y =< 13, σ > an =< 1, σ >. Substituting (31), (3) an (33) into (1), we obtain the parametric element stiff matrix S e = = + where Y 1+ Y 1+ Y Y 1+ ( = S e 0 + S e 0 = T + h y T ) 0 Y T Y T h y 1 0 T T σ k S e k + Y 1+ Se Se + (34) 0 1 T, Sk e = 1 0 (k) 13 (k) 13 (k) 1 T (35) SN e p +1 = h x 1 0 T, SN e p + = h y 1 0 T. (36) In view of (33) an (), the parametric element mass matrix is M e ( σ) = h y (1+)M c = M0 e + σ k Mk e (37) where M e 0 = M c an M e k = M c (k) The Assemble FEM Matrices The stanar way of generating matrices S( w, s) an M( w, s) is by a proceure calle stamping. The 3 3 element stiff matrix S e ( w, s) an the 3 3 element mass matrix M e ( w, s) are first generate for each element. Using the local-to-global vertex inex map, the entries of S e an M e are irectly ae (stampe) to the large matrices S( w, s) an M( w, s), respectively. The parametric stiff matrix is S( σ) = S 0 + σ k S (k) N f i ( σ)s (i) N + g j ( σ)s ( j) 3 (38) i=1 j=1 where matrices S 0, S (k) 1, S(i) an S ( j) 3 are respectively the reults of the stamping of the element matrices S0 e, Se k, Se +1 an Se + in (34), f i ( σ) = Y i, g j ( σ) = 1, (39) 1+ i 1+ j Y i =< (i) 13, σ >, i =< (i) 1, σ >, (40) N 1 an N are respectively the number of unique f i ( σ) an g j ( σ) for all the triangle elements in the computation omain, is the number of parameters in the vector σ. Similarly, the parametric mass matrix is M( σ) = M 0 + σ k M (k) 1, (41) where matrices M 0 an M (k) 1 are respectively the reults of the stamping of the element matrices M0 e an Me k in (37). 6.3 Regression Using the Kernel Metho The parametric forms in (38) an (41) are very similar to that in (4). But since N 1 an N are relate to the number of elements in the computational omain, so is the size of the reuce moel. Hence we have not obtaine a mask moel that is inepenent of the original problem size yet. In fact, this is the main reason why we can not irectly apply the technique in 9 to generate the reuce mask moel. The next critical task is to approximately represent the large number of f i ( σ) an g j ( σ) by the linear combination of a small number of the so-calle Kernel functions that are inepenent of the original problem size. Essentially we seek to approximate a high-imension space where the nonlinear function f i an g j resie with a much lower imension space. This kin of approximation has been well-stuie in the Machine Learning literatures an the so-calle Kernel Metho has been shown to be effective 17. The gist of the Kernel Metho is the following. Consier representing a scalar function f( x) : R n R of multi-variable argument. An interesting class of approximations are of the form N k f( x) = α k K( x, x k ) (4)

5 where x R n is the evaluation point, K( x,ȳ) : R n R n R is the Kernel, an the N k vectors x k R n are enote as the support vectors. The basic kernel methos 17 use simple function forms for the kernels an pick the coefficient α k base on certain loss functions. In this paper, we choose the kernel form to be the same as the nonlinear functions to be approximate. So the kernel regression form is where N k f i ( σ) α (i) N k k K 1(ξ k ), g j ( σ) β ( j) k K (ξ k ), (43) K 1 (ξ k ) = ξ k, K (ξ 1+ξ k ) = 1, (44) k 1+ξ k an ξ k =< k, σ >, N k is the number of kernels necessary to achieve the esire accuracy. As shown in section 8, 10 to 15 kernels can achieve 3 to 4 igits of accuracy. The coefficients α (i) k an β ( j) k are obtaine by the simple Least Square Fitting, the simplest loss function form 17. The support vectors k are selecte from all possible values of (i) 13 an (i) 1 in (40) by the K-mean Clustering Methos 17. It shoul be note that ue to the locality of the parametric mesh, each mesh noe is influence by a small number of parameters, typically in the D cases an 4 in the 3D cases. This means that the length of the vector (i) 13 an (i) 1 is for D cases an 4 for 3D cases, respectively. Essentially the number of clusters is inepenent of the number of the parameters in the mask structures. This is a key benefit from using the parametric mesh in section 5. Covering a relatively low-imensional space with the K-mean clustering usually means small cluster number N k in (43). Hence the mask moel size is largely etermine by the number of sampling solves in (5). Substituting (43) into (38), we obtain the final parametric form of the stiff matrix S( σ) S 0 + σ k S (k) 1 + N k K 1 (ξ k )S (k) + N k K (ξ k )S (k) 3 (45) where S (k) an S (k) 3 are the results of the stamping α (i) k Se +1 an β ( j) k Se +, respectively. 7. PARAMETRIC MASK MODEL Now we are reay to put everything together to present the final mask moel. Substituting (45) an (41) into (3), we obtain A 0 + σ k A (k) N k 1 + K 1 (ξ k )A (k) N k + K (ξ k )A (k) 3 ū = r, (46) where A 0 = S 0 M 0 B, A (k) 1 = S (k) 1 M(k) 1, A(k) = S (k), an A(k) 3 = S (k) 3. Using the congruence projection in (6), we obtain the reuce moel  0 + σ k  (k) 1 + N k K 1 (ξ k ) (k) N k + K (ξ k ) (k) 3 û = ˆr. (47) The usage of this mask moel is similar to that escribe in section 4. But there is a key ifference. In lithography simulation flow, we only care about the near fiel at the bottom of the mask. In view of (8), this can be accomplishe by using ū n = Λū = ΛPû = P n û, (48) where ū n is the portion of ū that correspons to the near fiel at the bottom of the mask, an the sparse matrix Λ selects those entries from ū. An important benefit from (48) is that we only nee to store P n = ΛP in the mask moel. Matrix P n is much smaller than matrix P. The algorithmic etails for the offline an the online stage are shown in Algorithm 1 an Algorithm, respectively. e want to emphasize again that the cost of online stage is O(N 3 s + ( + N k )N s ). This is clearly inepenent of the orginal problem size N. Algorithm 1: Offline Stage to generate mask moel Input: Range of parameters in σ; mesh; N s : number of samplings; tol: truncation tolerance for rank revealing QR Output:  0,  (k) 1,Â(k),Â(k) 3 ; ˆr; P n (1) Form matrices A 0,A (k) 1,A(i) j),a( 3 an r in (46) () foreach k = 1 : N s (3) Ranomly sample σ (k) using the given ranges (4) Solve (46) for ū (k) (5) P = P ū (k) (6) Run rank-revealing incremental QR to obtain rank r an the fullrank columns P 1,P,...,P r using the given truncation tolerance tol (7) if r < k, i.e., P is rank eficient (8) P = P 1,P,...,P r (9) Exit the sampling loop (10) Use P as projector an compute  0, (k) 1,Â(i) j),â( 3, ˆr as shown in (6) (11) P n = ΛP as shown in (48) Algorithm : Online Stage to evaluate mask moel Input: σ*;  0, (k) 1,Â(k),Â(k) 3 ; ˆr;P n Output: ū n : the near fiel at the bottom of the mask (1) Instantiate the compact mask moel in (47) using σ () Solve (47) for û (3) ū n = P n û as shown in (48) 8. EXPERIMENTAL EVIDENCE e use the mask in Figure 1 to valiate our ieas. For the 3nm noe, the nominal values of w 1, w an s are assume to be 18nm. They have ientically inepenent istribution with the variance being 10%. This 0% variation range is probably more than sufficient for practical consieration. It shoul be note that if we a correlation among these parameters, it will only change the sampling but not the main steps in algorithm 1 an. ith a ranomly generate parameter set σ, we substitute (38) an (41) into (3) an solve (3) for ū 1. This is treate as the accurate solution. This way, we can clearly see the error introuce separately by the Kernelbase fitting in (43) an the moel orer reuction in Algorithm 1. A) Kernel-base Regression Accuracy ith the same parameter set σ mentione above, we solve (46) for ū an then compute the relative L norm error ū 1 ū ū 1. The relative error vs. the number of kernels is shown in Fig. 4. B) Moel Orer Reuction Accuracy In this experiment, N k = 0 kernels are use in (47). The corresponing small error in kernel fitting ensures that the MoR error in (47) ominates the overall error, as is clearly seen in Figure 5. ith the increase of the number

6 Overall Fitting MoR Relative error Relative error Number of Kernels Figure 4: Relative error in regression. of samples N s in Algorithm 1, the relative error ue to moel orer reuction ecreases rapily in Figure 5. C) Mask Moel Accuracy Figure 5 also shows the overall error in the approximate solution provie by the new mask moel. ith about 8 kernels an 10 samples, the mask moel can achieve a 1% overall relative error or -igits of accuracy. D) Mask Moel Spee The main cost of using the reuce mask moel is to instantiate the small ense matrices in (47) an then invert one small ense matrix. Instantiating a few ense matrices an then inverting a ense matrix takes about 1e 4 secon. For the simple two-imensional phase-shift mask shown in Fig. 1, irect use of FEM approach woul result in a sparse matrix A( σ) in (46) with N being aroun The CPU time use by a irect sparse solver for such a system woul be in the orer of a few secons on a esktop PC. So we see a 4-orer of magnitue improvement in spee. In aition, since the CPU time at the online stage irectly relates to the number of samples in Algorithm 1, the smooth trae-off between the accuracy an the CPU time is clearly emonstrate in Figure 5. Further more, our empirical stuies inicate that number of samples is a weak function of the original problem size or the number of parameters. This is certainly a very attractive feature of the new mask moel propose in this paper. 9. CONCLUSIONS In this paper, we propose two key new ieas to improve our work on the parametric mask moel in 8: 1) Parametric unstructure mesh using the istance function an the blening function; ) Kernel-base regression to significantly reuce the number of parametric terms in the final system. Numerical experiments emonstrate that the new mask moel offers smooth accuracy-spee traeoff an hence can be tune to ifferent esign applications. The moel generation in Algorithm 1 is inherently incremental an both theoretical an practical ways to estimate the moel error are reaily available. In aition, the parametric mesh allows the ecoupling between the mesh generation an the parametrization of the FEM matrices, a consierable simplification from implementation point of view. 10. ACKNOLEDGMENTS e want to thank Dr. Joel Phillips from Caence Research Labs for bringing the Kernel Methos in 14, 17 to our attention. e want to thank Prof. Per-olof Persson from University of California at Berkeley for bringing to our attention the use of blening function in 16. An finally, we want to thank Dr. Apo Sezginer from Caence Design Systems for the valuable iscussions Number of Samples Figure 5: Convergence of the L norm relative error in fiel istribution for the mask shown in Figure 1. Each parameter varies between 10% an +10% from its nominal value of 18nm. MoR refers to the error cause by the moel reuction stage. Fit refers to the error cause by the regression fit in (43). Overall refers to the overall error ue to both. 11. REFERENCES 1 A.K.T. ong, Resolution Enhancement Techniques in Optical Lithography, SPIE, Bellingham, A, 001. A. K. ong an A.R. Neureuther, Rigorous three-imensional time-omain finite-ifference electromagnetic simulation for photolithograph applications, in IEEE Trans. on Semiconuctor Manufacturing, Nov. 1995, vol. 8, p S. Burger, R. Kohle, L. schierich, H. Nguyen, F. Schmit, R. Marz, an C. Nolscher, Rigorous simulation of 3D masks, in Proc. SPIE, 006, vol J.. Gooman, Introuction to Fourier Optics, Roberts an Companies, Greenwoo Village, CO, 005, thir eition. 5 Jaione Tirapu-Azpiroz; Paul Burchar; Eli Yablonovitch, Bounary layer moel to account for thick mask effects in photolithography, in Avance Lithography, Proc. SPIE, Feb. 003, p K. Aam an A.R. Neureuther, Domain ecomposition methos for the rapi electromagnetic simulation of photomask scattering, in J. Microlitho., Microfab., an Microsys., Oct. 00, vol. 14, p V. Singh, B. Hu, K. Toh, S. Bollepalli, S. agner, an Y. Boroovsky, Making a tillion pixels ance, Avance Lithography, Proc. SPIE, vol. 694, Feb henhai hu an Frank Schmit, An efficient an robust mask moel for lithography simulation, in Avance Lithography, Proc. SPIE, Feb. 008, vol. 695, p G. Rozza, D.B.P. Huynh, an A. Patera, Reuce basis approximation an a posteriori error estimation for affinely parameterize elliptic coercive partial ifferential equations, Arch. Comput. Methos Eng., vol. 15, pp. 9 75, J.R. Phillips, Variational interconnect analysis via PMTBR, in International Conference on CAD, San Jose, CA, Nov. 004, p J.M. Jin, The Finite Element Metho in Electromagnetics, John illeys an Sons Inc, New York, 00, secon eition. 1 F. Ihlenburg, Finite Element Analysis of Acoustic Scattering, Springer, J. Phillips, Projection-base approaches for moel reuction of weakly nonlinear, time-varying systems, IEEE Trans. on CAD, vol., pp. 171, Feb J.R. Phillips, J. Afonso, A. Oliveira, an L.M. Silveira, Analog macromoeling using kernel methos, in International Conference on CAD, San Jose, CA, Nov. 003, p J.A. Sethian, Level Set methos an Fast Marching Methos, Cambrige University Press, P. Persson, J. Peraire, an J. Bonet, Discontinuous galerkin solution of the navier-stokes equations on eformable omains, in Proc. 45th AIAA Aerospace Sciences Meeting an Exhibit, Jan T. Hastie, R. Tibshirani, an J.H. Frieman, Elements of Statistical Learning, Springer, 003.

Transient analysis of wave propagation in 3D soil by using the scaled boundary finite element method

Transient analysis of wave propagation in 3D soil by using the scaled boundary finite element method Southern Cross University epublications@scu 23r Australasian Conference on the Mechanics of Structures an Materials 214 Transient analysis of wave propagation in 3D soil by using the scale bounary finite

More information

Classifying Facial Expression with Radial Basis Function Networks, using Gradient Descent and K-means

Classifying Facial Expression with Radial Basis Function Networks, using Gradient Descent and K-means Classifying Facial Expression with Raial Basis Function Networks, using Graient Descent an K-means Neil Allrin Department of Computer Science University of California, San Diego La Jolla, CA 9237 nallrin@cs.ucs.eu

More information

A Neural Network Model Based on Graph Matching and Annealing :Application to Hand-Written Digits Recognition

A Neural Network Model Based on Graph Matching and Annealing :Application to Hand-Written Digits Recognition ITERATIOAL JOURAL OF MATHEMATICS AD COMPUTERS I SIMULATIO A eural etwork Moel Base on Graph Matching an Annealing :Application to Han-Written Digits Recognition Kyunghee Lee Abstract We present a neural

More information

6 Gradient Descent. 6.1 Functions

6 Gradient Descent. 6.1 Functions 6 Graient Descent In this topic we will iscuss optimizing over general functions f. Typically the function is efine f : R! R; that is its omain is multi-imensional (in this case -imensional) an output

More information

Characterizing Decoding Robustness under Parametric Channel Uncertainty

Characterizing Decoding Robustness under Parametric Channel Uncertainty Characterizing Decoing Robustness uner Parametric Channel Uncertainty Jay D. Wierer, Wahee U. Bajwa, Nigel Boston, an Robert D. Nowak Abstract This paper characterizes the robustness of ecoing uner parametric

More information

Research Article Inviscid Uniform Shear Flow past a Smooth Concave Body

Research Article Inviscid Uniform Shear Flow past a Smooth Concave Body International Engineering Mathematics Volume 04, Article ID 46593, 7 pages http://x.oi.org/0.55/04/46593 Research Article Invisci Uniform Shear Flow past a Smooth Concave Boy Abullah Mura Department of

More information

Multilevel Linear Dimensionality Reduction using Hypergraphs for Data Analysis

Multilevel Linear Dimensionality Reduction using Hypergraphs for Data Analysis Multilevel Linear Dimensionality Reuction using Hypergraphs for Data Analysis Haw-ren Fang Department of Computer Science an Engineering University of Minnesota; Minneapolis, MN 55455 hrfang@csumneu ABSTRACT

More information

Classical Mechanics Examples (Lagrange Multipliers)

Classical Mechanics Examples (Lagrange Multipliers) Classical Mechanics Examples (Lagrange Multipliers) Dipan Kumar Ghosh Physics Department, Inian Institute of Technology Bombay Powai, Mumbai 400076 September 3, 015 1 Introuction We have seen that the

More information

SURVIVABLE IP OVER WDM: GUARANTEEEING MINIMUM NETWORK BANDWIDTH

SURVIVABLE IP OVER WDM: GUARANTEEEING MINIMUM NETWORK BANDWIDTH SURVIVABLE IP OVER WDM: GUARANTEEEING MINIMUM NETWORK BANDWIDTH Galen H Sasaki Dept Elec Engg, U Hawaii 2540 Dole Street Honolul HI 96822 USA Ching-Fong Su Fuitsu Laboratories of America 595 Lawrence Expressway

More information

Particle Swarm Optimization Based on Smoothing Approach for Solving a Class of Bi-Level Multiobjective Programming Problem

Particle Swarm Optimization Based on Smoothing Approach for Solving a Class of Bi-Level Multiobjective Programming Problem BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 17, No 3 Sofia 017 Print ISSN: 1311-970; Online ISSN: 1314-4081 DOI: 10.1515/cait-017-0030 Particle Swarm Optimization Base

More information

A multiple wavelength unwrapping algorithm for digital fringe profilometry based on spatial shift estimation

A multiple wavelength unwrapping algorithm for digital fringe profilometry based on spatial shift estimation University of Wollongong Research Online Faculty of Engineering an Information Sciences - Papers: Part A Faculty of Engineering an Information Sciences 214 A multiple wavelength unwrapping algorithm for

More information

Learning convex bodies is hard

Learning convex bodies is hard Learning convex boies is har Navin Goyal Microsoft Research Inia navingo@microsoftcom Luis Raemacher Georgia Tech lraemac@ccgatecheu Abstract We show that learning a convex boy in R, given ranom samples

More information

1 Surprises in high dimensions

1 Surprises in high dimensions 1 Surprises in high imensions Our intuition about space is base on two an three imensions an can often be misleaing in high imensions. It is instructive to analyze the shape an properties of some basic

More information

Exercises of PIV. incomplete draft, version 0.0. October 2009

Exercises of PIV. incomplete draft, version 0.0. October 2009 Exercises of PIV incomplete raft, version 0.0 October 2009 1 Images Images are signals efine in 2D or 3D omains. They can be vector value (e.g., color images), real (monocromatic images), complex or binary

More information

PDE-BASED INTERPOLATION METHOD FOR OPTICALLY VISUALIZED SOUND FIELD. Kohei Yatabe and Yasuhiro Oikawa

PDE-BASED INTERPOLATION METHOD FOR OPTICALLY VISUALIZED SOUND FIELD. Kohei Yatabe and Yasuhiro Oikawa 4 IEEE International Conference on Acoustic, Speech an Signal Processing (ICASSP) PDE-BASED INTERPOLATION METHOD FOR OPTICALLY VISUALIZED SOUND FIELD Kohei Yatabe an Yasuhiro Oikawa Department of Intermeia

More information

Skyline Community Search in Multi-valued Networks

Skyline Community Search in Multi-valued Networks Syline Community Search in Multi-value Networs Rong-Hua Li Beijing Institute of Technology Beijing, China lironghuascut@gmail.com Jeffrey Xu Yu Chinese University of Hong Kong Hong Kong, China yu@se.cuh.eu.h

More information

BIJECTIONS FOR PLANAR MAPS WITH BOUNDARIES

BIJECTIONS FOR PLANAR MAPS WITH BOUNDARIES BIJECTIONS FOR PLANAR MAPS WITH BOUNDARIES OLIVIER BERNARDI AND ÉRIC FUSY Abstract. We present bijections for planar maps with bounaries. In particular, we obtain bijections for triangulations an quarangulations

More information

Shift-map Image Registration

Shift-map Image Registration Shift-map Image Registration Svärm, Linus; Stranmark, Petter Unpublishe: 2010-01-01 Link to publication Citation for publishe version (APA): Svärm, L., & Stranmark, P. (2010). Shift-map Image Registration.

More information

A Versatile Model-Based Visibility Measure for Geometric Primitives

A Versatile Model-Based Visibility Measure for Geometric Primitives A Versatile Moel-Base Visibility Measure for Geometric Primitives Marc M. Ellenrieer 1,LarsKrüger 1, Dirk Stößel 2, an Marc Hanheie 2 1 DaimlerChrysler AG, Research & Technology, 89013 Ulm, Germany 2 Faculty

More information

Refinement of scene depth from stereo camera ego-motion parameters

Refinement of scene depth from stereo camera ego-motion parameters Refinement of scene epth from stereo camera ego-motion parameters Piotr Skulimowski, Pawel Strumillo An algorithm for refinement of isparity (epth) map from stereoscopic sequences is propose. The metho

More information

Throughput Characterization of Node-based Scheduling in Multihop Wireless Networks: A Novel Application of the Gallai-Edmonds Structure Theorem

Throughput Characterization of Node-based Scheduling in Multihop Wireless Networks: A Novel Application of the Gallai-Edmonds Structure Theorem Throughput Characterization of Noe-base Scheuling in Multihop Wireless Networks: A Novel Application of the Gallai-Emons Structure Theorem Bo Ji an Yu Sang Dept. of Computer an Information Sciences Temple

More information

Regularized Laplacian Zero Crossings as Optimal Edge Integrators

Regularized Laplacian Zero Crossings as Optimal Edge Integrators Regularize aplacian Zero rossings as Optimal Ege Integrators R. KIMME A.M. BRUKSTEIN Department of omputer Science Technion Israel Institute of Technology Technion ity, Haifa 32, Israel Abstract We view

More information

UNIT 9 INTERFEROMETRY

UNIT 9 INTERFEROMETRY UNIT 9 INTERFEROMETRY Structure 9.1 Introuction Objectives 9. Interference of Light 9.3 Light Sources for 9.4 Applie to Flatness Testing 9.5 in Testing of Surface Contour an Measurement of Height 9.6 Interferometers

More information

Shift-map Image Registration

Shift-map Image Registration Shift-map Image Registration Linus Svärm Petter Stranmark Centre for Mathematical Sciences, Lun University {linus,petter}@maths.lth.se Abstract Shift-map image processing is a new framework base on energy

More information

5th International Conference on Advanced Design and Manufacturing Engineering (ICADME 2015)

5th International Conference on Advanced Design and Manufacturing Engineering (ICADME 2015) 5th International Conference on Avance Design an Manufacturing Engineering (ICADME 25) Research on motion characteristics an application of multi egree of freeom mechanism base on R-W metho Xiao-guang

More information

A Plane Tracker for AEC-automation Applications

A Plane Tracker for AEC-automation Applications A Plane Tracker for AEC-automation Applications Chen Feng *, an Vineet R. Kamat Department of Civil an Environmental Engineering, University of Michigan, Ann Arbor, USA * Corresponing author (cforrest@umich.eu)

More information

Architecture Design of Mobile Access Coordinated Wireless Sensor Networks

Architecture Design of Mobile Access Coordinated Wireless Sensor Networks Architecture Design of Mobile Access Coorinate Wireless Sensor Networks Mai Abelhakim 1 Leonar E. Lightfoot Jian Ren 1 Tongtong Li 1 1 Department of Electrical & Computer Engineering, Michigan State University,

More information

Intensive Hypercube Communication: Prearranged Communication in Link-Bound Machines 1 2

Intensive Hypercube Communication: Prearranged Communication in Link-Bound Machines 1 2 This paper appears in J. of Parallel an Distribute Computing 10 (1990), pp. 167 181. Intensive Hypercube Communication: Prearrange Communication in Link-Boun Machines 1 2 Quentin F. Stout an Bruce Wagar

More information

Coupling the User Interfaces of a Multiuser Program

Coupling the User Interfaces of a Multiuser Program Coupling the User Interfaces of a Multiuser Program PRASUN DEWAN University of North Carolina at Chapel Hill RAJIV CHOUDHARY Intel Corporation We have evelope a new moel for coupling the user-interfaces

More information

Advanced method of NC programming for 5-axis machining

Advanced method of NC programming for 5-axis machining Available online at www.scienceirect.com Proceia CIRP (0 ) 0 07 5 th CIRP Conference on High Performance Cutting 0 Avance metho of NC programming for 5-axis machining Sergej N. Grigoriev a, A.A. Kutin

More information

Study of Network Optimization Method Based on ACL

Study of Network Optimization Method Based on ACL Available online at www.scienceirect.com Proceia Engineering 5 (20) 3959 3963 Avance in Control Engineering an Information Science Stuy of Network Optimization Metho Base on ACL Liu Zhian * Department

More information

AN INVESTIGATION OF FOCUSING AND ANGULAR TECHNIQUES FOR VOLUMETRIC IMAGES BY USING THE 2D CIRCULAR ULTRASONIC PHASED ARRAY

AN INVESTIGATION OF FOCUSING AND ANGULAR TECHNIQUES FOR VOLUMETRIC IMAGES BY USING THE 2D CIRCULAR ULTRASONIC PHASED ARRAY AN INVESTIGATION OF FOCUSING AND ANGULAR TECHNIQUES FOR VOLUMETRIC IMAGES BY USING THE D CIRCULAR ULTRASONIC PHASED ARRAY S. Monal Lonon South Bank University; Engineering an Design 103 Borough Roa, Lonon

More information

Almost Disjunct Codes in Large Scale Multihop Wireless Network Media Access Control

Almost Disjunct Codes in Large Scale Multihop Wireless Network Media Access Control Almost Disjunct Coes in Large Scale Multihop Wireless Network Meia Access Control D. Charles Engelhart Anan Sivasubramaniam Penn. State University University Park PA 682 engelhar,anan @cse.psu.eu Abstract

More information

Generalized Edge Coloring for Channel Assignment in Wireless Networks

Generalized Edge Coloring for Channel Assignment in Wireless Networks Generalize Ege Coloring for Channel Assignment in Wireless Networks Chun-Chen Hsu Institute of Information Science Acaemia Sinica Taipei, Taiwan Da-wei Wang Jan-Jan Wu Institute of Information Science

More information

Calculation on diffraction aperture of cube corner retroreflector

Calculation on diffraction aperture of cube corner retroreflector November 10, 008 / Vol., No. 11 / CHINESE OPTICS LETTERS 8 Calculation on iffraction aperture of cube corner retroreflector Song Li (Ó Ø, Bei Tang (», an Hui Zhou ( ï School of Electronic Information,

More information

SMART IMAGE PROCESSING OF FLOW VISUALIZATION

SMART IMAGE PROCESSING OF FLOW VISUALIZATION SMAR IMAGE PROCESSING OF FLOW VISUALIZAION H Li (A Rinoshika) 1, M akei, M Nakano 1, Y Saito 3 an K Horii 4 1 Department of Mechanical Systems Engineering, Yamagata University, Yamagata 99-851, JAPAN Department

More information

Selection Strategies for Initial Positions and Initial Velocities in Multi-optima Particle Swarms

Selection Strategies for Initial Positions and Initial Velocities in Multi-optima Particle Swarms ACM, 2011. This is the author s version of the work. It is poste here by permission of ACM for your personal use. Not for reistribution. The efinitive version was publishe in Proceeings of the 13th Annual

More information

On the Placement of Internet Taps in Wireless Neighborhood Networks

On the Placement of Internet Taps in Wireless Neighborhood Networks 1 On the Placement of Internet Taps in Wireless Neighborhoo Networks Lili Qiu, Ranveer Chanra, Kamal Jain, Mohamma Mahian Abstract Recently there has emerge a novel application of wireless technology that

More information

Bends, Jogs, And Wiggles for Railroad Tracks and Vehicle Guide Ways

Bends, Jogs, And Wiggles for Railroad Tracks and Vehicle Guide Ways Ben, Jogs, An Wiggles for Railroa Tracks an Vehicle Guie Ways Louis T. Klauer Jr., PhD, PE. Work Soft 833 Galer Dr. Newtown Square, PA 19073 lklauer@wsof.com Preprint, June 4, 00 Copyright 00 by Louis

More information

A Classification of 3R Orthogonal Manipulators by the Topology of their Workspace

A Classification of 3R Orthogonal Manipulators by the Topology of their Workspace A Classification of R Orthogonal Manipulators by the Topology of their Workspace Maher aili, Philippe Wenger an Damien Chablat Institut e Recherche en Communications et Cybernétique e Nantes, UMR C.N.R.S.

More information

Digital fringe profilometry based on triangular fringe patterns and spatial shift estimation

Digital fringe profilometry based on triangular fringe patterns and spatial shift estimation University of Wollongong Research Online Faculty of Engineering an Information Sciences - Papers: Part A Faculty of Engineering an Information Sciences 4 Digital fringe profilometry base on triangular

More information

Loop Scheduling and Partitions for Hiding Memory Latencies

Loop Scheduling and Partitions for Hiding Memory Latencies Loop Scheuling an Partitions for Hiing Memory Latencies Fei Chen Ewin Hsing-Mean Sha Dept. of Computer Science an Engineering University of Notre Dame Notre Dame, IN 46556 Email: fchen,esha @cse.n.eu Tel:

More information

AnyTraffic Labeled Routing

AnyTraffic Labeled Routing AnyTraffic Labele Routing Dimitri Papaimitriou 1, Pero Peroso 2, Davie Careglio 2 1 Alcatel-Lucent Bell, Antwerp, Belgium Email: imitri.papaimitriou@alcatel-lucent.com 2 Universitat Politècnica e Catalunya,

More information

Kinematic Analysis of a Family of 3R Manipulators

Kinematic Analysis of a Family of 3R Manipulators Kinematic Analysis of a Family of R Manipulators Maher Baili, Philippe Wenger an Damien Chablat Institut e Recherche en Communications et Cybernétique e Nantes, UMR C.N.R.S. 6597 1, rue e la Noë, BP 92101,

More information

Approximation with Active B-spline Curves and Surfaces

Approximation with Active B-spline Curves and Surfaces Approximation with Active B-spline Curves an Surfaces Helmut Pottmann, Stefan Leopolseer, Michael Hofer Institute of Geometry Vienna University of Technology Wiener Hauptstr. 8 10, Vienna, Austria pottmann,leopolseer,hofer

More information

EXTRACTION OF MAIN AND SECONDARY ROADS IN VHR IMAGES USING A HIGHER-ORDER PHASE FIELD MODEL

EXTRACTION OF MAIN AND SECONDARY ROADS IN VHR IMAGES USING A HIGHER-ORDER PHASE FIELD MODEL EXTRACTION OF MAIN AND SECONDARY ROADS IN VHR IMAGES USING A HIGHER-ORDER PHASE FIELD MODEL Ting Peng a,b, Ian H. Jermyn b, Véronique Prinet a, Josiane Zerubia b a LIAMA & PR, Institute of Automation,

More information

A Highly Scalable Parallel Boundary Element Method for Capacitance Extraction

A Highly Scalable Parallel Boundary Element Method for Capacitance Extraction A Highly Scalable Parallel Bounary Element Metho for Capacitance Extraction The MIT Faculty has mae this article openly available. Please share how this access benefits you. Your story matters. Citation

More information

Performance Modelling of Necklace Hypercubes

Performance Modelling of Necklace Hypercubes erformance Moelling of ecklace ypercubes. Meraji,,. arbazi-aza,, A. atooghy, IM chool of Computer cience & harif University of Technology, Tehran, Iran {meraji, patooghy}@ce.sharif.eu, aza@ipm.ir a Abstract

More information

Interference and diffraction are the important phenomena that distinguish. Interference and Diffraction

Interference and diffraction are the important phenomena that distinguish. Interference and Diffraction C H A P T E R 33 Interference an Diffraction 33- Phase Difference an Coherence 33-2 Interference in Thin Films 33-3 Two-Slit Interference Pattern 33-4 Diffraction Pattern of a Single Slit * 33-5 Using

More information

Figure 1: Schematic of an SEM [source: ]

Figure 1: Schematic of an SEM [source:   ] EECI Course: -9 May 1 by R. Sanfelice Hybri Control Systems Eelco van Horssen E.P.v.Horssen@tue.nl Project: Scanning Electron Microscopy Introuction In Scanning Electron Microscopy (SEM) a (bunle) beam

More information

Online Appendix to: Generalizing Database Forensics

Online Appendix to: Generalizing Database Forensics Online Appenix to: Generalizing Database Forensics KYRIACOS E. PAVLOU an RICHARD T. SNODGRASS, University of Arizona This appenix presents a step-by-step iscussion of the forensic analysis protocol that

More information

Linear First-Order PDEs

Linear First-Order PDEs MODULE 2: FIRST-ORDER PARTIAL DIFFERENTIAL EQUATIONS 9 Lecture 2 Linear First-Orer PDEs The most general first-orer linear PDE has the form a(x, y)z x + b(x, y)z y + c(x, y)z = (x, y), (1) where a, b,

More information

Animated Surface Pasting

Animated Surface Pasting Animate Surface Pasting Clara Tsang an Stephen Mann Computing Science Department University of Waterloo 200 University Ave W. Waterloo, Ontario Canaa N2L 3G1 e-mail: clftsang@cgl.uwaterloo.ca, smann@cgl.uwaterloo.ca

More information

Using Vector and Raster-Based Techniques in Categorical Map Generalization

Using Vector and Raster-Based Techniques in Categorical Map Generalization Thir ICA Workshop on Progress in Automate Map Generalization, Ottawa, 12-14 August 1999 1 Using Vector an Raster-Base Techniques in Categorical Map Generalization Beat Peter an Robert Weibel Department

More information

Optimization of cable-stayed bridges with box-girder decks

Optimization of cable-stayed bridges with box-girder decks Avances in Engineering Software 31 (2000) 417 423 www.elsevier.com/locate/avengsoft Optimization of cable-staye briges with box-girer ecks L.M.C. Simões*, J.H.J.O. Negrão Department of Civil Engineering,

More information

Optimal Oblivious Path Selection on the Mesh

Optimal Oblivious Path Selection on the Mesh Optimal Oblivious Path Selection on the Mesh Costas Busch Malik Magon-Ismail Jing Xi Department of Computer Science Rensselaer Polytechnic Institute Troy, NY 280, USA {buschc,magon,xij2}@cs.rpi.eu Abstract

More information

Learning Subproblem Complexities in Distributed Branch and Bound

Learning Subproblem Complexities in Distributed Branch and Bound Learning Subproblem Complexities in Distribute Branch an Boun Lars Otten Department of Computer Science University of California, Irvine lotten@ics.uci.eu Rina Dechter Department of Computer Science University

More information

Tight Wavelet Frame Decomposition and Its Application in Image Processing

Tight Wavelet Frame Decomposition and Its Application in Image Processing ITB J. Sci. Vol. 40 A, No., 008, 151-165 151 Tight Wavelet Frame Decomposition an Its Application in Image Processing Mahmu Yunus 1, & Henra Gunawan 1 1 Analysis an Geometry Group, FMIPA ITB, Banung Department

More information

Modifying ROC Curves to Incorporate Predicted Probabilities

Modifying ROC Curves to Incorporate Predicted Probabilities Moifying ROC Curves to Incorporate Preicte Probabilities Cèsar Ferri DSIC, Universitat Politècnica e València Peter Flach Department of Computer Science, University of Bristol José Hernánez-Orallo DSIC,

More information

Cluster Center Initialization Method for K-means Algorithm Over Data Sets with Two Clusters

Cluster Center Initialization Method for K-means Algorithm Over Data Sets with Two Clusters Available online at www.scienceirect.com Proceia Engineering 4 (011 ) 34 38 011 International Conference on Avances in Engineering Cluster Center Initialization Metho for K-means Algorithm Over Data Sets

More information

A Revised Simplex Search Procedure for Stochastic Simulation Response Surface Optimization

A Revised Simplex Search Procedure for Stochastic Simulation Response Surface Optimization 272 INFORMS Journal on Computing 0899-1499 100 1204-0272 $05.00 Vol. 12, No. 4, Fall 2000 2000 INFORMS A Revise Simplex Search Proceure for Stochastic Simulation Response Surface Optimization DAVID G.

More information

Iterative Computation of Moment Forms for Subdivision Surfaces

Iterative Computation of Moment Forms for Subdivision Surfaces Iterative Computation of Moment Forms for Subivision Surfaces Jan P. Hakenberg ETH Zürich Ulrich Reif TU Darmstat Figure : Solis boune by Catmull-Clark an Loop subivision surfaces with centrois inicate

More information

Generalized Low Rank Approximations of Matrices

Generalized Low Rank Approximations of Matrices Machine Learning, 2005 2005 Springer Science + Business Meia, Inc.. Manufacture in The Netherlans. DOI: 10.1007/s10994-005-3561-6 Generalize Low Rank Approximations of Matrices JIEPING YE * jieping@cs.umn.eu

More information

Software Reliability Modeling and Cost Estimation Incorporating Testing-Effort and Efficiency

Software Reliability Modeling and Cost Estimation Incorporating Testing-Effort and Efficiency Software Reliability Moeling an Cost Estimation Incorporating esting-effort an Efficiency Chin-Yu Huang, Jung-Hua Lo, Sy-Yen Kuo, an Michael R. Lyu -+ Department of Electrical Engineering Computer Science

More information

Figure 1: 2D arm. Figure 2: 2D arm with labelled angles

Figure 1: 2D arm. Figure 2: 2D arm with labelled angles 2D Kinematics Consier a robotic arm. We can sen it commans like, move that joint so it bens at an angle θ. Once we ve set each joint, that s all well an goo. More interesting, though, is the question of

More information

Here are a couple of warnings to my students who may be here to get a copy of what happened on a day that you missed.

Here are a couple of warnings to my students who may be here to get a copy of what happened on a day that you missed. Preface Here are my online notes for my Calculus I course that I teach here at Lamar University. Despite the fact that these are my class notes, they shoul be accessible to anyone wanting to learn Calculus

More information

Distributed Decomposition Over Hyperspherical Domains

Distributed Decomposition Over Hyperspherical Domains Distribute Decomposition Over Hyperspherical Domains Aron Ahmaia 1, Davi Keyes 1, Davi Melville 2, Alan Rosenbluth 2, Kehan Tian 2 1 Department of Applie Physics an Applie Mathematics, Columbia University,

More information

FINDING OPTICAL DISPERSION OF A PRISM WITH APPLICATION OF MINIMUM DEVIATION ANGLE MEASUREMENT METHOD

FINDING OPTICAL DISPERSION OF A PRISM WITH APPLICATION OF MINIMUM DEVIATION ANGLE MEASUREMENT METHOD Warsaw University of Technology Faculty of Physics Physics Laboratory I P Joanna Konwerska-Hrabowska 6 FINDING OPTICAL DISPERSION OF A PRISM WITH APPLICATION OF MINIMUM DEVIATION ANGLE MEASUREMENT METHOD.

More information

B-Splines and NURBS Week 5, Lecture 9

B-Splines and NURBS Week 5, Lecture 9 CS 430/585 Computer Graphics I B-Splines an NURBS Week 5, Lecture 9 Davi Breen, William Regli an Maxim Peysakhov Geometric an Intelligent Computing Laboratory Department of Computer Science Drexel University

More information

Comparative Study of Projection/Back-projection Schemes in Cryo-EM Tomography

Comparative Study of Projection/Back-projection Schemes in Cryo-EM Tomography Comparative Stuy of Projection/Back-projection Schemes in Cryo-EM Tomography Yu Liu an Jong Chul Ye Department of BioSystems Korea Avance Institute of Science an Technology, Daejeon, Korea ABSTRACT In

More information

Estimating Velocity Fields on a Freeway from Low Resolution Video

Estimating Velocity Fields on a Freeway from Low Resolution Video Estimating Velocity Fiels on a Freeway from Low Resolution Vieo Young Cho Department of Statistics University of California, Berkeley Berkeley, CA 94720-3860 Email: young@stat.berkeley.eu John Rice Department

More information

Image compression predicated on recurrent iterated function systems

Image compression predicated on recurrent iterated function systems 2n International Conference on Mathematics & Statistics 16-19 June, 2008, Athens, Greece Image compression preicate on recurrent iterate function systems Chol-Hui Yun *, Metzler W. a an Barski M. a * Faculty

More information

Improving Performance of Sparse Matrix-Vector Multiplication

Improving Performance of Sparse Matrix-Vector Multiplication Improving Performance of Sparse Matrix-Vector Multiplication Ali Pınar Michael T. Heath Department of Computer Science an Center of Simulation of Avance Rockets University of Illinois at Urbana-Champaign

More information

Non-homogeneous Generalization in Privacy Preserving Data Publishing

Non-homogeneous Generalization in Privacy Preserving Data Publishing Non-homogeneous Generalization in Privacy Preserving Data Publishing W. K. Wong, Nios Mamoulis an Davi W. Cheung Department of Computer Science, The University of Hong Kong Pofulam Roa, Hong Kong {wwong2,nios,cheung}@cs.hu.h

More information

THE APPLICATION OF ARTICLE k-th SHORTEST TIME PATH ALGORITHM

THE APPLICATION OF ARTICLE k-th SHORTEST TIME PATH ALGORITHM International Journal of Physics an Mathematical Sciences ISSN: 2277-2111 (Online) 2016 Vol. 6 (1) January-March, pp. 24-6/Mao an Shi. THE APPLICATION OF ARTICLE k-th SHORTEST TIME PATH ALGORITHM Hua Mao

More information

Image Segmentation using K-means clustering and Thresholding

Image Segmentation using K-means clustering and Thresholding Image Segmentation using Kmeans clustering an Thresholing Preeti Panwar 1, Girhar Gopal 2, Rakesh Kumar 3 1M.Tech Stuent, Department of Computer Science & Applications, Kurukshetra University, Kurukshetra,

More information

Tracking and Regulation Control of a Mobile Robot System With Kinematic Disturbances: A Variable Structure-Like Approach

Tracking and Regulation Control of a Mobile Robot System With Kinematic Disturbances: A Variable Structure-Like Approach W. E. Dixon e-mail: wixon@ces.clemson.eu D. M. Dawson e-mail: awson@ces.clemson.eu E. Zergeroglu e-mail: ezerger@ces.clemson.eu Department of Electrical & Computer Engineering, Clemson University, Clemson,

More information

Frequency Domain Parameter Estimation of a Synchronous Generator Using Bi-objective Genetic Algorithms

Frequency Domain Parameter Estimation of a Synchronous Generator Using Bi-objective Genetic Algorithms Proceeings of the 7th WSEAS International Conference on Simulation, Moelling an Optimization, Beijing, China, September 15-17, 2007 429 Frequenc Domain Parameter Estimation of a Snchronous Generator Using

More information

Investigation into a new incremental forming process using an adjustable punch set for the manufacture of a doubly curved sheet metal

Investigation into a new incremental forming process using an adjustable punch set for the manufacture of a doubly curved sheet metal 991 Investigation into a new incremental forming process using an ajustable punch set for the manufacture of a oubly curve sheet metal S J Yoon an D Y Yang* Department of Mechanical Engineering, Korea

More information

Learning Polynomial Functions. by Feature Construction

Learning Polynomial Functions. by Feature Construction I Proceeings of the Eighth International Workshop on Machine Learning Chicago, Illinois, June 27-29 1991 Learning Polynomial Functions by Feature Construction Richar S. Sutton GTE Laboratories Incorporate

More information

A Convex Clustering-based Regularizer for Image Segmentation

A Convex Clustering-based Regularizer for Image Segmentation Vision, Moeling, an Visualization (2015) D. Bommes, T. Ritschel an T. Schultz (Es.) A Convex Clustering-base Regularizer for Image Segmentation Benjamin Hell (TU Braunschweig), Marcus Magnor (TU Braunschweig)

More information

New Version of Davies-Bouldin Index for Clustering Validation Based on Cylindrical Distance

New Version of Davies-Bouldin Index for Clustering Validation Based on Cylindrical Distance New Version of Davies-Boulin Inex for lustering Valiation Base on ylinrical Distance Juan arlos Roas Thomas Faculta e Informática Universia omplutense e Mari Mari, España correoroas@gmail.com Abstract

More information

EFFICIENT ON-LINE TESTING METHOD FOR A FLOATING-POINT ADDER

EFFICIENT ON-LINE TESTING METHOD FOR A FLOATING-POINT ADDER FFICINT ON-LIN TSTING MTHOD FOR A FLOATING-POINT ADDR A. Droz, M. Lobachev Department of Computer Systems, Oessa State Polytechnic University, Oessa, Ukraine Droz@ukr.net, Lobachev@ukr.net Abstract In

More information

A Duality Based Approach for Realtime TV-L 1 Optical Flow

A Duality Based Approach for Realtime TV-L 1 Optical Flow A Duality Base Approach for Realtime TV-L 1 Optical Flow C. Zach 1, T. Pock 2, an H. Bischof 2 1 VRVis Research Center 2 Institute for Computer Graphics an Vision, TU Graz Abstract. Variational methos

More information

Adjacency Matrix Based Full-Text Indexing Models

Adjacency Matrix Based Full-Text Indexing Models 1000-9825/2002/13(10)1933-10 2002 Journal of Software Vol.13, No.10 Ajacency Matrix Base Full-Text Inexing Moels ZHOU Shui-geng 1, HU Yun-fa 2, GUAN Ji-hong 3 1 (Department of Computer Science an Engineering,

More information

Handling missing values in kernel methods with application to microbiology data

Handling missing values in kernel methods with application to microbiology data an Machine Learning. Bruges (Belgium), 24-26 April 2013, i6oc.com publ., ISBN 978-2-87419-081-0. Available from http://www.i6oc.com/en/livre/?gcoi=28001100131010. Hanling missing values in kernel methos

More information

Navigation Around an Unknown Obstacle for Autonomous Surface Vehicles Using a Forward-Facing Sonar

Navigation Around an Unknown Obstacle for Autonomous Surface Vehicles Using a Forward-Facing Sonar Navigation Aroun an nknown Obstacle for Autonomous Surface Vehicles sing a Forwar-Facing Sonar Patrick A. Plonski, Joshua Vaner Hook, Cheng Peng, Narges Noori, Volkan Isler Abstract A robotic boat is moving

More information

Fast Fractal Image Compression using PSO Based Optimization Techniques

Fast Fractal Image Compression using PSO Based Optimization Techniques Fast Fractal Compression using PSO Base Optimization Techniques A.Krishnamoorthy Visiting faculty Department Of ECE University College of Engineering panruti rishpci89@gmail.com S.Buvaneswari Visiting

More information

EXACT SIMULATION OF A BOOLEAN MODEL

EXACT SIMULATION OF A BOOLEAN MODEL Original Research Paper oi:10.5566/ias.v32.p101-105 EXACT SIMULATION OF A BOOLEAN MODEL CHRISTIAN LANTUÉJOULB MinesParisTech 35 rue Saint-Honoré 77305 Fontainebleau France e-mail: christian.lantuejoul@mines-paristech.fr

More information

Finite Automata Implementations Considering CPU Cache J. Holub

Finite Automata Implementations Considering CPU Cache J. Holub Finite Automata Implementations Consiering CPU Cache J. Holub The finite automata are mathematical moels for finite state systems. More general finite automaton is the noneterministic finite automaton

More information

Generalized Edge Coloring for Channel Assignment in Wireless Networks

Generalized Edge Coloring for Channel Assignment in Wireless Networks TR-IIS-05-021 Generalize Ege Coloring for Channel Assignment in Wireless Networks Chun-Chen Hsu, Pangfeng Liu, Da-Wei Wang, Jan-Jan Wu December 2005 Technical Report No. TR-IIS-05-021 http://www.iis.sinica.eu.tw/lib/techreport/tr2005/tr05.html

More information

Bayesian localization microscopy reveals nanoscale podosome dynamics

Bayesian localization microscopy reveals nanoscale podosome dynamics Nature Methos Bayesian localization microscopy reveals nanoscale poosome ynamics Susan Cox, Ewar Rosten, James Monypenny, Tijana Jovanovic-Talisman, Dylan T Burnette, Jennifer Lippincott-Schwartz, Gareth

More information

State Indexed Policy Search by Dynamic Programming. Abstract. 1. Introduction. 2. System parameterization. Charles DuHadway

State Indexed Policy Search by Dynamic Programming. Abstract. 1. Introduction. 2. System parameterization. Charles DuHadway State Inexe Policy Search by Dynamic Programming Charles DuHaway Yi Gu 5435537 503372 December 4, 2007 Abstract We consier the reinforcement learning problem of simultaneous trajectory-following an obstacle

More information

filtering LETTER An Improved Neighbor Selection Algorithm in Collaborative Taek-Hun KIM a), Student Member and Sung-Bong YANG b), Nonmember

filtering LETTER An Improved Neighbor Selection Algorithm in Collaborative Taek-Hun KIM a), Student Member and Sung-Bong YANG b), Nonmember 107 IEICE TRANS INF & SYST, VOLE88 D, NO5 MAY 005 LETTER An Improve Neighbor Selection Algorithm in Collaborative Filtering Taek-Hun KIM a), Stuent Member an Sung-Bong YANG b), Nonmember SUMMARY Nowaays,

More information

Using Ray Tracing for Site-Specific Indoor Radio Signal Strength Analysis 1

Using Ray Tracing for Site-Specific Indoor Radio Signal Strength Analysis 1 Using Ray Tracing for Site-Specific Inoor Raio Signal Strength Analysis 1 Michael Ni, Stephen Mann, an Jay Black Computer Science Department, University of Waterloo, Waterloo, Ontario, NL G1, Canaa Abstract

More information

Polygon Simplification by Minimizing Convex Corners

Polygon Simplification by Minimizing Convex Corners Polygon Simplification by Minimizing Convex Corners Yeganeh Bahoo 1, Stephane Durocher 1, J. Mark Keil 2, Saee Mehrabi 3, Sahar Mehrpour 1, an Debajyoti Monal 1 1 Department of Computer Science, University

More information

Non-Uniform Sensor Deployment in Mobile Wireless Sensor Networks

Non-Uniform Sensor Deployment in Mobile Wireless Sensor Networks 01 01 01 01 01 00 01 01 Non-Uniform Sensor Deployment in Mobile Wireless Sensor Networks Mihaela Carei, Yinying Yang, an Jie Wu Department of Computer Science an Engineering Floria Atlantic University

More information

A shortest path algorithm in multimodal networks: a case study with time varying costs

A shortest path algorithm in multimodal networks: a case study with time varying costs A shortest path algorithm in multimoal networks: a case stuy with time varying costs Daniela Ambrosino*, Anna Sciomachen* * Department of Economics an Quantitative Methos (DIEM), University of Genoa Via

More information

Technical Report TR Navigation Around an Unknown Obstacle for Autonomous Surface Vehicles Using a Forward-Facing Sonar

Technical Report TR Navigation Around an Unknown Obstacle for Autonomous Surface Vehicles Using a Forward-Facing Sonar Technical Report Department of Computer Science an Engineering niversity of Minnesota 4-192 Keller Hall 2 nion Street SE Minneapolis, MN 55455-159 SA TR 15-5 Navigation Aroun an nknown Obstacle for Autonomous

More information

Dual Arm Robot Research Report

Dual Arm Robot Research Report Dual Arm Robot Research Report Analytical Inverse Kinematics Solution for Moularize Dual-Arm Robot With offset at shouler an wrist Motivation an Abstract Generally, an inustrial manipulator such as PUMA

More information