-
Notifications
You must be signed in to change notification settings - Fork 0
/
2023-02-12.html
executable file
·535 lines (454 loc) · 22.5 KB
/
2023-02-12.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta http-equiv="X-UA-Compatible" content="IE=Edge">
<meta name="description" content="">
<meta name="keywords" content="">
<meta name="author" content="">
<meta name="viewport" content="width=device-width, initial-scale=1, maximum-scale=1">
<title>Finite-boundary element coupling in MATLAB</title>
<!--
Template 2085 Neuron
http://www.tooplate.com/view/2085-neuron
-->
<link rel="stylesheet" href="css/bootstrap.min.css">
<link rel="stylesheet" href="css/all.css">
<link rel="stylesheet" href="css/magnific-popup.css">
<!-- Main css -->
<link rel="stylesheet" href="css/style.css">
<link href="https://fonts.googleapis.com/css?family=Lora|Merriweather:300,400" rel="stylesheet">
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=UA-108246943-1"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'UA-108246943-1');
</script>
<!-- Google AdSense -->
<script async src="https://pagead2.googlesyndication.com/pagead/js/adsbygoogle.js?client=ca-pub-8299683413193339" crossorigin="anonymous"></script>
<!--<script async src="https://fundingchoicesmessages.google.com/i/pub-8299683413193339?ers=1" nonce="AuY4vKpDElA9N7YwmFfMjg"></script>
<script nonce="AuY4vKpDElA9N7YwmFfMjg">(function() {function signalGooglefcPresent() {if (!window.frames['googlefcPresent']) {if (document.body) {const iframe = document.createElement('iframe'); iframe.style = 'width: 0; height: 0; border: none; z-index: -1000; left: -1000px; top: -1000px;'; iframe.style.display = 'none'; iframe.name = 'googlefcPresent'; document.body.appendChild(iframe);} else {setTimeout(signalGooglefcPresent, 0);}}}signalGooglefcPresent();})();</script>-->
</head>
<body>
<!-- Tikz for mathjax -->
<link rel="stylesheet" type="text/css" href="https://tikzjax.com/v1/fonts.css">
<script src="https://tikzjax.com/v1/tikzjax.js"></script>
<!-- PRE LOADER -->
<div class="preloader">
<div class="sk-spinner sk-spinner-wordpress">
<span class="sk-inner-circle"></span>
</div>
</div>
<!-- Navigation section -->
<div class="navbar navbar-default navbar-static-top" role="navigation">
<div class="container">
<div class="navbar-header">
<button class="navbar-toggle" data-toggle="collapse" data-target=".navbar-collapse">
<span class="icon icon-bar"></span>
<span class="icon icon-bar"></span>
<span class="icon icon-bar"></span>
</button>
<a href="index.html" class="navbar-brand">H. Montanelli</a>
</div>
<div class="collapse navbar-collapse">
<ul class="nav navbar-nav navbar-right">
<li><a href="about.html">About</a></li>
<li><a href="blog.html">Blog</a></li>
<li><a href="publications.html">Publications</a></li>
<li><a href="research.html">Research</a></li>
<li><a href="students.html">Students</a></li>
<li><a href="talks.html">Talks</a></li>
<li><a href="teaching.html">Teaching</a></li>
<li><a href="https://www.paypal.com/donate/?hosted_button_id=UCLCSJLFL433E"><b>Donate</b></a></li>
</ul>
</div>
</div>
</div>
<!-- Home Section -->
<section id="home" class="main-about parallax-section">
<div class="overlay"></div>
<div class="container">
<div class="row">
<div class="col-md-12 col-sm-12">
<h1>Finite-boundary element coupling in MATLAB</h1>
</div>
</div>
</div>
</section>
<!-- Blog Single Post Section -->
<section id="about">
<div class="container">
<div class="row">
<div class="col-md-offset-1 col-md-10 col-sm-12">
<p><i>February 12, 2023 — Support my next blog post, <a href="https://www.paypal.com/donate/?hosted_button_id=UCLCSJLFL433E">buy me a coffee</a> ☕.</i></p>
<p>In this post, I talk about coupling the finite and boundary element methods for solving <a href="https://en.wikipedia.org/wiki/Wave_propagation">wave propagation</a> problems in an inhomogeneous medium.</p>
<h2>Acoustic waves in an inhomogeneous medium</h2>
<center>
<img src="/blog/fembem-tik1.png" class="img-responsive">
</center>
<p>We consider the following model with a total field \(u\) coupled with a scattered field \(w^s\):
$$
\begin{align}
& \Delta u + k^2nu = 0 \quad \text{in $B\setminus\overline{D}$}, \\[.4em]
& u = 0 \quad \text{on $\partial D$}, \\[0.4em]
& \Delta w^s + k^2w^s = 0 \quad \text{in $\mathbb{R}^2\setminus\overline{B}$}, \\[.4em]
& \text{$w^s$ is radiating}, \\[0.4em]
& \displaystyle w^s = u - u^i \;\,\text{and}\;\, \frac{\partial w^s}{\partial\nu} = \frac{\partial (u - u^i)}{\partial\nu} \quad \text{on $\partial B$}.
\end{align}
$$
The part of the domain where the refractive index \(n\neq1\) models an inhomogeneous medium. For the applications I am interested in, \(n\) is often a (smooth) <a href="https://en.wikipedia.org/wiki/Stochastic_process">random function</a>.
<p>To solve this system, we must be able to solve interior problems of the form
$$
\begin{align}
& \Delta u + k^2nu = f \quad \text{in $B\setminus\overline{D}$}, \\[0.4em]
& u = g \quad \text{on $\partial D$}, \\[0.4em]
& \frac{\partial u}{\partial\nu} = h \quad \text{on $\partial B$},
\end{align}
$$
with a <a href="https://en.wikipedia.org/wiki/Finite_element_method">fine element method</a>, as well as exterior problems of the form
$$
\begin{align}
& \Delta w + k^2w = 0 \quad \text{in $\mathbb{R}^2\setminus\overline{B}$}, \\[0.4em]
& w = g \quad \text{on $\partial B$}, \\[0.4em]
& \text{$w$ is radiating},
\end{align}
$$
with a <a href="https://en.wikipedia.org/wiki/Boundary_element_method">boundary element method</a>.
<h2>Interior problems & finite elements</h2>
<h4>Theory (disk)</h4>
<center>
<img src="/blog/fembem-tik2.png" class="img-responsive">
</center>
<p>Let \(D=\mathcal{D}(0,R)\), the disk of radius \(R>0\) centered at the origin, and consider
\begin{align*}
& \Delta u + k^2nu = f \quad \text{in $D$}, \\[0.4em]
& u = g \quad \text{on $\partial D$},
\end{align*}
for some \(g\) such that \(\Delta g = 0\) in \(D\). Let us define \(\widetilde{u} = u - g\). Then,
$$
\begin{align}
& \Delta \widetilde{u} + k^2n\widetilde{u} = f - k^2ng \quad \text{in $D$}, \\[0.4em]
& \widetilde{u} = 0 \quad \text{on $\partial D$}.
\end{align}
$$
The variational problem is to find \(\widetilde{u}\in H^1_0(D)\) such that
$$
\int_D\left(-\nabla\widetilde{u}\nabla v + k^2n\widetilde{u}v\right) = \int_D (f - k^2ng)v,
$$
for all \(v\in H^1_0(D)\), where
$$
H^1_0(D) = \left\{v \in H^1(D) \,:\, v|_{\partial D} = 0\right\}.
$$
Once the problem is solved for \(\widetilde{u}\), we recover \(u=\widetilde{u}+g\).
</p>
<h4>Experiment (disk)</h4>
<p>Let us take \(R=1\), \(k=2\pi\), and
$$
n(x,y)=1, \quad f(x,y)=x^2, \quad g(x,y)=4.
$$
Using the <a href="https://github.com/matthieuaussal/gypsilab">gypsilab</a> toolbox, we may compute the solution as follows.
<!------- Code ------->
<pre>
R = 1;
k = 2*pi;
n = @(X) 0*X(:, 1) + 1;
f = @(X) X(:, 1).^2;
g = @(X) 0*X(:, 1) + 4;
N = 1000;
mesh = mshDisk(N, R);
D = dom(mesh, 3);
V = fem(mesh, 'P1');
V = dirichlet(V, mesh.bnd);
L = -integral(D, grad(V), grad(V)) ...
+ k^2*integral(D, V, n, V);
F = integral(D, V, @(X) f(X) - k^2*n(X).*g(X));
u = L\F;
</pre>
<img src="/blog/fembem1.jpg" class="img-responsive">
<!------- Code ------->
<p>In <a href="https://www.chebfun.org/">Chebfun</a>, we may use the <code>diskfun.helmholtz</code> code.</p>
<!------- Code ------->
<pre>
F = diskfun(@(x,y) f([x, y]) - k^2*g([x, y]));
BC = @(theta) 0*theta;
u = diskfun.helmholtz(F, k, BC, N);
</pre>
<img src="/blog/fembem2.jpg" class="img-responsive">
<!------- Code ------->
<h4>Theory (annulus)</h4>
<center>
<img src="/blog/fembem-tik3.png" class="img-responsive">
</center>
<p>Let us now consider a harder problem. Let \(D=\mathcal{D}(0, R_1)\) and \(B=\mathcal{D}(0,R_2)\) for some \(R_2>R_1>0\). Define \(C=B\setminus\overline{D}\) and consider
$$
\begin{align}
& \Delta u + k^2nu = f \quad \text{in $C$}, \\[0.4em]
& u = g \quad \text{on $\partial D$}, \\[0.4em]
& \frac{\partial u}{\partial\nu} = h \quad \text{on $\partial B$},
\end{align}
$$
for some \(g\) such that \(\Delta g = 0\) in \(C\) and \(\partial g/\partial\nu = 0\) on \(\partial B\). Let us define \(\widetilde{u} = u - g\). Then,
$$
\begin{align}
& \Delta \widetilde{u} + k^2n\widetilde{u} = f - k^2ng \quad \text{in $C$}, \\[0.4em]
& \widetilde{u} = 0 \quad \text{on $\partial D$}, \\[0.4em]
& \frac{\partial\widetilde{u}}{\partial\nu} = h \quad \text{on $\partial B$}.
\end{align}
$$
The problem is to find \(\widetilde{u}\in H^1_0(C)\) such that
$$
\int_{C}\left(-\nabla\widetilde{u}\nabla v + k^2n\widetilde{u}v\right) = \int_{C} (f - k^2ng)v - \int_{\partial B}hv,
$$
for all \(v\in H^1_0(C)\), where (slight abuse of notation)
$$
H^1_0(C) = \left\{v \in H^1(C) \,:\, v|_{\partial D} = 0\right\}.
$$
Once the problem is solved for \(\widetilde{u}\), we recover \(u=\widetilde{u}+g\).</p>
<h4>Experiment (annulus)</h4>
<p>Let us take \(R_1=0.1\), \(R_2=1\), \(k=2\pi\),
$$
n(x,y)=\cos(y), \quad f(x,y)=x^2,
$$
and \(g(x,y) = h(x,y) = 1\).
</p>
<!------- Code ------->
<pre>
R1 = 0.1;
R2 = 1;
k = 2*pi;
n = @(X) cos(X(:, 2));
f = @(X) X(:, 1).^2;
g = @(X) 0*X(:, 1) + 1;
h = @(X) 0*X(:, 1) + 1;
N = 2000;
mesh = mshAnnulus(N, R1, R2);
C = dom(mesh, 3);
[dD, dB] = bndAnnulus(mesh.bnd);
dB = dom(dB, 3);
V = fem(mesh, 'P1');
V = dirichlet(V, dD);
L = -integral(C, grad(V), grad(V)) ...
+ k^2*integral(C, V, n, V);
F = integral(C, V, @(X) f(X) - k^2*n(X).*g(X)) ...
- integral(dB, V, h);
u = L\F;
</pre>
<img src="/blog/fembem3.jpg" class="img-responsive">
<!------- Code ------->
<h2>Exterior problems & boundary elements</h2>
<h4>Theory</h4>
<center>
<img src="/blog/fembem-tik4.png" class="img-responsive">
</center>
<p>Let \(B=\mathcal{D}(0,R)\) and consider
$$
\begin{align}
& \Delta w + k^2w = 0 \quad \text{in $\mathbb{R}^2\setminus\overline{B}$}, \\[0.4em]
& w = g \quad \text{on $\partial B$}, \quad \text{$w$ is radiating},
\end{align}
$$
for some function \(g\). Let us look for \(w\) as a single-layer potential,
$$
w(\boldsymbol{x}) = \int_{\partial B} G(\boldsymbol{x},\boldsymbol{y})\varphi(\boldsymbol{y})dS(\boldsymbol{y}) = (\mathcal{S}\varphi)(\boldsymbol{x}).
$$
The problem is to find \(\varphi\in H^{-1/2}(\partial B)\) such that
$$
\int_{\partial B}(\mathcal{S}\varphi)(\boldsymbol{x})\psi(\boldsymbol{x})dS(\boldsymbol{x}) = \int_{\partial B}g(\boldsymbol{x})\psi(\boldsymbol{x})dS(\boldsymbol{x}),
$$
for all \(\psi\in H^{-1/2}(\partial B)\).</p>
<h4>Experimemt</h4>
<p>Let us take \(R=0.5\), \(k=2\pi\), and
$$
g(x,y) = -e^{ikx}.
$$
</p>
<!------- Code ------->
<pre>
R = 0.5;
k = 2*pi;
g = @(X) -exp(1i*k*X(:,1));
G = @(X,Y) femGreenKernel(X, Y, '[H0(kr)]', k);
N = 100;
mesh = mshCircle(N, R);
dB = dom(mesh, 3);
W = fem(mesh, 'P1');
S = (1i/4)*integral(dB, dB, W, G, W);
S = S - 1/(2*pi)*regularize(dB, dB, W, ...
'[log(r)]', W);
F = integral(dB, W, g);
phi = S\F;
</pre>
<img src="/blog/fembem4.jpg" class="img-responsive">
<!------- Code ------->
<h2>Finite-boundary element coupling</h2>
<h4>Theory</h4>
<center>
<img src="/blog/fembem-tik1.png" class="img-responsive">
</center>
<p>We now come back to the initial problem:
$$
\begin{align}
& \Delta u + k^2nu = 0 \quad \text{in $B\setminus\overline{D}$}, \\[.4em]
& u = 0 \quad \text{on $\partial D$}, \\[0.4em]
& \Delta w^s + k^2w^s = 0 \quad \text{in $\mathbb{R}^2\setminus\overline{B}$}, \\[.4em]
& \text{$w^s$ is radiating}, \\[0.4em]
& \displaystyle w^s = u - u^i \;\,\text{and}\;\, \frac{\partial w^s}{\partial\nu} = \frac{\partial (u - u^i)}{\partial\nu} \quad \text{on $\partial B$}.
\end{align}
$$
Let \(D=\mathcal{D}(0, R_1)\), \(B=\mathcal{D}(0,R_2)\) for some \(R_2>R_1>0\), and \(C=B\setminus\overline{D}\). Combining the theory of the previous sections for interior and exterior problems, we define a bilinear form \(A\) with entries
$$
\begin{pmatrix}
\displaystyle \int_{\partial B}(\mathcal{S}\varphi)\psi & \displaystyle -\int_{\partial B}u\psi \\[.4em]
\displaystyle \int_{\partial B}\left[\left(\mathcal{D}^* - \frac{\mathcal{I}}{2}\right)\varphi\right]v & \displaystyle \int_C\left(-\nabla u\nabla v + k^2nuv\right)
\end{pmatrix},
$$
while the linear form \(F\) has entries
$$
\begin{pmatrix}
\displaystyle -\int_{\partial B} u^i\psi \\[.4em]
\displaystyle -\int_{\partial B}\frac{\partial u^i}{\partial \nu}v
\end{pmatrix},
$$
for \(\varphi,\psi\in H^{-1/2}(\partial B)\) and \(u,v\in H^1_0(C)\). We may write the system as
$$
\left\{
\begin{array}{l}
A_{11}\varphi + A_{12} u = F_1, \\[.4em]
A_{21}\varphi + A_{22} u = F_2.
\end{array}
\right.
$$
This yields
$$
\left\{
\begin{array}{l}
-A_{21}A_{11}^{-1}A_{12}u + A_{22}u = F_2 - A_{21}A_{11}^{-1}F_1, \\[.4em]
\varphi = A_{11}^{-1}(F_1 - A_{12}u).
\end{array}
\right.
$$
We use GMRES for the first equation (<code>mgcr</code> in <a href="https://github.com/matthieuaussal/gypsilab">gypsilab</a>).</p>
<h4>Experiment</h4>
<p>Let us take \(R_1=0.5\), \(R_2=1\), \(k=2\pi\), and
$$
u^i(x,y) = e^{ikx}, \quad n(x,y)=1.
$$
Since we take \(n=1\) everywhere, we should recover the numerical solution of the previous section. The MATLAB code is a little bit more involved. We are going to break it down. We first set up the parameters.</p>
<!------- Code ------->
<pre>
R1 = 0.5;
R2 = 1;
k = 2*pi;
lambda = 2*pi/k;
ui = @(X) exp(1i*k*X(:, 1));
dui = @(X) 1i*k*X(:,1)/R2.*ui(X);
G = @(X,Y) femGreenKernel(X, Y, '[H0(kr)]', k);
dG{1} = @(X,Y) femGreenKernel(X, Y, ...
'grady[H0(kr)]1', k);
dG{2} = @(X,Y) femGreenKernel(X, Y, ...
'grady[H0(kr)]2', k);
dG{3} = @(X,Y) femGreenKernel(X, Y, ...
'grady[H0(kr)]3', k);
n = @(X) 0*X(:, 1) + 1;
</pre>
<!------- Code ------->
<p>We define the mesh and the finite/boundary elements.</p>
<!------- Code ------->
<pre>
N = 1000;
mesh = mshAnnulus(N, R1, R2);
C = dom(mesh, 3);
[dD, dB] = bndAnnulus(mesh.bnd);
V = fem(mesh, 'P1');
V = dirichlet(V, dD);
W = fem(dB, 'P1');
dB = dom(dB, 3);
</pre>
<!------- Code ------->
<p>This is the tricky part where we define the projection between finite and boundary elements, as well as the operators. (The negative sign in the double-layer variable <code>D</code> is to adjust the direction of the normal.)</p>
<!------- Code ------->
<pre>
[~, IW, IV] = intersect(W.unk, V.unk, 'rows');
P = sparse(IW, IV, 1, size(W.unk, 1), size(V.unk, 1));
L = -integral(C, grad(V), grad(V)) ...
+ k^2*integral(C, V, n, V);
S = (1i/4)*integral(dB, dB, W, G, W);
S = S - 1/(2*pi)*regularize(dB, dB, W, ...
'[log(r)]', W);
D = (1i/4)*integral(dB, dB, W, dG, ntimes(W));
D = D - 1/(2*pi)*regularize(dB, dB, W, ...
'grady[log(r)]', ntimes(W));
D = -D;
I = integral(dB, W, W);
</pre>
<!------- Code ------->
<p>We assemble the system and solve it.</p>
<!------- Code ------->
<pre>
A11 = S;
A21 = P'*(D.' - I/2);
A12 = -I*P;
A22 = L;
F1 = -integral(dB, W, ui);
F2 = -integral(dB, V, dui);
fun = @(V) -A21*(A11\(A12*V)) + A22*V;
u = mgcr(fun, F2 - A21*(A11\F1), [], 1e-3, 1000);
phi = A11\(F1 - A12*u);
</pre>
<img src="/blog/fembem5.jpg" class="img-responsive">
<img src="/blog/fembem6.jpg" class="img-responsive">
<!------- Code ------->
<hr>
<h4>Blog posts about boundary elements</h4>
<p>2023 <a href="2023-10-02.html">Strongly singular integrals over curved elements</a></p>
<p>2023 <a href="2023-02-12.html">Finite-boundary element coupling in MATLAB</a></p>
<p>2021 <a href="2021-11-25.html">Weakly singular integrals over curved elements</a></p>
</div>
</div>
</div>
</section>
<!-- Footer Section -->
<footer>
<div class="container">
<div class="row">
<div class="col-md-4 col-md-offset-1 col-sm-6">
<h3>Contact</h3>
<p style="color:white"><i class="fa fa-send-o"></i>[email protected]</p><br>
<a href="https://www.inria.fr/en"><img src="/images/inria.png" class="img-responsive"></a>
</div>
<div class="col-md-4 col-md-offset-1 col-sm-6">
<h3 style="opacity:0;">Contact</h3>
<p style="color:white">Copyright © 2024 Hadrien Montanelli</p><br>
<a href="https://www.ip-paris.fr/en"><img src="/images/polytechnique.png" class="img-responsive"></a>
</div>
<div class="clearfix col-md-12 col-sm-12">
<hr style="color:white;">
</div>
<div class="col-md-12 col-sm-12">
<ul class="social-icon">
<li><a href="https://github.com/Hadrien-Montanelli" class="fa-brands fa-github" style="font-size:30px;color:white"></a></li>
<li><a href="https://scholar.google.com/citations?user=Bjmkfe8AAAAJ&hl=en&oi=sra/" class="fa-brands fa-google" style="font-size:30px;color:white"></a></li>
<li><a href="https://www.linkedin.com/in/hadrien-montanelli/" class="fa-brands fa-linkedin" style="font-size:30px;color:white"></a></li>
<li><a href="https://orcid.org/0009-0005-1742-9828" class="fa-brands fa-orcid" style="font-size:30px;color:white"></a></li>
<li><a href="https://twitter.com/drmontanelli" class="fa-brands fa-x-twitter" style="font-size:30px;color:white;"></a></li>
</ul>
</div>
</div>
</div>
</footer>
<!-- Back top -->
<a href="#back-top" class="go-top"><i class="fa fa-angle-up"></i></a>
<!-- SCRIPTS -->
<script src="js/jquery.js"></script>
<script src="js/bootstrap.min.js"></script>
<script src="js/particles.min.js"></script>
<script src="js/app.js"></script>
<script src="js/jquery.parallax.js"></script>
<script src="js/smoothscroll.js"></script>
<script src="js/custom.js"></script>
<script src="https://polyfill.io/v3/polyfill.min.js?features=es6"></script>
<script id="MathJax-script" async src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js"></script>
</body>
</html>