# American Institute of Mathematical Sciences

August  2019, 13(4): 827-862. doi: 10.3934/ipi.2019038

## Nash strategies for the inverse inclusion Cauchy-Stokes problem

 1 Université Cȏte d'Azur, Inria, CNRS, LJAD, UMR 7351, Parc Valrose, Nice 06108, France 2 Université de Tunis El Manar, Ecole Nationale d'Ingénieurs de Tunis, LAMSIN, BP 37, 1002 Tunis Belvedere, Tunisia

* Corresponding author: A. Habbal

Received  October 2018 Revised  March 2019 Published  May 2019

We introduce a new algorithm to solve the problem of detecting unknown cavities immersed in a stationary viscous fluid, using partial boundary measurements. The considered fluid obeys a steady Stokes regime, the cavities are inclusions and the boundary measurements are a single compatible pair of Dirichlet and Neumann data, available only on a partial accessible part of the whole boundary. This inverse inclusion Cauchy-Stokes problem is ill-posed for both the cavities and missing data reconstructions, and designing stable and efficient algorithms is not straightforward. We reformulate the problem as a three-player Nash game. Thanks to an identifiability result derived for the Cauchy-Stokes inclusion problem, it is enough to set up two Stokes boundary value problems, then use them as state equations. The Nash game is then set between 3 players, the two first targeting the data completion while the third one targets the inclusion detection. We used a level-set approach to get rid of the tricky control dependence of functional spaces, and we provided the third player with the level-set function as strategy, with a cost functional of Kohn-Vogelius type. We propose an original algorithm, which we implemented using Freefem++. We present 2D numerical experiments for three different test-cases.The obtained results corroborate the efficiency of our 3-player Nash game approach to solve parameter or shape identification for Cauchy problems.

Citation: Abderrahmane Habbal, Moez Kallel, Marwa Ouni. Nash strategies for the inverse inclusion Cauchy-Stokes problem. Inverse Problems & Imaging, 2019, 13 (4) : 827-862. doi: 10.3934/ipi.2019038
##### References:

show all references

##### References:
An example of the geometric configuration of the problem : the whole domain including cavities is denoted by $\Omega$. It contains an inclusion $\omega^*$. The boundary of $\Omega$ is composed of $\Gamma_{\!\! c}$, an accessible part where over-specified data are available, and an inaccessible part $\Gamma_{\!\!i }$ where the data are missing
Different situations
Test case A. Reconstruction of the inclusion shape and missing boundary data with noise free Dirichlet data over $\Gamma_{ c}$. (a) initial contour is $\phi^{(0)}_1$ (b) exact inclusion shape -green line- and computed one - blue dashed- (c) exact -line- and computed -dashed line- first component of the velocity over $\Gamma_{i }$ (d) exact -line- and computed -dashed line- second component of the velocity over $\Gamma_{i }$ (e) exact -line- and computed -dashed line- first component of the normal stress over $\Gamma_{i }$ (f) exact -line- and computed -dashed line- second component of the normal stress over $\Gamma_{i }$
Test case A. Reconstruction of the inclusion shape and missing boundary data with noise free Dirichlet data over $\Gamma_{ c}$.(a) initial contour is $\phi^{(0)}_2$ (b) exact inclusion shape -green line- and computed one - blue dashed- (c) exact -line- and computed -dashed line- first component of the velocity over $\Gamma_{i }$ (d) exact -line- and computed -dashed line- second component of the velocity over $\Gamma_{i }$ (e) exact -line- and computed -dashed line- first component of the normal stress over $\Gamma_{i }$ (f) exact -line- and computed -dashed line- second component of the normal stress over $\Gamma_{i }$
Test case A. Reconstruction of the inclusion shape and missing boundary data with noisy Dirichlet data over $\Gamma_{ c}$ with noise levels $\sigma = \lbrace 1\%, 3\%, 5\% \rbrace$.(a) initial contour is $\phi^{(0)}_2$ (b) exact inclusion shape -green line- and computed ones for different noise levels (c) exact and computed first components of the velocity over $\Gamma_{i }$ (d) exact and computed second components of the velocity over $\Gamma_{i }$ (e) exact and computed first components of the normal stress over $\Gamma_{i }$ (f) exact and computed second components of the normal stress over $\Gamma_{i }$
Test case B. Reconstruction of the inclusion shape and missing boundary data with noise free Dirichlet data over $\Gamma_{ c}$. (a) initial contour is $\phi^{(0)}_2$ (b) exact inclusion shape -green line- and computed one - blue dashed- (c) exact -line- and computed -dashed line- first component of the velocity over $\Gamma_{i }$ (d) exact -line- and computed -dashed line- second component of the velocity over $\Gamma_{i }$ (e) exact -line- and computed -dashed line- first component of the normal stress over $\Gamma_{i }$ (f) exact -line- and computed -dashed line- second component of the normal stress over $\Gamma_{i }$
Test case B. Reconstruction of the inclusion shape and missing boundary data with noisy Dirichlet data over $\Gamma_{ c}$ with levels $\sigma = \lbrace 1\%, 3\%, 5\% \rbrace$.(a) initial contour is $\phi^{(0)}_2$ (b) exact inclusion shape -green line- and computed ones for different noise levels (c) exact and computed first components of the velocity over $\Gamma_{i }$ (d) exact and computed second components of the velocity over $\Gamma_{i }$ (e) exact and computed first components of the normal stress over $\Gamma_{i }$ (f) exact and computed second components of the normal stress over $\Gamma_{i }$
Test case C. Reconstruction of the inclusion shape and missing boundary data with noise free Dirichlet data over $\Gamma_{ c}$. (a) initial contour is $\phi^{(0)}_2$ (b) exact inclusion shape -green line- and computed one - blue dashed- (c) exact -line- and computed -dashed line- first component of the velocity over $\Gamma_{i }$ (d) exact -line- and computed -dashed line- second component of the velocity over $\Gamma_{i }$ (e) exact -line- and computed -dashed line- first component of the normal stress over $\Gamma_{i }$ (f) exact -line- and computed -dashed line- second component of the normal stress over $\Gamma_{i }$
Test case C. Reconstruction of the inclusion shape and missing boundary data with noisy Dirichlet data over $\Gamma_{ c}$ with levels $\sigma = \lbrace 1\%, 3\%, 5\% \rbrace$.(a) initial contour is $\phi^{(0)}_2$ (b) exact inclusion shape -green line- and computed ones for different noise levels (c) exact and computed first components of the velocity over $\Gamma_{i }$ (d) exact and computed second components of the velocity over $\Gamma_{i }$ (e) exact and computed first components of the normal stress over $\Gamma_{i }$ (f) exact and computed second components of the normal stress over $\Gamma_{i }$
Test case A. (Left) sensitivity of the reconstruction w.r.t. the mesh size (on abscissae : the number of F.E. nodes on the boundary $\partial \Omega$). (Right) sensitivity of the reconstruction w.r.t. the distance to the inaccessible boundary $\Gamma_{i }$ (on abscissae : the distance of the center of the circular inclusion from $\Gamma_{i }$)
Test case C. (Left) Mesh used for solving the direct problem with the P1bubble-P1 finite element, in order to construct the synthetic data. (Right) Mesh used for solving the coupled inverse problem with P2-P1 finite element, using the P1 bubble-P1 synthetic data
Assessing Inverse-Crime-Free reconstruction. Test case C. Top: initial and optimal contour. Middle: the two components of the velocity on $\Gamma_i$. Bottom: the two components of the normal stress on $\Gamma_i$ ($err_D = 0.0615048,$ $err_N = 0.124296,$ and $err_O = 0.113156)$
Test-case A. $L^2$ relative errors on missing data on $\Gamma_i$ (on Dirichlet and Neumann data), and the error between the reconstructed and the real shape of the inclusion for various noise levels
 Noise level $\sigma=0\%$ $\sigma=1\%$ $\sigma=3\%$ $\sigma=5\%$ $err_D$ 0.010 0.015 0.039 0.063 $err_N$ 0.031 0.033 0.051 0.07 $err_O$ 0.032 0.043 0.066 0.117
 Noise level $\sigma=0\%$ $\sigma=1\%$ $\sigma=3\%$ $\sigma=5\%$ $err_D$ 0.010 0.015 0.039 0.063 $err_N$ 0.031 0.033 0.051 0.07 $err_O$ 0.032 0.043 0.066 0.117
Test-case C. $L^2$-errors on missing data over $\Gamma_{i }$ (on Dirichlet and Neumann data), and the error between the reconstructed and the real shape for various noise levels
 Noise level $\sigma=0\%$ $\sigma=1\%$ $\sigma=3\%$ $\sigma=5\%$ $err_D$ 0.042 0.044 0.046 0.08 $err_N$ 0.095 0.1 0.13 0.16 $err_O$ 0.099 0.11 0.13 0.15
 Noise level $\sigma=0\%$ $\sigma=1\%$ $\sigma=3\%$ $\sigma=5\%$ $err_D$ 0.042 0.044 0.046 0.08 $err_N$ 0.095 0.1 0.13 0.16 $err_O$ 0.099 0.11 0.13 0.15
Relative errors on the reconstructed missing data and inclusion shape for the Stokes problem (with noise free measurements), compared for a classical Nash algorithm and Algorithm 2: (left) test-case A (right) test-case C
 Case A Classical algorithm Algorithm 2 $err_D$ 0.058 0.033 $err_N$ 0.106 0.032 $err_O$ 0.358 0.140 Case C Classical algorithm Algorithm 2 $err_D$ 0.067 0.058 $err_N$ 0.208 0.122 $err_O$ 0.566 0.167
 Case A Classical algorithm Algorithm 2 $err_D$ 0.058 0.033 $err_N$ 0.106 0.032 $err_O$ 0.358 0.140 Case C Classical algorithm Algorithm 2 $err_D$ 0.067 0.058 $err_N$ 0.208 0.122 $err_O$ 0.566 0.167
 [1] Moez Kallel, Maher Moakher, Anis Theljani. The Cauchy problem for a nonlinear elliptic equation: Nash-game approach and application to image inpainting. Inverse Problems & Imaging, 2015, 9 (3) : 853-874. doi: 10.3934/ipi.2015.9.853 [2] Jaroslav Haslinger, Raino A. E. Mäkinen, Jan Stebel. Shape optimization for Stokes problem with threshold slip boundary conditions. Discrete & Continuous Dynamical Systems - S, 2017, 10 (6) : 1281-1301. doi: 10.3934/dcdss.2017069 [3] John Sebastian Simon, Hirofumi Notsu. A shape optimization problem constrained with the Stokes equations to address maximization of vortices. Evolution Equations & Control Theory, 2022  doi: 10.3934/eect.2022003 [4] Roman Chapko, B. Tomas Johansson. Integral equations for biharmonic data completion. Inverse Problems & Imaging, 2019, 13 (5) : 1095-1111. doi: 10.3934/ipi.2019049 [5] Marion Darbas, Jérémy Heleine, Stephanie Lohrengel. Numerical resolution by the quasi-reversibility method of a data completion problem for Maxwell's equations. Inverse Problems & Imaging, 2020, 14 (6) : 1107-1133. doi: 10.3934/ipi.2020056 [6] Yang Han. On the cauchy problem for the coupled Klein Gordon Schrödinger system with rough data. Discrete & Continuous Dynamical Systems, 2005, 12 (2) : 233-242. doi: 10.3934/dcds.2005.12.233 [7] Li Liang. Increasing stability for the inverse problem of the Schrödinger equation with the partial Cauchy data. Inverse Problems & Imaging, 2015, 9 (2) : 469-478. doi: 10.3934/ipi.2015.9.469 [8] Hiroyuki Hirayama, Mamoru Okamoto. Random data Cauchy problem for the nonlinear Schrödinger equation with derivative nonlinearity. Discrete & Continuous Dynamical Systems, 2016, 36 (12) : 6943-6974. doi: 10.3934/dcds.2016102 [9] Junichi Minagawa. On the uniqueness of Nash equilibrium in strategic-form games. Journal of Dynamics & Games, 2020, 7 (2) : 97-104. doi: 10.3934/jdg.2020006 [10] Yi Yang, Jianwei Ma, Stanley Osher. Seismic data reconstruction via matrix completion. Inverse Problems & Imaging, 2013, 7 (4) : 1379-1392. doi: 10.3934/ipi.2013.7.1379 [11] Julius Fergy T. Rabago, Jerico B. Bacani. Shape optimization approach for solving the Bernoulli problem by tracking the Neumann data: A Lagrangian formulation. Communications on Pure & Applied Analysis, 2018, 17 (6) : 2683-2702. doi: 10.3934/cpaa.2018127 [12] Xiaoqiang Dai, Shaohua Chen. Global well-posedness for the Cauchy problem of generalized Boussinesq equations in the control problem regarding initial data. Discrete & Continuous Dynamical Systems - S, 2021, 14 (12) : 4201-4211. doi: 10.3934/dcdss.2021114 [13] Dean A. Carlson. Finding open-loop Nash equilibrium for variational games. Conference Publications, 2005, 2005 (Special) : 153-163. doi: 10.3934/proc.2005.2005.153 [14] Athanasios Kehagias. A note on the Nash equilibria of some multi-player reachability/safety games. Journal of Dynamics & Games, 2022, 9 (1) : 117-122. doi: 10.3934/jdg.2021028 [15] Kazuhiro Ishige. On the existence of solutions of the Cauchy problem for porous medium equations with radon measure as initial data. Discrete & Continuous Dynamical Systems, 1995, 1 (4) : 521-546. doi: 10.3934/dcds.1995.1.521 [16] Angkana Rüland, Eva Sincich. Lipschitz stability for the finite dimensional fractional Calderón problem with finite Cauchy data. Inverse Problems & Imaging, 2019, 13 (5) : 1023-1044. doi: 10.3934/ipi.2019046 [17] Francesca Crispo, Paolo Maremonti. A remark on the partial regularity of a suitable weak solution to the Navier-Stokes Cauchy problem. Discrete & Continuous Dynamical Systems, 2017, 37 (3) : 1283-1294. doi: 10.3934/dcds.2017053 [18] Ruihong Ji, Yongfu Wang. Mass concentration phenomenon to the 2D Cauchy problem of the compressible Navier-Stokes equations. Discrete & Continuous Dynamical Systems, 2019, 39 (2) : 1117-1133. doi: 10.3934/dcds.2019047 [19] Zhilei Liang, Jiangyu Shuai. Existence of strong solution for the Cauchy problem of fully compressible Navier-Stokes equations in two dimensions. Discrete & Continuous Dynamical Systems - B, 2021, 26 (10) : 5383-5405. doi: 10.3934/dcdsb.2020348 [20] Qun Liu, Lihua Fu, Meng Zhang, Wanjuan Zhang. Two-dimensional seismic data reconstruction using patch tensor completion. Inverse Problems & Imaging, 2020, 14 (6) : 985-1000. doi: 10.3934/ipi.2020052

2020 Impact Factor: 1.639