Contents

# Shape optimization

If the cost function is potential energy, Corollaries 2.13 , 2.18 can be used directly. We now consider the case where the cost function is mean compliance , that is $$F_c(u^{\varphi _t})=\int _{\Omega (t)}fu^{\varphi _t}dx+\int _{\Gamma _N(t)}gu^{\varphi _t} ds \tag{3.1}$$ and the problem (2.1) is linear ((2.28) ), then we have $$F_c(u^{\varphi _t})=-2\mathcal{E}(u^{\varphi _t};f,g,\Omega (t)) \tag{3.2}$$ Hence the shape sensitivity of mean compliance expressed by the GJ-integral is $$\left .\frac{d}{dt}F_c(u^{\varphi _t})\right |_{t=0} =2\left \{R_{\Omega }(u,\mu _{\varphi })+\int _{\partial \Omega }fu(\mu _{\varphi }\cdot n)ds\right \} \tag{3.3}$$ The shape derivative $u'^{\varphi }$ is defined by $$u'^{\varphi }(x)=\dot{u}^{\varphi }(x)-\nabla u(x)\cdot \mu _{\varphi }(x)$$ For $u\in W^{1,p}(\Omega ), v\in W^{1,q}(\mathbb{R}^d), \frac{1}{p}+\frac{1}{q}=1,p\ge 1, q\ge 1$, the following holds $$\left .\frac{d}{dt}\int _{\Omega (t)}u^{\varphi _t}v\, dx\right |_{t=0} =\int _{\Omega }u'^{\varphi }v\,dx+\int _{\partial \Omega }uv(\mu _{\varphi }\cdot n)ds$$ If the variational problem (2.1) is linear (2.28) and Corollary 2.13 holds, then we have the following.
[O-K00] Assume the problem (2.1) is linear. For a given function $\Phi \in H^1(\mathbb{R}^d;\mathbb{R}^m)$, let $u_{\Phi }^{\varphi _t}$ be the solution of the problem $$\mathcal{E}(u_{\Phi }^{\varphi _t};\Phi ,0,\Omega (t)) =\min _{v\in V^{\varphi _t}(\Omega )}\mathcal{E}(v;\Phi ,0,\Omega (t))$$ and consider $\delta R_{\omega }(u,\mu )[w]=\lim _{\epsilon →0}\epsilon ^{-1}\left \{R_{\omega }(u+\epsilon w,\mu )-R_{\omega }(u,\mu )\right \}$. Then we have $$\left .\frac{d}{dt}\int _{\Omega }\Phi u^{\varphi _t}\, dx\right |_{t=0} =\delta R_{\Omega }(u,\mu _{\varphi })[u_{\Phi }]+\int _{\partial \Omega }fu_{\Phi }(\mu _{\varphi }\cdot n)\,ds \tag{3.4}$$

Notes
[Corollary 2.81, Sok92], The mapping $t \mapsto u^{\varphi _t}\circ \varphi _t$ is weakly differentiable in $L^2(\Omega )$) and the weak material derivative 3.4 under Poisson with Dirichlet boundary condition. (see also [Corollary 3.1, N-S13]).
By defining $\delta P_{\omega }(u,\mu )[w]$ like $\delta R_{\omega }(u,\mu )[w]$, we get the following.
Corollary 3.2
If $u$ and $w$ is regular inside $\omega$, then $\delta P_{\omega }(u,\mu )[w]+\delta R_{\omega }(u,\mu )[w]=0$ for all $\mu \in W^{1,\infty }(\mathbb{R}^d;\mathbb{R}^d)$.
Consider the Poisson equation with Dirichlet boundary condition and its Green's function $\Delta _xG^{\varphi _t}(x,y)=-\delta (x-y)$($\delta$ is delta function); then we have $u^{\varphi _t}(z)=\int _{\Omega }G^{\varphi _t}(z,x)f(x)dy, u_{\Phi }^{\varphi _t}(z)=\int _{\Omega }G^{\varphi _t}(z,y)\Phi (y)dy$. From (2.46) , Theorem 3.1 and Corollary 3.2 we obtain Hadamard variational formula [Had68] $$G'^{\varphi }(x,y)= \int _{\partial \Omega }\frac{\partial G}{\partial n_z}(x,z) \frac{\partial G}{\partial n_z}(y,z)(\mu (z)\cdot n(z))ds_z. \tag{3.5}$$ Thus Theorem 3.1 is a generalization of Hadamard variational formula. Here we consider the const function by $\widehat{g}(z) \in C^2(\mathbb{R}^m)$ $$F_g(u^{\varphi _t})=\int _{\Omega (t)}\widehat{g}(u^{\varphi _t})\, dx, \tag{3.6}$$ When $\widehat{g}(z)=(z-u_{d})^2$, the cost function $F_g$ is least mean-square error. From Theorem 3.1 , $u'^{\varphi }$ belongs to $L^2(\Omega ;\mathbb{R}^m)$, we have $$\left .\frac{d}{dt}F_g(u^{\varphi _t})\right |_{t=0} =\int _{\Omega }\nabla _z\widehat{g}(u)u'^{\varphi }dx+\int _{\partial \Omega } \widehat{g}(u)(\mu _{\varphi }\cdot n)\, ds. \tag{3.7}$$ Let $u_g$ be the solution of the problem $$\mathcal{E}(u_g;\nabla _z \widehat{g}(u),0,\Omega ) =\min _{v\in V(\Omega )} \mathcal{E}(v;\nabla _z \widehat{g}(u),0,\Omega ) \tag{3.8}$$ Then from Theorem 3.1 , the first term in the right-hand side of (3.7) is written by $\delta R(u,\mu _{\varphi })[u_g]$ and we arrive at \begin{eqnarray} \left .\frac{d}{dt}F_g(u^{\varphi _t})\right |_{t=0} &=&\delta R_{\Omega }(u,\mu _{\varphi })[u_g]\notag \\ &&+\int _{\partial \Omega } (fu_g+\widehat{g}(u)-\nabla _z\widehat{g}(u)u)(\mu _{\varphi }\cdot n)ds \tag{3.9} \end{eqnarray} The method of erasing $u'^{\varphi }$ with $u_g$ is called the adjoint variable method'' and it is often used in shape optimization.

Information about the page: The current position is painted circle in the diagram below. Blue is the main MaKR and orange is a duplicate for MaKR's public use, where dashed line means the connection to the private area The dashed lines are only connections to main MaKR.

[A-P06] G. Allaire and O. Pantz, Structural optimization with FreeFem++, Struct. Multidiscip. Opt, 32 (2006), 173--181.
[Al07] G. Allaire, Conception optimale de structures, Springer, 2007.
[Az94] H. Azegami, Solution to domain optimization problems, Trans. Japan Soc. Mech. Engrs. Series A, 60, No.574 (1994), 1479--1486. (in Japanese)
[A-W96] H. Azegami and Z. Wu, Domain optimization analysis in linear elastic problems: Approach using traction method, JSME Inter. J. Series A, 39 (1996), 272--278.
[Az17] H. Azegami. Solution of shape optimization problem and its application to product design, Mathematical Analysis of Continuum Mechanics and Industrial Applications, Springer, 2017, 83--98.
[B-S04] M.P. Bends{\o }e and O. Sigmund, Topology optimization: theory, methods, and applications, Springer, 2004.
[Bu04] H.D. Bui, Fracture mechanics -- Inverse problems and solutions, Springer, 2006.
[Ch67] G.P. Cherepanov, On crack propagation in continuous media, Prikl. Math. Mekh., 31 (1967), 476--493.
[Cir88] P.G. Ciarlet, Mathematical elasticity: Three-dimensional elasticity, North-Holland, 1988.
[Co85] R. Correa and A. Seeger, Directional derivative of a minimax function. Nonlinear Anal., 9(1985), 13--22.
[D-Z88] M.C. Delfour and J.-P. Zolésio, Shape sensitivity analysis via min max differentiability, SIAM J. Control and Optim., 26(1988), 834--862.
[D-D81] Ph. Destuynder and M. Djaoua, Sur une interprétation de l'intégrale de Rice en théorie de la rupture fragile. Math. Meth. in Appl. Sci., 3 (1981), 70--87.
[E-G04] A. Em and J.-L. Guermond, Theory and practice of finite elements, Springer, 2004.
[Es56] J.D. Eshelby, The Continuum theory of lattice defects, Solid State Physics, 3 (1956), 79--144.
[F-O78] D. Fujiwara and S. Ozawa, The Hadamard variational formula for the Green functions of some normal elliptic boundary value problems, Proc. Japan Acad., 54 (1978), 215--220.
[G-S52] P.R. Garabedian and M. Schiffer, Convexity of domain functionals, J.Anal.Math., 2 (1952), 281--368.
[Gr21] A.A. Griffith, The phenomena of rupture and flow in solids, Phil. Trans. Roy. Soc. London, Series A 221 (1921), 163--198.
[Gr24] A.A. Griffith, The theory of rupture, Proc. 1st.Intern. Congr. Appl. Mech., Delft (1924) 55--63.
[Gr85] P. Grisvard, Elliptic problems in nonsmooth domains, Pitman, 1985.
[Gr92] P. Grisvard, Singularities in boundary value problems, Springer, 1992.
[Had68] J. Hadamard, Mémoire sur un problème d'analyse relatif à l'équilibre des plaques élastiques encastrées, Mémoire des savants étragers, 33 (1907), 515--629.
[Hau86] E.J. Haug, K.K. Choi and V. Komkov, Design sensitivity analysis of structural systems, Academic Press, 1986.
[ffempp] F. Hecht, New development in freefem++. J. Numer. Math. 20 (2012), 251--265. 65Y15, (FreeFem++ URL:http://www.freefem.org)
[Kato] T. Kato, Perturbation theory for linear operators, Springer, 1980.
[K-W06] M. Kimura. and I. Wakano, New mathematical approach to the energy release rate in crack extension, Trans. Japan Soc. Indust. Appl. Math., 16(2006) 345--358. (in Japanese) \bibitem {K-W11} M. Kimura and I. Wakano, Shape derivative of potential energy and energy release rate in rracture mechanics, J. Math-for-industry, 3A (2011), 21--31.
[Kne05] D. Knees, Regularity results for quasilinear elliptic systems of power-law growth in nonsmooth domains: boundary, transmission and crack problems. PhD thesis, Universität Stuttgart, 2005. http://elib.uni-stuttgart.de/opus/volltexte/2005/2191/.
[Ko06] V.A. Kovtunenko, Primal-dual methods of shape sensitivity analysis for curvilinear cracks with nonpenetration, IMA Jour. Appl. Math. 71 (2006), 635--657.
[K-O18] V.A. Kovtunenko and K. Ohtsuka, Shape differentiability of Lagrangians and application to stokes problem, SIAM J. Control Optim. 56 (2018), 3668--3684.
[M-P01] B. Mohammadi and O. Pironneau, Applied shape optimization for fluids. Oxford University Press, 2001.
[Na94] S.Nazarov and B.A.Plamenevsky, Elliptic problems in domains with piecewise smooth boundaries, de Gruyter Expositions in Mathematics 13. Walter de Gruyter \& Co., 1994.
[Nec67] J. Nečas, Direct methods in the theory of elliptic equations, Springer, 2012. Translated from Méthodes directes en théorie des équations elliptiques, 1967, Masson''.
[Noe18] E. Noether, Invariante variationsprobleme, göttinger nachrichten, Mathematisch-Physikalische Klasse (1918), 235--257.
[N-S13] A.A. Novotny and J. Sokolowski, Topological derivatives in shape optimization, Springer, 2013.
[Oh81] K. Ohtsuka, Generalized J-integral and three dimensional fracture mechanics I, Hiroshima Math. J., 11(1981), 21--52.
[Oh85] K. Ohtsuka, Generalized J-integral and its applications. I. -- Basic theory, Japan J. Appl. Math., 2 (1985), 329--350.
[O-K00] K. Ohtsuka and A. Khludnev, Generalized J-integral method for sensitivity analysis of static shape design, Control \& Cybernetics, 29 (2000), 513--533.
[Oh02] K. Ohtsuka, Comparison of criteria on the direction of crack extension, J. Comput. Appl. Math., 149 (2002), 335--339.
[Oh02-2] K. Ohtsuka, Theoretical and numerical analysis on 3-dimensional brittle fracture, Mathematical Modeling and Numerical Simulation in Continuum Mechanics, Springer, 2002, 233--251.
[Oh09] K. Ohtsuka, Criterion for stable/unstable quasi-static crack extension by extended griffith energy balance theory, Theor. Appl. Mech. Japan, 57 (2009), 25--32.
[Oh12] K. Ohtsuka, Shape optimization for partial differential equations/system with mixed boundary conditions, RIMS K\^oky\^uroku 1791 (2012), 172--181.
[OT-K12] K. Ohtsuka and M. Kimura, Differentiability of potential energies with a parameter and shape sensitivity analysis for nonlinear case: the p-Poisson problem, Japan J. Indust. Appl. Math., 29 (2012), 23--35.
[Oh14] K. Ohtsuka and T. Takaishi, Finite element anaysis using mathematical programming language FreeFem++, Kyoritsu Shuppan, 2014. (in Japanese)
[Oh17] K. Ohtsuka, Shape optimization by GJ-integral: Localization method for composite material, Mathematical Analysis of Continuum Mechanics and Industrial Applications, Springer, 2017, 73--109.
[Oh18] K. Ohtsuka, Shape optimization by Generalized J-integral in Poisson's equation with a mixed boundary condition, Mathematical Analysis of Continuum Mechanics and Industrial Applications II, Springer, 2018, 73--83.
[Pr10] A.N. Pressley, Elementary differential geometry, Springer, 2010.
[Ri68] J.R. Rice, A path-independent integral and the approximate analysis of strain concentration by notches and cracks, J. Appl. Mech., 35(1968), 379--386.
[Ri68-2] J.R. Rice, Mathematical analysis in the mechanics of fracture, Fracture Volume II, Academic Press, 1968, 191--311.
[Pi84] O. Pironneau, Optimal shape design for elliptic systems, Springer-Verlag, 1984.
[Sa99] J.A. Samareh, A survey of shape parameterization techniques, NASA Report CP-1999-209136 (1999), 333--343.
[Sc91] B.-W. Schulze, Pseudo-differential operators on manifolds with singularities, North-Holland, 1991.
[Sok92] J. Sokolowski and J.-P. Zolesio, Introduction to shape optimization, Springer, 1992.
[St14] K. Sturm, On shape optimization with non-linear partial differential equations, Doctoral thesis, Technische Universiltät of Berlin, 2014. https://d-nb.info/106856959X/34
[Sumi] Y. Sumi, Mathematical and computational analyses of cracking formation, Springer, 2014.
[Zei/2B] E. Zeidler. Nonlinear functional analysis and its applications II/B, Springer, 1990.
[Z-S73] O.C. Zienkiewicz and J.S. Campbell, Shape optimization and sequential linear programming, Optimum Structural Design, Wiley, 1973, 109--126.