In applied mathematics,discontinuous Galerkin methods (DG methods) form a class ofnumerical methods for solvingdifferential equations. They combine features of thefinite element and thefinite volume framework and have been successfully applied tohyperbolic,elliptic,parabolic and mixed form problems arising from a wide range of applications. DG methods have in particular received considerable interest for problems with a dominant first-order part, e.g. inelectrodynamics,fluid mechanics andplasma physics. Indeed, the solutions of such problems may involve strong gradients (and even discontinuities) so that classical finite element methods fail, while finite volume methods are restricted to low order approximations.

Discontinuous Galerkin methods were first proposed and analyzed in the early 1970s as a technique to numerically solve partial differential equations. In 1973 Reed and Hill introduced a DG method to solve the hyperbolic neutron transport equation.

The origin of the DG method for elliptic problems cannot be traced back to a single publication as features such as jump penalization in the modern sense were developed gradually. However, among the early influential contributors wereBabuška,J.-L. Lions, Joachim Nitsche and Miloš Zlámal. DG methods for elliptic problems were already developed in a paper by Garth Baker in the setting of 4th order equations in 1977. A more complete account of the historical development and an introduction to DG methods for elliptic problems is given in a publication by Arnold, Brezzi, Cockburn and Marini. A number of research directions and challenges on DG methods are collected in the proceedings volume edited by Cockburn, Karniadakis and Shu.

Much like thecontinuous Galerkin (CG) method, the discontinuous Galerkin (DG) method is afinite element method formulated relative to aweak formulation of a particular model system. Unlike traditional CG methods that areconforming, the DG method works over a trial space of functions that are onlypiecewise continuous, and thus often comprise more inclusivefunction spaces than the finite-dimensional inner product subspaces utilized in conforming methods.

As an example, consider thecontinuity equation for a scalar unknown${\displaystyle \rho }$ in a spatial domain${\displaystyle \mathrm{\Omega }}$ without "sources" or "sinks" :

${\displaystyle \frac{\mathrm{\partial }\rho }{\mathrm{\partial }t}+\mathrm{\nabla }\cdot \mathbf{J}=0,}$

where${\displaystyle \mathbf{J}}$ is the flux of${\displaystyle \rho }$.

Now consider the finite-dimensional space of discontinuous piecewise polynomial functions over the spatial domain${\displaystyle \mathrm{\Omega }}$ restricted to a discretetriangulation${\displaystyle {\mathrm{\Omega }}_h}$, written as

${\displaystyle S_h^p({\mathrm{\Omega }}_h)=\{v_{|{\mathrm{\Omega }}_{e_i}}\in P^p({\mathrm{\Omega }}_{e_i}),\mathrm{\forall }{\mathrm{\Omega }}_{e_i}\in {\mathrm{\Omega }}_h\}}$

for${\displaystyle P^p({\mathrm{\Omega }}_{e_i})}$ the space of polynomials with degrees less than or equal to${\displaystyle p}$ over element${\displaystyle {\mathrm{\Omega }}_{e_i}}$ indexed by${\displaystyle i}$. Then for finite element shape functions${\displaystyle N_j\in P^p}$ the solution is represented by

${\displaystyle {\rho }_h^i=\sum _{j=1}^{\text{dofs}}{\rho }_j^i(t)N_j^i(\mathit{x}),\mathrm{\forall }\mathit{x}\in {\mathrm{\Omega }}_{e_i}.}$

Then similarly choosing a test function

${\displaystyle {\phi }_h^i(\mathit{x})=\sum _{j=1}^{\text{dofs}}{\phi }_j^i{N}_j^i(\mathit{x}),\mathrm{\forall }\mathit{x}\in {\mathrm{\Omega }}_{e_i},}$

multiplying the continuity equation by${\displaystyle {\phi }_h^i}$ andintegrating by parts in space, the semidiscrete DG formulation becomes:

${\displaystyle \frac{d}{dt}{\int }_{{\mathrm{\Omega }}_{e_i}}{\rho }_h^i{\phi }_h^{i}d\mathit{x}+{\int }_{\mathrm{\partial }{\mathrm{\Omega }}_{e_i}}{\phi }_h^i{\mathbf{J}}_h\cdot \mathit{n}d\mathit{x}={\int }_{{\mathrm{\Omega }}_{e_i}}{\mathbf{J}}_h\cdot \mathrm{\nabla }{\phi }_h^{i}d\mathit{x}.}$

A scalarhyperbolic conservation law is of the form

where one tries to solve for the unknown scalar function${\displaystyle u\equiv u(t,x)}$, and the functions${\displaystyle f,u_0}$ are typically given.

The${\displaystyle x}$-space will be discretized as

Furthermore, we need the following definitions

${\displaystyle h_k:=|I_k|,h:=\underset{k}{sup}{h}_k,{\hat{x}}_k:=x_k+\frac{h_k}{2}.}$

We derive the basis representation for the function space of our solution${\displaystyle u}$.The function space is defined as

${\displaystyle S_h^p:=\left\{v\in L^2(\mathbb{R}):v{|}_{I_k}\in {\mathrm{\Pi }}_p\right\}\text{for}p\in {\mathbb{N}}_0,}$

where${\displaystyle {v|}_{I_k}}$ denotes therestriction of${\displaystyle v}$ onto the interval${\displaystyle I_k}$, and${\displaystyle {\mathrm{\Pi }}_p}$ denotes the space of polynomials of maximaldegree${\displaystyle p}$.The index${\displaystyle h}$ should show the relation to an underlying discretization given by${\displaystyle {\left(x_k\right)}_k}$.Note here that${\displaystyle v}$ is not uniquely defined at the intersection points${\displaystyle (x_k{)}_k}$.

At first we make use of a specific polynomial basis on the interval${\displaystyle [-1,1]}$, theLegendre polynomials${\displaystyle (P_n{)}_{n\in {\mathbb{N}}_0}}$, i.e.,

${\displaystyle P_0(x)=1,P_1(x)=x,P_2(x)=\frac{1}{2}(3x^2-1),\dots }$

Note especially the orthogonality relations

${\displaystyle {⟨P_i,P_j⟩}_{L^2([-1,1])}=\frac{2}{2i+1}{\delta }_{ij}\mathrm{\forall }i,j\in {\mathbb{N}}_0.}$

Transformation onto the interval${\displaystyle [0,1]}$, and normalization is achieved by functions${\displaystyle ({\phi }_i{)}_i}$

${\displaystyle {\phi }_i(x):=\sqrt{2i+1}{P}_i(2x-1)\text{for}x\in [0,1],}$

which fulfill the orthonormality relation

${\displaystyle {⟨{\phi }_i,{\phi }_j⟩}_{L^2([0,1])}={\delta }_{ij}\mathrm{\forall }i,j\in {\mathbb{N}}_0.}$

Transformation onto an interval${\displaystyle I_k}$ is given by${\displaystyle {\left({\overline{\phi }}_{ki}\right)}_i}$

${\displaystyle {\overline{\phi }}_{ki}:=\frac{1}{\sqrt{h_k}}{\phi }_i\left(\frac{x-x_k}{h_k}\right)\text{for}x\in I_k,}$

which fulfill

${\displaystyle {⟨{\overline{\phi }}_{ki},{\overline{\phi }}_{kj}⟩}_{L^2(I_k)}={\delta }_{ij}\mathrm{\forall }i,j\in {\mathbb{N}}_0\mathrm{\forall }k.}$

For${\displaystyle L^{\mathrm{\infty }}}$-normalization we define${\displaystyle {\phi }_{ki}:=\sqrt{h_k}{\overline{\phi }}_{ki}}$, and for${\displaystyle L^1}$-normalization we define${\displaystyle {\tilde{\phi }}_{ki}:=\frac{1}{\sqrt{h_k}}{\overline{\phi }}_{ki}}$, s.t.

${\displaystyle \Vert {\phi }_{ki}{\Vert }_{L^{\mathrm{\infty }}(I_k)}=\Vert {\phi }_i{\Vert }_{L^{\mathrm{\infty }}([0,1])}=:c_{i,\mathrm{\infty }}\text{and}\Vert {\tilde{\phi }}_{ki}{\Vert }_{L^1(I_k)}=\Vert {\phi }_i{\Vert }_{L^1([0,1])}=:c_{i,1}.}$

Finally, we can define the basis representation of our solutions${\displaystyle u_h}$

Note here, that${\displaystyle u_h}$ is not defined at the interface positions.

Besides, prism bases are employed for planar-like structures, and are capable for 2-D/3-D hybridation.

The conservation law is transformed into its weak form by multiplying with test functions, and integration over test intervals

By using partial integration one is left with

${\displaystyle \begin{array}{r}\frac{\mathrm{d}}{\mathrm{d}t}{u}_{ki}(t)+f(u(t,x_{k+1})){\tilde{\phi }}_{ki}(x_{k+1})-f(u(t,x_k)){\tilde{\phi }}_{ki}(x_k)-{⟨f(u(t,\cdot )),{\tilde{\phi }}_{ki}^{\prime }⟩}_{L^2(I_k)}=0\text{for}\mathrm{\forall }k\mathrm{\forall }i\le p.\end{array}}$

The fluxes at the interfaces are approximated by numerical fluxes${\displaystyle g}$ with

${\displaystyle g_k:=g(u_k^{-},u_k^{+}),u_k^{\pm }:=u(t,x_k^{\pm }),}$

where${\displaystyle u_k^{\pm }}$ denotes the left- and right-hand sided limits.Finally, theDG-Scheme can be written as

${\displaystyle \begin{array}{r}\frac{\mathrm{d}}{\mathrm{d}t}{u}_{ki}(t)+g_{k+1}{\tilde{\phi }}_{ki}(x_{k+1})-g_k{\tilde{\phi }}_{ki}(x_k)-{⟨f(u(t,\cdot )),{\tilde{\phi }}_{ki}^{\prime }⟩}_{L^2(I_k)}=0\text{for}\mathrm{\forall }k\mathrm{\forall }i\le p.\end{array}}$

A scalar elliptic equation is of the form

This equation is the steady-state heat equation, where${\displaystyle u}$ is the temperature. Space discretization is the same as above. We recall that the interval${\displaystyle (a,b)}$ is partitioned into${\displaystyle N+1}$ intervals of length${\displaystyle h}$.

We introduce jump${\displaystyle [\cdot ]}$ and average${\displaystyle \{\cdot \}}$ of functions at the node${\displaystyle x_k}$:

${\displaystyle [v]{|}_{x_k}=v(x_k^{+})-v(x_k^{-}),\{v\}{|}_{x_k}=0.5(v(x_k^{+})+v(x_k^{-}))}$

The interior penalty discontinuous Galerkin (IPDG) method is: find${\displaystyle u_h}$ satisfying

${\displaystyle A(u_h,v_h)+A_{\mathrm{\partial }}(u_h,v_h)=\ell (v_h)+{\ell }_{\mathrm{\partial }}(v_h)}$

where the bilinear forms${\displaystyle A}$ and${\displaystyle A_{\mathrm{\partial }}}$ are

${\displaystyle A(u_h,v_h)=\sum _{k=1}^{N+1}{\int }_{x_{k-1}}^{x_k}{\mathrm{\partial }}_x{u}_h{\mathrm{\partial }}_x{v}_h-\sum _{k=1}^N\{{\mathrm{\partial }}_x{u}_h{\}}_{x_k}[v_h{]}_{x_k}+\epsilon \sum _{k=1}^N\{{\mathrm{\partial }}_x{v}_h{\}}_{x_k}[u_h{]}_{x_k}+\frac{\sigma }{h}\sum _{k=1}^N[u_h{]}_{x_k}[v_h{]}_{x_k}}$

and

${\displaystyle A_{\mathrm{\partial }}(u_h,v_h)={\mathrm{\partial }}_x{u}_h(a)v_h(a)-{\mathrm{\partial }}_x{u}_h(b)v_h(b)-\epsilon {\mathrm{\partial }}_x{v}_h(a)u_h(a)+\epsilon {\mathrm{\partial }}_x{v}_h(b)u_h(b)+\frac{\sigma }{h}(u_h(a)v_h(a)+u_h(b)v_h(b))}$

The linear forms${\displaystyle \ell }$ and${\displaystyle {\ell }_{\mathrm{\partial }}}$ are

${\displaystyle \ell (v_h)={\int }_a^{b}f{v}_h}$

and

${\displaystyle {\ell }_{\mathrm{\partial }}(v_h)=-\epsilon {\mathrm{\partial }}_x{v}_h(a)g(a)+\epsilon {\mathrm{\partial }}_x{v}_h(b)g(b)+\frac{\sigma }{h}(g(a)v_h(a)+g(b)v_h(b))}$

The penalty parameter${\displaystyle \sigma }$ is a positive constant. Increasing its value will reduce the jumps in the discontinuous solution. The term${\displaystyle \epsilon }$ is chosen to be equal to${\displaystyle -1}$ for the symmetric interior penalty Galerkin method; it is equal to${\displaystyle +1}$ for the non-symmetric interior penalty Galerkin method.

Thedirect discontinuous Galerkin (DDG) method is a new discontinuous Galerkin method for solving diffusion problems. In 2009, Liu and Yan first proposed the DDG method for solving diffusion equations.^[1][2] The advantage of this method compared with Discontinuous Galerkin method is that the direct discontinuous Galerkin method derives the numerical format by directly taking the numerical flux of the function and the first derivative term without introducing intermediate variables. We can still get reasonable numerical results by using this method, and due to the simpler derivation process, the amount of calculation is greatly reduced.

The direct discontinuous finite element method is a branch of the Discontinuous Galerkin methods. It mainly includes transforming the problem into variational form, regional unit splitting, constructing basis functions, forming and solving discontinuous finite element equations, and convergence and error analysis.

For example, consider a nonlinear diffusion equation, which is one-dimensional:

${\displaystyle U_t-{(a(U)\cdot U_x)}_x=0in(0,1)\times (0,T)}$, in which${\displaystyle U(x,0)=U_0(x)on(0,1)}$

Firstly, define${\displaystyle \left\{I_j=\left(x_{j-\frac{1}{2}},x_{j+\frac{1}{2}}\right),j=\mathrm{1...}N\right\}}$, and${\displaystyle \mathrm{\Delta }{x}_j=x_{j+\frac{1}{2}}-x_{j-\frac{1}{2}}}$. Therefore we have done the space discretization of${\displaystyle x}$. Also, define.

We want to find an approximation${\displaystyle u}$ to${\displaystyle U}$ such that${\displaystyle \mathrm{\forall }t\in \left[0,T\right]}$,${\displaystyle u\in {\mathbb{V}}_{\mathrm{\Delta }x}}$,

${\displaystyle {\mathbb{V}}_{\mathrm{\Delta }x}:=\left\{v\in L^2\left(0,1\right):{v|}_{I_j}\in P^k\left(I_j\right),j=1,...,N\right\}}$,${\displaystyle P^k\left(I_j\right)}$ is the polynomials space in${\displaystyle I_j}$ with degree at most${\displaystyle k}$.

Flux:${\displaystyle h:=h\left(U,U_x\right)=a\left(U\right)U_x}$.

${\displaystyle U}$: the exact solution of the equation.

Multiply the equation with a smooth function${\displaystyle v\in H^1\left(0,1\right)}$ so that we obtain the following equations:

${\displaystyle {\int }_{I_j}{U}_{t}vdx-h_{j+\frac{1}{2}}{v}_{j+\frac{1}{2}}+h_{j-\frac{1}{2}}{v}_{j-\frac{1}{2}}+\int a\left(U\right)U_x{v}_{x}dx=0}$,

${\displaystyle {\int }_{I_j}U\left(x,0\right)v\left(x\right)dx={\int }_{I_j}{U}_0\left(x\right)v\left(x\right)dx}$

Here${\displaystyle v}$ is arbitrary, the exact solution${\displaystyle U}$ of the equation is replaced by the approximate solution${\displaystyle u}$, that is to say, the numerical solution we need is obtained by solving the differential equations.

Choosing a proper numerical flux is critical for the accuracy of DDG method.

The numerical flux needs to satisfy the following conditions:

♦ It is consistent with${\displaystyle h={b\left(u\right)}_x=a\left(u\right)u_x}$

♦ The numerical flux is conservative in the single value on${\displaystyle x_{j+\frac{1}{2}}}$.

♦ It has the${\displaystyle L^2}$-stability;

♦ It can improve the accuracy of the method.

Thus, a general scheme for numerical flux is given:

${\displaystyle \hat{h}=D_{x}b(u)={\beta }_0\frac{\left[b\left(u\right)\right]}{\mathrm{\Delta }x}+\overline{{b\left(u\right)}_x}+\sum _{m=1}^{\frac{k}{2}}{\beta }_m{\left(\mathrm{\Delta }x\right)}^{2m-1}\left[{\mathrm{\partial }}_x^{2m}b\left(u\right)\right]}$

In this flux,${\displaystyle k}$ is the maximum order of polynomials in two neighboring computing units.${\displaystyle [\cdot ]}$ is the jump of a function. Note that in non-uniform grids,${\displaystyle \mathrm{\Delta }x}$ should be${\displaystyle \left(\frac{\mathrm{\Delta }{x}_j+\mathrm{\Delta }{x}_{j+1}}{2}\right)}$ and${\displaystyle \frac{1}{N}}$ in uniform grids.

Denote that the error between the exact solution${\displaystyle U}$ and the numerical solution${\displaystyle u}$ is${\displaystyle e=u-U}$ .

We measure the error with the following norm:

${\displaystyle \left|\left|\left|v(\cdot ,t)\right|\right|\right|={\left({\int }_0^1{v}^{2}dx+\left(1-\gamma \right){\int }_0^t\sum _{j=1}^N{\int }_{I_j}{v}_x^{2}dxd\tau +\alpha {\int }_0^t\sum _{j=1}^N{\left[v\right]}^2/\mathrm{\Delta }x\cdot d\tau \right)}^{0.5}}$

and we have${\displaystyle \left|\left|\left|U(\cdot ,T)\right|\right|\right|\le \left|\left|\left|U(\cdot ,0)\right|\right|\right|}$,${\displaystyle \left|\left|\left|u(\cdot ,T)\right|\right|\right|\le \left|\left|\left|U(\cdot ,0)\right|\right|\right|}$

• Galerkin method

• D.N. Arnold, F. Brezzi, B. Cockburn and L.D. Marini,Unified analysis of discontinuous Galerkin methods for elliptic problems, SIAM J. Numer. Anal. 39(5):1749–1779, 2002.
• G. Baker,Finite element methods for elliptic equations using nonconforming elements, Math. Comp. 31 (1977), no. 137, 45–59.
• A. Cangiani, Z. Dong, E.H. Georgoulis, and P. Houston,hp-Version Discontinuous Galerkin Methods on Polygonal and Polyhedral Meshes, SpringerBriefs in Mathematics, (December 2017).
• W. Mai, J. Hu, P. Li, and H. Zhao, “An efficient and stable 2-D/3-D hybrid discontinuous Galerkin time-domain analysis with adaptive criterion for arbitrarily shaped antipads in dispersive parallel-plate pair,”IEEE Trans. Microw. Theory Techn., vol. 65, no. 10, pp. 3671–3681, Oct. 2017.
• W. Maiet al., “A straightforward updating criterion for 2-D/3-D hybrid discontinuous Galerkin time-domain method controlling comparative error,”IEEE Trans. Microw. Theory Techn., vol. 66, no. 4, pp. 1713–1722, Apr. 2018.
• B. Cockburn, G. E. Karniadakis and C.-W. Shu (eds.),Discontinuous Galerkin methods. Theory, computation and applications, Lecture Notes in Computational Science and Engineering, 11. Springer-Verlag, Berlin, 2000.
• P. Lesaint, and P. A. Raviart. "On a finite element method for solving the neutron transport equation." Mathematical aspects of finite elements in partial differential equations 33 (1974): 89–123.
• D.A. Di Pietro and A. Ern,Mathematical Aspects of Discontinuous Galerkin Methods. Mathématiques et Applications, Vol. 69, Springer-Verlag, Berlin, 2011.
• J.S. Hesthaven and T. Warburton,Nodal Discontinuous Galerkin Methods: Algorithms, Analysis, and Applications. Springer Texts in Applied Mathematics 54. Springer Verlag, New York, 2008.
• B. Rivière,Discontinuous Galerkin Methods for Solving Elliptic and Parabolic Equations: Theory and Implementation. SIAM Frontiers in Applied Mathematics, 2008.
• CFD Wikihttp://www.cfd-online.com/Wiki/Discontinuous_Galerkin
• W.H. Reed and T.R. Hill,Triangular mesh methods for the neutron transport equation, Tech. Report LA-UR-73–479, Los Alamos Scientific Laboratory, 1973.

• Numerical differential equations
• Partial differential equations
• Finite element method

• Articles with short description
• Short description is different from Wikidata