Mainpage.dox

/*!
    @mainpage LifeV - A Parallel Finite Element Library


<BR>
\section introduction Introduction
LifeV (to be read Life Five) is an open source library for the numerical solution of
partial differential equations with the finite element method.
The choice of the finite element method is motivated by its rigorous mathematical background,
with a sound control of the accuracy, coupled with excellent flexibility in managing
complex geometries.

LifeV is written in C++ and is entirely coded with an Object Oriented approach
and advanced programming features. The project started in 1999 from the collaboration
of the Department of Mathematics at EPFL, Lausanne, Switzerland, of the
Department of Mathematics at Politecnico di Milano, Italy and the INRIA Rocquencourt, France,
under the supervision of Alfio Quarteroni.
Nowadays, LifeV is concurrently  developed and maintained by CMCS@EPFL Lausanne
(group coordinator), MOX@Politecnico di Milano, REO and ESTIME@INRIA and
 the E(CM)2 group at the Department of Mathematics and Computer Science at Emory University,
 Atlanta (GA), USA.

As for now, the library includes solvers for incompressible fluid dynamics,
(linear) structural problems, transport in porous media, fluid-structure interaction,
electrical conduction in the heart. Even though this library is a research code oriented to
the development and test of new numerical methods and algorithms,
it is intended to be an effective tool for solving complex "real-life"
engineering problems. One of the main field of applications (although not the only one)
is cardiovascular mathematics.

The code is distributed under the LGPL license and is downloadable on <a href="http://cmcsforge.epfl.ch">CMCSFORGE</a>. <BR>
There are two distributed versions of LifeV, one is parallel, for complex three dimensional
engineering problems, the other is serial,
for one, two and three dimensional problems (prototyping, testing). These pages
review the parallel version.<BR>
For other details, see the <a href="http://www.lifev.org">LifeV web site</a>.



<BR>
\section overview Overview of LifeV
In the following we briefly describe the main features of LifeV.

\subsection boundaryConditions Boundary Conditions
LifeV provides classes and methods to prescribe the most common boundary conditions
(<i>Dirichlet, Neumann, Robin</i>).
Normal and tangential boundary conditions are also allowed for vectorial (3D) fields.
In addition, a few kinds of boundary conditions specific for flow problems are available.
Boundary conditions are prescribed by modifying the system matrix and the right-hand side after the assembling phase.
The parameters of the boundary condition can be defined as functions of the space and time, or as finite element fields.
The regions where to apply boundary conditions are determined by markers associated to the boundary faces of the mesh.
Boundary conditions settings are stored in the \c BCHandler class (\c BCHandler.hpp), whereas the file
\c BCManage.hpp contains the routines used to prescribe the conditions.
The \c BCInterface class (\c BCInterface.hpp) contains tools to set the boundary conditions directly from data files.




\subsection timeDiscretization Time discretization
LifeV provides the class \c TimeAdvance to integrate in time a generic non-linear PDE
with derivates in time of order <i>m = 1, 2</i>. <BR>
This class is a virtual class, that defines the main features of a generic time advancing scheme:
<ul>
    <li> it stores the unknown and its first, and second if <i>m=2</i>, derivatives;</li>
    <li> it provides methods for the extrapolation (in time) of the unknown, and of its first derivative
         if <i> m=2 </i>;</li>
    <li> it computes the right-hand side associated to the discretization of the first derivative,
         and of the second derivative if <i>m=2</i>;</li>
    <li> it provides methods to update the states of the stored unknown.</li>
    </ul>

The needed coefficients, variables and methods are specified in derived classes.<BR>
LifeV provides two different implementations, namely:
<ul>
  <li><i>Backward Differentiation Formulae (BDF)</i> schemes (\c TimeAdvanceBDF.hpp) of order
       <i> p </i> (with <i> p<6</i>)  for <i>m=1,2</i>;</li>
  <li> the family of methods obtained from the <i>Newmark</i> schemes for
       second order problems (<i> m=2</i>), and the <i> theta-methods </i> for the first order
       problems (<i>m=1</i>) (\c TimeAdvanceNewmark.hpp).</li>
</ul>


\subsection spaceDiscretization Space discretization
LifeV provides classes to perform easily and quickly the assembly of the most common problems
(e.g. advection-diffusion-reaction, Stokes and Navier-Stokes). These classes
are called Assemblers and can be considered as building blocks that can be combined at will.
The user has moreover access to lower level structures and instructions in
order to provide the maximal flexibility.


\subsection algebraicSolver Algebraic solvers and preconditioners
Linear systems are solved using the preconditioned GMRES method. LifeV provides the following preconditioners:
<ul>
    <li>Additive Schwarz;
    <li>Multilevel Schwarz;
    <li>Multigrid preconditioner;
</ul>
It is also possible to develop new preconditioners using the abstract class \c Preconditioner (\c Preconditioner.hpp).


\subsection physicalSolver Physical solvers
In the following we briefly mention all the available physical solvers.

<ul>
    <li> <b>Advection Diffusion Reaction solver:</b> this solver (\c ADRAssembler.hpp) deals with the simplest PDE usually studied, but
         also maybe the most widely used. Indeed, this is the PDE that governs phenomena where diffusion, transport
         and reaction are relevant. This solver is used to solve this kind of problems in a simple and
         efficient manner. It can also deal with time-dependent PDEs.


    <li> <b>Darcy solver:</b> phenomenological law which describes the flow of a fluid through a porous medium.
         The implementation of the solver (\c DarcySolver.hpp) uses the dual-mixed-hybrid formulation, entailing good approximation of the
         velocity field as well as of the pressure field. The global system in saddle point form is recast to an equivalent
         positive definite system, using the hybridization and static condensation procedure. The finite element
         spaces used are piecewise constant for the pressure and lowest order Raviart-Thomas for the velocity.


    <li> <b>1D Fluid-Structure Interaction solver:</b> a solver for the Euler equations for blood flow is available (\c OneDFSISolver.hpp).
         It is based on a second order Taylor-Galerkin scheme, featuring excellent control of the dispersion error.
         In its current implementation, it can be used to study the pressure pulse wave propagation in an artery,
         assuming that the blood behaves as a Newtonian fluid, while the arterial wall behaves as a linearly elastic
         or viscoelastic structure. The presence of tapering in the arterial geometry can also be taken into account.


        <li> <b>3D Fluid-Structure Interaction solver:</b> the FSI model consists of Navier-Stokes equations coupled with an elasticity equation by imposing the continuity of the velocity
        and of the forces at the interface.
        In LifeV both partitioned and monolithic FSI solver are implemented. For the formers the coupling obtained with separate fluid and structure
        solver iterations, while for the latters the fully coupled
        matrix is assembled and solved using a single "monolithic" solver (\c FSIMonolithic.hpp). Concerning the partitioned strategies the
        available choices are Dirichlet-Neumann and Robin-Neumann fixed--point algorithms, or the Newton algorithm applied to the
        "Dirichlet-Neumann" fixed--point formulation. Various monolithic solver variants are obtained depending on the way the
        various nonlinearities (due mainly to the convective term of the fluid equations and to the domain motion) are treated. The main difference is
        among the geometric explicit (GE), for which the geometry is extrapolated from the previous time steps, and the geometric implicit (GI).
        Many parallel preconditioners for the monolithic problem are available, featuring the splitting of the matrix into several factors.

        <li> <b>Heart electrophysiological solver:</b> both bidomain (\c HeartBidomainSolver.hpp) and monodomain (\c HeartMonodomainSolver.hpp) models are available as models for the electrophysiology behavior of cardiac tissue.
        These models consist on anisotropic reaction-diffusion equations governing the propagation of electrical potentials,
        coupled with a system of ODEs describing the physics of the cellular membrane and time evolution of ionic quantities.
        Several variants for membrane models are already present in the library: Luo-Rudy phase I, Rogers-McCulloch and
        Mitchell-Schaeffer models. The solver currently employs piecewise linear finite elements for all involved fields
        in the coupling.


        <li> <b>Venant-Kirchhoff solver:</b> the structural model available in LifeV
        is a linear elastic constitutive law. It is the linear approximation of the nonlinear St.
        Venant-Kirchhoff model.
        It is implemented in the \c VenantKirchhoffSolver class (\c VenantKirchhoffSolver.hpp). It can be used both for
        purely structural dynamics problems (traction of a cube) or for coupled or multiscale problems.
        In fact, it has been integrated in the FSI (both segregated and monolithic) and in the
        Multiscale frameworks. At the moment, in the LifeV developers community there is a big effort
        to implement a general interface (i.e. a general solver) for solving structural mechanics problems.
        In particular, the efforts are focused on the implementation of other material laws
        and to exploit the general interface in the FSI and Multiscale frameworks.
    </ul>

\subsection geometricalMultiscale Geometrical Multiscale Framework
The geometrical multiscale framework provides a general and abstract way to couple different dimensionally-heterogeneous physical solvers.
In the framework, the local specific mathematical equations (partial differential equations, differential algebraic equations, etc.) and
the numerical approximation (finite elements, finite differences, etc.) of the heterogeneous compartments are hidden behind generic operators.
Consequently, the resulting global interface problem is formulated and solved in a completely transparent way. The coupling between models of
different dimensional scale (three-dimensional, one-dimensional, etc.) and type (Navier--Stokes, fluid-structure interaction, etc.) is
addressed writing the interface equations in terms of scalar quantities, i.e., area, flow rate, and mean (total) normal stress. In the
resulting flexible framework the heterogeneous models are treated as black boxes, each one equipped with a specific number of compatible
interfaces such that (i) the arrangement of the compartments in the network can be easily manipulated, thus allowing a high level of
customization in the design and optimization of the global geometrical multiscale model, (ii) the parallelization of the solution of the
different compartments is straightforward, leading to the opportunity to make use of the latest high-performance computing facilities, and
(iii) new models can be easily added and connected to the existing ones.

At the present time the framework includes the following models
<ol>
    <li> Fluid3D
    <li> FSI3D
    <li> OneDimensional
    <li> Multiscale
    <li> Windkessel0D
</ol>

which can be coupled through the following interface equations (or just closed by boundary conditions)
<ol>
    <li> BoundaryCondition
    <li> MeanNormalStress
    <li> MeanNormalStressArea
    <li> MeanNormalStressValve
    <li> MeanTotalNormalStress
    <li> MeanTotalNormalStressArea
</ol>

while the interface problem can be solved through the following algorithms
<ol>
    <li> Aitken
    <li> Broyden
    <li> Explicit
    <li> Newton
</ol>


\subsection tools Tools
In the following we briefly describe some of the main tools inside the library.

<ul>
        <li> <b>Importers and Exporters:</b> LifeV can input mesh files stored in some
        common formats encountered in this field: GMSH, Inria, Netgen, mesh++.
        LifeV is designed with large parallel simulations in mind, which are usually performed on
        high performance computing hardware.
        In many such cases, the low amount of memory per compute node requires a change in the workflow.
        LifeV allows to perform the mesh partitioning phase offline, on a workstation where memory
        isn't an issue and use this data later, during online simulations on modern supercomputers.
        For postprocessing, it's possible to save data in the following formats, using the common interface provided in \c Exporter.hpp:
        <ul>
            <li>Ensight (\c ExporterEnsight.hpp);
            <li>HDF5 (\c ExporterHDF5.hpp);
            <li>VTK (\c ExporterVTK.hpp).
        </ul>
        All of them are supported by ParaView, on which we rely for creating high quality visualizations.

        <li> <b>Grammar parser for algebraic expression:</b> the class \c Parser (\c Parser.hpp)provides a general
        interface to parse any simple algebraic function given as a string.
        This tool is based on \c Spirit which is a set of C++ libraries for parsing and output
        generation implemented as Domain Specific Embedded Languages (DSEL)
        using Expression Templates and Template Meta-Programming.
</ul>






<BR>
\section license License
    LifeV is available for download under the terms of the <a href="http://www.gnu.org/copyleft/lesser.html">
    GNU Lesser General Public License</a>.



<BR>
\section downloads Downloads
    LifeV can be downloaded from the <a href="http://www.lifev.org/downloads">LifeV downloads page</a>.
    For additional documentation about the LifeV architecture please check the
    <a href="http://www.lifev.org/documentation">LifeV documentation page</a>.


<BR>
\section email Mailing lists
    We recommend that users and developers subscribe to the following mailing lists as appropriate.
    <ul>
        <li><a href="http://groups.google.com/group/lifev-user">LifeV-Users Mailing List</a> -
        Discussion forum for users of %LifeV. (Recommended)
        <li><a href="http://groups.google.com/group/lifev-dev">LifeV-Developers Mailing List</a> -
        Discussion forum for %LifeV developers. (Developers)
        <li><a href="http://groups.google.com/group/lifev-cvs">LifeV-Checkins Mailing List</a> -
        GIT Commit Messages (Active  Developers)
    </ul>


<BR>
\section contributors LifeV contributors
    LifeV is the joint collaboration between four institutions:
    <ul>
        <li> École Polytechnique Fédérale de Lausanne (CMCS), Switzerland;
        <li> Politecnico di Milano (MOX), Italy;
        <li> INRIA (REO, ESTIME), France;
        <li> Emory University (Math&CS), U.S.A.
    </ul>

    The following people are active developers:
    <ul>
        <li> Alessandro Melani, MOX - Politecnico di Milano
        <li> Alessandro Veneziani, Math&CS - Emory University
        <li> Alessio Fumagalli, MOX - Politecnico di Milano
        <li> Antonio Cervone, MOX - Politecnico di Milano
        <li> Carlo de Falco, MOX - Politecnico di Milano
        <li> Claudia Colciago, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Christian Vergara, MOX - Politecnico di Milano
        <li> Cristiano Malossi, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Gianmarco Mengaldo, MOX - Politecnico di Milano
        <li> Guido Iori, MOX - Politecnico di Milano
        <li> Gwenol Grandperrin, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Huanhuan Yang, Math&CS - Emory University
        <li> Jean Bonnemain, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Laura Cattaneo, MOX - Politecnico di Milano
        <li> Luca Bertagna, Math&CS - Emory University
        <li> Luca Formaggia, MOX - Politecnico di Milano
        <li> Lucia Mirabella, CFM Lab - Georgia Institute of Technology
        <li> Matteo Pozzoli, MOX - Politecnico di Milano
        <li> Mauro Perego, Sandia National Laboratories
        <li> Michel Kern, ESTIME - INRIA
        <li> Nur Fadel, MOX - Politecnico di Milano
        <li> Paolo Crosetto, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Radu Popescu, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Ricardo Ruiz Baier, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Samuel Quinodoz, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Simone Deparis, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Simone Pezzuto, MOX - Politecnico di Milano
        <li> Simone Rossi, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Tiziano Passerini, Math&CS - Emory University
        <li> Toni Lassila, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Tricerri Paolo, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Umberto Emanuele Villa, Lawrence Livermore National Laboratory
    </ul>

    The following people are former developers:
    <ul>
        <li> Alexandra Moura, formerly at MOX, Politecnico di Milano
        <li> Alexis Aposporidis, formerly at Math&CS - Emory University
        <li> Andrea Manzoni, formerly at CMCS - École Polytechnique Fédérale de Lausanne
        <li> Andreas Fischle
        <li> Christoph Winkelmann, formerly at CMCS, École Polytechnique Fédérale de Lausanne
        <li> Christophe Prud'homme, formerly at CMCS, École Polytechnique Fédérale de Lausanne
        <li> Daniele A. Di Pietro, formerly at CMCS, École Polytechnique Fédérale de Lausanne
        <li> Fabio Nobile, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Federico Bonelli
        <li> Gilles Fourestey, formerly at CMCS - École Polytechnique Fédérale de Lausanne
        <li> Jean-Frédéric Gerbeau - REO, INRIA
        <li> Laura Iapichino, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Lorenzo Tamellini, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Marco Discacciati, formerly at CMCS, École Polytechnique Fédérale de Lausanne
        <li> Mariarita de Luca, formerly at MOX, Politecnico di Milano
        <li> Marta D'Elia, formerly at Math&CS - Emory University
        <li> Martin Prosi, formerly at MOX, Politecnico di Milano
        <li> Matteo Astorino, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Matteo Lesinigo, CMCS - École Polytechnique Fédérale de Lausanne
        <li> Miguel Ángel Fernández - REO, INRIA
        <li> Nicola Parolini, MOX - Politecnico di Milano
        <li> Paolo Ferrandi, formerly at MOX - Politecnico di Milano
        <li> Sara Minisini, formerly at MOX - Politecnico di Milano
        <li> Simone Cassani
        <li> Susanna Carcano, MOX - Politecnico di Milano
        <li> Vincent Martin, formerly at MOX - Politecnico di Milano
        <li> Zhen Wang, Math&CS - Emory University
    </ul>



<BR>
\section questions For other questions and comments...
Please visit the <a href="http://www.lifev.org">LifeV web site</a>.
 */