Computational science 

(also scientific computing or scientific computation (SC)) is a

rapidly growing multidisciplinary field that uses advanced computing capabilities to understand
and solve complex problems. Computational science fuses three distinct elements:[1]

 Algorithms (numerical and non-numerical) and modeling and simulation software developed

to solve science (e.g., biological, physical, and social), engineering,
and humanities problems
 Computer and information science that develops and optimizes the advanced
system hardware, software, networking, and data management components needed to solve
computationally demanding problems
 The computing infrastructure that supports both the science and engineering problem solving
and the developmental computer and information science
In practical use, it is typically the application of computer simulation and other forms
of computation from numerical analysis and theoretical computer science to solve problems in
various scientific disciplines.
The field is different from theory and laboratory experiment which are the traditional forms of
science and engineering. The scientific computing approach is to gain understanding, mainly
through the analysis of mathematical models implemented on computers.
Scientists and engineers develop computer programs, application software, that model systems
being studied and run these programs with various sets of input parameters. In some cases,
these models require massive amounts of calculations (usually floating-point) and are often
executed on supercomputers or distributed computing platforms.
Numerical analysis is an important underpinning for techniques used in computational science.

Contents
  [hide] 

 1Applications of computational science

o 1.1Numerical simulations
o 1.2Model fitting and data analysis
o 1.3Computational optimization
 2Methods and algorithms
 3Reproducibility and open research computing
 4Journals
 5Education
 6Related fields
 7See also
 8References
 9Additional sources
 10External links

Applications of computational science[edit]

Problem domains for computational science/scientific computing include:

Numerical simulations[edit]
Numerical simulations have different objectives depending on the nature of the task being
simulated:

 Reconstruct and understand known events (e.g., earthquake, tsunamis and other natural
disasters).
 Predict future or unobserved situations (e.g., weather, sub-atomic particle behaviour, and
primordial explosions).
Model fitting and data analysis[edit]
 Appropriately tune models or solve equations to reflect observations, subject to model
constraints (e.g. oil exploration geophysics, computational linguistics).
 Use graph theory to model networks, such as those connecting individuals, organizations,
websites, and biological systems.
Computational optimization[edit]
Main article: Mathematical optimization

 Optimize known scenarios (e.g., technical and manufacturing processes, front-end

engineering).
 Machine learning

Methods and algorithms[edit]

Algorithms and mathematical methods used in computational science are varied. Commonly
applied methods include:

 Numerical analysis
 Application of Taylor series as convergent and asymptotic series
 Computing derivatives by Automatic differentiation (AD)
 Computing derivatives by finite differences
 Finite element method
 Graph theoretic suites
 High order difference approximations via Taylor series and Richardson extrapolation
 Methods of integration on a uniform mesh: rectangle rule (also called midpoint
rule), trapezoid rule, Simpson's rule
 Runge Kutta method for solving ordinary differential equations
 Monte Carlo methods
 Molecular dynamics
 Linear programming
 Branch and cut
 Branch and Bound
 Numerical linear algebra
 Computing the LU factors by Gaussian elimination
 Cholesky factorizations
 Discrete Fourier transform and applications.
 Newton's method
 Space mapping
 Time stepping methods for dynamical systems
Both historically and today, Fortran remains popular for most applications of scientific computing.
[2][3]
 Other programming languages and computer algebra systems commonly used for the more
mathematical aspects of scientific computing applications include GNU Octave, Haskell,[2] Julia,
[2]
 Maple,[3] Mathematica,[4] MATLAB, Python (with third-party SciPy library), Perl (with third-
party PDL library),[citation needed] R, SciLab, and TK Solver. The more computationally intensive
aspects of scientific computing will often use some variation of C or Fortran and optimized
algebra libraries such as BLAS or LAPACK.
Computational science application programs often model real-world changing conditions, such as
weather, air flow around a plane, automobile body distortions in a crash, the motion of stars in a
galaxy, an explosive device, etc. Such programs might create a 'logical mesh' in computer
memory where each item corresponds to an area in space and contains information about that
space relevant to the model. For example, in weather models, each item might be a square
kilometer; with land elevation, current wind direction, humidity, temperature, pressure, etc. The
program would calculate the likely next state based on the current state, in simulated time steps,
solving equations that describe how the system operates; and then repeat the process to
calculate the next state.
The term computational scientist is used to describe someone skilled in scientific computing. This
person is usually a scientist, an engineer or an applied mathematician who applies high-
performance computing in different ways to advance the state-of-the-art in their respective
applied disciplines in physics, chemistry or engineering. Scientific computing has increasingly
also impacted on other areas including economics, biology and medicine.
Computational science is now commonly considered a third mode of science, complementing
and adding to experimentation/observation and theory.[5] The essence of computational science
is numerical algorithm[6]and/or computational mathematics. In fact, substantial effort in
computational sciences has been devoted to the development of algorithms, the efficient
implementation in programming languages, and validation of computational results. A collection
of problems and solutions in computational science can be found in Steeb, Hardy, Hardy and
Stoop, 2004.[7]

Reproducibility and open research computing[edit]

The complexity of computational methods is a threat to the reproducibility of research.[8][9] Jon
Claerbout has become prominent for pointing out that reproducible research requires archiving
and documenting all raw data and all code used to obtain a result.[10][11][12] Nick Barnes, in
the Science Code Manifesto, proposed five principles that should be followed when software is
used in open science publication.[13] Tomi Kauppinen et al. established and defined Linked Open
Science, an approach to interconnect scientific assets to enable transparent, reproducible and
transdisciplinary research.[14]

Journals[edit]
Most scientific journals do not accept software papers because a description of a reasonably
mature software usually does not meet the criterion of novelty.[citation needed] Outside computer science
itself, there are only few journals dedicated to scientific software. Established journals
like Elsevier's Computer Physics Communications publish papers that are not open-access
(though the described software usually is). To fill this gap, a new journal entitled Open research
computation was announced in 2010;[15] it closed in 2012 without having published a single paper,
for a lack of submissions probably due to excessive quality requirements.[16] A new initiative was
launched in 2012, the Journal of Open Research Software. [17] In 2015, a new journal [18] dedicated
to the replication of computational results has been started on GitHub.

Education[edit]
Scientific computation is most often studied through an applied mathematics or computer
science program, or within a standard mathematics, sciences, or engineering program. At some
institutions a specialization in scientific computation can be earned as a "minor" within another
program (which may be at varying levels). However, there are increasingly many bachelor's and
master's programs in computational science. Some schools also offer the Ph.D. in computational
science, computational engineering, computational science and engineering, scientific
computation, or all three degrees bachelor's, master's and PhD in scientific computing.[19]
There are also programs in areas such as computational physics, computational chemistry, etc.