Vous êtes sur la page 1sur 3

4/14/2013

OVERVIEW
Numerical analysis is the study • Babylonian tablet YBC 7289
of algorithms for the problems of
continuous mathematics.
Numerical analysis continues
long tradition of practical
mathematical calculations.
Numerical analysis naturally
finds applications in all fields of
engineering and the physical
sciences, but in the 21st century,
the life sciences and even the arts
have adopted elements of
scientific computations.

Engr. Kristine P. Ortega

OVERVIEW GENERAL INTRODUCTION


SOME APPLICATIONS: Numerical Analysis overall goal is the design
•ordinary differential equations appear in the and analysis of techniques to give approximate
movement of heavenly bodies solutions to hard problems.
•optimization occurs in portfolio management CONSIDER SOME PROBLEMS AND METHODS:
•numerical linear algebra is essential to •If a company wants to put a toothpaste commercial on
quantitative psychology television, it might produce five commercials and then
•stochastic differential equations and Markov choose the best one by testing each one on a focus
chains are essential in simulating living cells for group. This would be an example of a Monte Carlo
medicine and biology optimization method.

Engr. Kristine P. Ortega Engr. Kristine P. Ortega

GENERAL INTRODUCTION GENERAL INTRODUCTION


•To send a rocket to the moon, rocket scientists will •Airlines use sophisticated optimization
need a rocket simulator. This simulator will essentially
be an integrator for an ordinary differential equation.
algorithms to decide ticket prices, airplane and
•Car companies can improve the crash safety of their crew assignments and fuel needs. This field is
vehicles by using computer simulations of car crashes. also called operations research.
These simulations are essentially solving partial •Insurance companies use numerical programs
differential equations numerically.
for actuarial analysis
•Hedge funds (private investment funds) use tools from
all fields of numerical analysis to calculate the value of
stocks and derivatives more precisely than other
market participants.
Engr. Kristine P. Ortega Engr. Kristine P. Ortega

1
4/14/2013

HISTORY DIRECT AND ITERATIVE METHODS


•The field of numerical analysis predates the Direct methods compute the solution to a
invention of modern computers by many problem in a finite number of steps.
centuries.
Iterative methods form successive
•To facilitate computations by hand, large books approximations that converge to the exact
were produced with formulas and tables of data
such as interpolation points and function solution only in the limit.
coefficients.
•The mechanical calculator was also developed
as a tool for hand computation.

Engr. Kristine P. Ortega Engr. Kristine P. Ortega

THE GENERATION AND PROPAGATION OF


DISCRETIZATION
ERRORS
Continuous problems must sometimes be The study of errors forms an important part of numerical
analysis. There are several ways in which error can be introduced
replaced by a discrete problem whose solution is in the solution of the problem.
known to approximate that of the continuous •Round-off
problem; this process is called discretization. Round-off errors arise because it is impossible to
represent all real numbers exactly on a finite-state
machine.
•Truncation and discretization error
Truncation errors are committed when an iterative
method is terminated and the approximate solution
differs from the exact solution.

Engr. Kristine P. Ortega Engr. Kristine P. Ortega

THE GENERATION AND PROPAGATION OF


ERRORS
AREAS OF STUDY
Discretization induces a discretization error Seven core areas:
because the solution of the discrete problem does
not coincide with the solution of the continuous • Nonlinear equations
problem. – Finding the zeros of a given function. The problem
•Numerical stability and well posedness translates into finding the roots of the equation,
or the points along the x-axis where the function
An algorithm is numerically stable if an error,
crosses. The roots may exist as real numbers. In
once it is generated, does not grow too much
during the calculation. cases where the real roots do not exist, there are
corresponding imaginary roots.
An algorithm that solves a well-conditioned
problem may or may not be numerically stable.
Engr. Kristine P. Ortega Engr. Kristine P. Ortega

2
4/14/2013

AREAS OF STUDY AREAS OF STUDY


• System of linear equations • Interpolation and approximation
– Often encountered in various science and – curve fitting problems that contribute in things
engineering problems. The problem arises like designing the surface of an aircraft. The
frequently as the original problem reduces to techniques are also applied in other problems,
linear equations at some stage in the solution. For such as forecasting, pattern matching, and
example, mesh modeling using the finite-element routing.
method in a fluid dynamics results in a system of
hundreds of linear equations. As the size of the
matrix in this problem is large, the solution needs
to be tackled using a fast numerical method.
Engr. Kristine P. Ortega Engr. Kristine P. Ortega

AREAS OF STUDY AREAS OF STUDY


• Differentiation and integration • Eigenvalues and eigenvectors
– fundamental topics that arise in many problems. • Ordinary differential equations
Good approximations are needed to these two – Arise in modeling and simulation. Exact solutions
topics as their exact values may not be easy to are difficult to obtain as the models are subject to
obtain. variation because of the presence of constraints
and nonlinear factors. Therefore, numerical
methods are needed in their successful
implementation.

Engr. Kristine P. Ortega Engr. Kristine P. Ortega

AREAS OF STUDY
• Partial differential equations
– Modeling and simulation involving partial
differential equations commonly use numerical
techniques as their fundamental elements.
Numerical methods contribute to provide the
desired solutions in most cases as the exact
solutions are not practical for implementation on
the computer.

Engr. Kristine P. Ortega

Vous aimerez peut-être aussi