Scribd
Upload
25 views

Brief Explanation of Integration Schemes

This document provides an overview of explicit and implicit integration schemes for numerically solving differential equations. Explicit integration (like Euler's method) allows calculating future states directly from present states, but time steps must be small for stability. Implicit integration (like Backward Euler) requires solving an equation involving future and present states, making each step more computationally expensive, but is unconditionally stable for any time step size. The document uses the test equation dx/dt = λx to demonstrate: for explicit methods, the time step must satisfy Δt ≤ 2/|λ| for stability, while implicit methods are unconditionally stable for any Δt as long as <(λ) ≤ 0.

Uploaded by

Dac Ho Van
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
25 views

Brief Explanation of Integration Schemes

This document provides an overview of explicit and implicit integration schemes for numerically solving differential equations. Explicit integration (like Euler's method) allows calculating future states directly from present states, but time steps must be small for stability. Implicit integration (like Backward Euler) requires solving an equation involving future and present states, making each step more computationally expensive, but is unconditionally stable for any time step size. The document uses the test equation dx/dt = λx to demonstrate: for explicit methods, the time step must satisfy Δt ≤ 2/|λ| for stability, while implicit methods are unconditionally stable for any Δt as long as <(λ) ≤ 0.

Uploaded by

Dac Ho Van
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Brief Explanation of Integration Schemes

John T. Foster

Explicit Integration
Lets use a simple example to frame our investigation of explicit integration. Take a look
at Equation 1
x = f (x, t)
(1)
If we discretize t as we would in any numerical scheme t = t0 , t0 + t, t0 + 2t, . . . , t0 + nt
we can transform Equation 1 into the discrete time equivalent shown in Equation 2.
x n = f (xn , tn )

(2)

Now lets approximate the derivative at a given step n as follows.


x n =

(xn+1 xn )
t

(3)

Now substituting Equation 3 into Equation 1 we have the following.


(xn+1 xn )
= f (tn , xn )
t
xn+1 = xn + tf (tn , xn )

(4)

Using this formulation we can explicitly write down the solution for each consecutive step
in time given initial conditions. (i.e. x0 = c)
x1 = tf (t0 , c) + c
x2 = tf (t1 , x1 ) + x1
..
.

The particular method is known as Eulers method which is the simplest explicit integration
scheme. It allows for the calculation of xn+1 in terms of xn in a straightforward manner
at each time step, but is sensitive to stability issues (like all explicit methods) which will
be discussed next.
1

Stability of Explicit Integration


Lets attempt to solve the equation below using Eulers method.
x = x(t)

(5)

If we assume x(0) = c we know Equation 5 has the exact solution as follows


x(t) = cet

(6)

We also know that Equation 5 is only bounded (stable) when <() 0. Lets keep this in
mind as we investigate the stability of the numerical method. Discretize Equation 5 and
plug into Eulers formula.
xn+1 = txn + xn
= (1 + t)xn
= (1 + t)2 xn1
..
.
= (1 + t)n+1 x0

(7)

Equation 7 shows us that as n increases the only way xn+1 will not grow indefinitely is for
the following to hold true.
|1 + t| 1
(8)
Solving Equation 8 shows us that we must choose at time step that satisfies Equation 9
for the solution to be stable.
2
t
(9)
||
Therefore, even though the calculation of xn+1 at each step is simple in an explicit scheme
the time step must be small enough to ensure stability.

Implicit Integration
We will now use a Backward Euler method to demonstrate an implicit integration scheme.
We will only make a slight change to how we approximate the derivative.
x n =

(xn xn1 )
t

Substitute Equation 10 into Equation 2


(xn xn1 )
= f (xn , tn )
t
xn = tf (xn , tn ) + xn1
2

(10)

If we reindex n, we have the equation.


xn+1 = tf (xn+1 , tn+1 ) + xn

(11)

Since xn+1 appears on both sides of the Equation 11 it is said to be implicit in xn+1 .
This sometimes requires unique solution techniques to solve for xn+1 at each time step. So
each time step is computationally more expensive than an explicit method, but implicit
methods have advantages in stability.

Stability of Implicit Integration


Discretizing Equation 5 and plugging into the Backward Euler equation we have the following.
xn+1 = txn+1 + xn
xn+1 txn+1 = xn
(1 t)xn+1 = xn
xn
1 t
xn1
=
(1 t)2

xn+1 =

x0
(1 t)n+1

(12)

Equation 12 is stable as long as the following inequality holds true.


|1 t| 1

(13)

Because of the stability criterion on that requires <() 0, Equation 13 is always


true. This is said to be unconditionally stable. Herein lies the advantages of implicit
schemes.

You might also like