Calc3.9

This is an old revision of this page, as edited by JacobB (talk | contribs) at 05:33, February 27, 2010. It may differ significantly from current revision.

Frequently we encounter equations in which we are solving for a function instead of a number. If such an equation involves an unknown function as well as its derivatives, it is called a differential equation. These equations have numerous applications in a variety of fields, including physics, engineering, chemistry, and biology. In this lecture, we will dipping our toes into the theory of ordinary differential equations, which means differential equations in which the unknown function(s) involves only one variable.

The Basics

We can immediately solve some of the most basic equations, such as  , where k is a constant, from our knowledge of the basic derivatives.

Integrable Equations

The easiest differential equations to solve, at least in theory, are those where we can manipulate the equation algebraically to get   on one side and some expression only involving x and constants on the other side. Then, we can just integrate both sides to get y. Integrable equations may not be presented in this format right away; often times, some manipulation is required to reach an integrable format. Take a look at  . We can manipulate this to get

 

and integrate to get

 

Let's check that this fits:

 

as expected.

Separable Equations

Most equations are not integrable. But sometimes, we're very lucky, and we find that we separate the dy and the dx from each other, and that we can arrange the equation such that only terms of y appear on one side, and only terms of x on the other. Let's do an example of this by solving  .

We can rewrite this

 

and separate it as

 

Now it's a simple matter of integrating!

 

 

which is an implicit equation with explicit solution

 

where we change the arbitrary constant c.

Not all separable equations are this obviously separable; sometimes much manipulation is needed to get it into a separable form.

First-Order Equations

Constant Co-efficients

Variable Co-efficients

Second-Order Equations

A second order equation is an equation which contains a   term, but no higher derivatives. We will consider only linear equations for now: this means that the coefficients of the derivatives of y do not contain terms in y.

Constant Co-efficients

When we solved  , we found that   was a solution. Note that   is a root of the equation  . Therefore, it makes sense to suppose that solutions of the equation   may be found with  , where   are roots of the equation  . Let r represent a specific root (either the first or second), and observe:

 

 

as desired. Hence we have found two possible solutions. An interesting thing happens if we consider linear combinations of these two solutions:

 

Taking all the relevant derivatives, we have

 

and examining the behavior of u in the LHS of the differential equation

 

 

reveals u is also a solution.

What about when the discriminant is 0, that is, when  ? In this case, lets drop the subscript and just call our root  .

Obviously,   will still be a solution, but is there a second solution, as there was before? The key insight needed to answer this question is that for a double root  , not only is the characteristic polynomial  , so is the derivative  . Therefore, it makes sense to consider the derivative of the solution   with respect to r as another possible solution. Let's check:

 

 .

As before, any linear combination   is also a solution.

Variable Co-efficients

Non-homogeneous Equations

Infinite Series

Keeping in mind how many functions can be represented with infinite sums created with Taylor series, we assume a function has an infinite series representation.

For example, let's solve   by assuming there exists   such that  . Then our differential equation becomes

 

Grouping like terms, we have

 

so that for  , the coefficient of   on the right hand side is  . So we need only set this equal to   to find an infinite series representation of our solution. If we're given the values of  , as we frequently are in applications (remember what we said about having two "degrees of freedom" in a DEQ involving second-order derivatives?) we can solve this completely. Let's say  . We have

     , since    , since   and so on.

This method may not seem particularly useful. After all, unless we can find a closed form solution for the  , we are going to have to keeping solving this linear equations in   forever to get the solution. However, it may happen, on occasion, that we can find such a closed form representation. Or, it may happen that we find no closed form, but recognize it as similar or equivalent to a function which has been studied for decades or centuries, so that all the properties we wish to know about are easily available to us. Or, we may not care about a closed form - near 0, those higher powers will become very small, and a low-order approximation becomes highly accurate.

Functional Transformations