Skip to main content

Ordinary Differential Equations

Section 1.9 Behavior of solutions without solving

While most differential equations can’t be solved by hand, that doesn’t mean we can’t describe the behavior of their solutions.

Subsection 1.9.1 Direction fields

Definition 1.9.1. Direction fields (also called slope fields).

A direction field for (1.1.1) is a collection of slopes at points \((t,y)\) on solution curves. Visually, direction fields represent the approximate "flow of solutions" among a family of solution curves.
We now consider the differential equation \(ty'=y^2+1\text{.}\) First isolate \(y'\) to obtain \(y'=\dfrac{y^2+1}{t}\text{.}\) We see that we have put this differential equation into the form (1.1.1) with \(f(t,y)=\dfrac{y^2+1}{t}\text{.}\) That function \(f\) gives us the slope at each point \((t,y)\) in the plane. It turns out we can write an explicit general solution as \(y(t)=\tan(c+\log(t))\text{.}\) We now use this to explore the initial value problem \(ty'=y^2+1, \quad y(2)=1\text{.}\)

Example 1.9.2.

Modify the above code to plot the slope field and solution of the IVP \(y'=y^2\sin(t), \quad y(1)=-1\) which has general solution \(y(t)=\dfrac{1}{c+\cos(t)}\)

Subsection 1.9.2 Euler’s method

From calculus, we know that tangent lines (or linear approximation) can be used to estimate a differentiable function at some value \(t=a\text{.}\) In a similar manner, we would like to approximate the solution of an IVP.

Example 1.9.4.

Use Euler’s method with \(h=0.5\) and \(h=0.2\) to approximate the solution of the IVP
\begin{equation*} y'=f(t,y):=y-t^2+1, \quad y(0)=0.5 \end{equation*}
for \(0 \le t \le 2\text{.}\) It turns out that the exact solution to this IVP is \(y(t)=(t+1)^2-0.5e^t\text{.}\) We will learn how to find such solutions later on.

Subsection 1.9.3 Autonomous ODEs

Recall the form of an autonomous first-order ODE (1.1.3). Since the independent variable \(t\) does not appear in the right-hand side explicitly, we can study the behavior of its solutions and sketch typical solution curves without finding an explicit or implicit solution. Recall from calculus that we can sketch a function by knowing where it is increasing and decreasing, which we find by locating the critical points, i.e. the values of \(y\) so that \(y'=f(y)=0\text{.}\) There are three types of critical points. A critical point \(y=c\) is...
  1. asymptotically stable if as \(t \rightarrow \infty\text{,}\) the solution curve tends toward the horizontal line \(y=c\text{,}\)
  2. unstable if as \(t \rightarrow \infty\text{,}\) the solution curve moves away from the horizontal line \(y=c\text{,}\) and
  3. semi-stable if the solution curve moves either towards or away from the line \(y=c\text{,}\) depending on on whether the solution it is above or below that line at time \(t\text{.}\)

Example 1.9.5.

Find the critical points of the differential equation \(y'=y^3-4y^2\text{.}\)

Example 1.9.6.

Consider the differential equation \(\dfrac{\mathrm{d}y}{\mathrm{d}t}=y^2(y^4-16)\text{.}\)
  1. Find the critical points. Classify each critical point as either asymptotically stable, unstable, or semi-stable. Sketch the appropriate phase portrait.
  2. Sketch the typical solution curves in regions in the plane determined by the graphs of the equilibrium solutions.

Example 1.9.7.

Suppose two chemicals \(A\) and \(B\) are combined to form the compound \(X\) as was studied in Section 1.5. Initially, there are \(30\) units of \(A\) and \(40\) units of \(B\text{.}\) For every \(2\) units of \(B\text{,}\) \(3\) units of \(A\) are used to form the compound \(X\text{.}\) It is observed that \(6\) units of \(X\) are formed after \(10\) minutes.
  1. Set up the boundary value problem.
  2. Find the order of the differential equation. Is it (non)linear, (non)automous, and (non)separable?
  3. What is the limiting amount of \(X\) after a long time? How much of \(A\) and \(B\) remain after a long time?