Skip to main content
Logo image

Section Interactions Between Variables

In the last section, we saw uncoupled systems, where each equation evolved completely on its own. But many real-world systems aren’t so tidy. Often one quantity affects another β€” or they push on each other in both directions.
In this section, we’ll step up the complexity: first looking at systems where one variable drives another, and then meeting the case where both variables interact fully.

Subsection Partially Coupled Systems

Suppose we have two variables, \(x(t)\) and \(y(t)\text{,}\) but this time \(y\) depends on \(x\text{.}\) In other words, \(x\) evolves independently, while \(y\) β€œlistens” to \(x\text{.}\)
Here’s an example:
\begin{align*} \frac{dx}{dt} \amp = -x \qquad x(0) = 1\\ \frac{dy}{dt} \amp = -2y + x \qquad y(0) = 0 \end{align*}
Notice what’s happening: \(x\) follows a simple exponential decay. Meanwhile, \(y\) is influenced by two things β€” its own decay and whatever value \(x\) has at the time. This is called a partially coupled system: one-way interaction.

🌌 Example 283. Solving a Partially Coupled System.

Solve the system:
\begin{align*} \frac{dx}{dt} \amp = -x \qquad x(0) = 1\\ \frac{dy}{dt} \amp = -2y + x \qquad y(0) = 0 \end{align*}
Solution.
Step 1: Solve the independent equation for \(x\text{.}\) This is a familiar exponential decay:
\begin{equation*} \frac{dx}{dt} = -x \quad \Rightarrow \quad x(t) = e^{-t}. \end{equation*}
Step 2: Substitute \(x(t) = e^{-t}\) into the \(y\)-equation:
\begin{equation*} \frac{dy}{dt} = -2y + e^{-t}. \end{equation*}
This is a linear first-order equation. Use an integrating factor to solve:
\begin{equation*} y' + 2y = e^{-t}, \qquad \mu(t) = e^{2t}. \end{equation*}
Multiply by \(e^{2t}\) and rewrite the left side as a derivative:
\begin{equation*} \frac{d}{dt}\left(e^{2t} y \right) = e^{t}. \end{equation*}
Integrate:
\begin{equation*} e^{2t} y = \int e^{t} dt = e^{t} + C. \end{equation*}
Solve for \(y(t)\text{:}\)
\begin{equation*} y(t) = e^{-t} + C e^{-2t}. \end{equation*}
Use the initial condition \(y(0) = 0\text{:}\)
\begin{equation*} 0 = 1 + C \quad \Rightarrow \quad C = -1. \end{equation*}
So the solution is:
\begin{equation*} x(t) = e^{-t}, \qquad y(t) = e^{-t} - e^{-2t}. \end{equation*}
\(x(t)\) decays smoothly, while \(y(t)\) rises briefly (pulled by \(x\)) before eventually decaying as well. This β€œone-way influence” is the hallmark of a partially coupled system.

Checkpoint 284. πŸ“–β“ Partially Coupled Equations.

(a) πŸ“–β“ Partial Coupling is One-Way.
    In a partially coupled system, one variable affects the other, but not vice versa.
  • True.

  • Exactly. In these systems, one equation is independent and can be solved first. The second equation depends on the first.
  • False.

  • Exactly. In these systems, one equation is independent and can be solved first. The second equation depends on the first.
(b) πŸ“–β“ Which Should You Solve First?
You are given the system:
\begin{align*} \frac{dx}{dt} \amp = -x \\ \frac{dy}{dt} \amp = -2y + x \end{align*}
Which equation should you solve first?
  • \(dy/dt\text{,}\) because it depends on both variables.
  • Not quite. You need to know \(x(t)\) before you can solve for \(y(t)\text{.}\)
  • It doesn’t matter, either can be solved independently.
  • Actually, only \(x\) evolves independently. \(y\) depends on it.
  • \(dx/dt\text{,}\) because it is independent of \(y\text{.}\)
  • Correct. Solve the independent equation for \(x\) first, then use that in the equation for \(y\text{.}\)
  • Neither equation can be solved until both are known.
  • Actually, this is a one-way system, \(x\) drives \(y\text{,}\) not the other way around.

Subsection Fully Coupled Systems (Conceptual)

We’ve seen systems where nothing interacts (uncoupled) and where one quantity drives another (partially coupled). Now we reach the most intricate case: fully coupled systems.
In a fully coupled system, both variables affect each other. Neither equation stands alone β€” any change in \(x\) alters \(y\)’s future, and \(y\) feeds back into \(x\text{.}\)
Consider:
\begin{align*} \frac{dx}{dt} \amp = x + y \qquad x(0) = 2\\ \frac{dy}{dt} \amp = x - y \qquad y(0) = 0 \end{align*}
Here, \(x\) and \(y\) are fully entangled β€” you can’t solve one without the other. Instead of peeling off one equation at a time, we’ll eventually treat the system as a single object.

Checkpoint 285. πŸ“–β“ What Makes It β€œFully Coupled”?

What must be true for a system to be fully coupled?
  • Both equations involve both variables.
  • Each equation has only one variable.
  • That would actually make the system uncoupled.
  • One equation is independent, and the other depends on it.
  • That’s partially coupled, not fully coupled.

Subsection πŸ“€ Wrap-Up

πŸ—οΈ \(\textbf{Key Takeaways...}\)
  • In a partially coupled system, one equation is independent and the other depends on it.
  • You can solve these without any new ideas by solving independent equation first, then substitute its solution into the dependent equation.
  • These systems model one-way influences, like a light level driving plant growth.
  • In fully coupled systems, both variables influence each other. Neither equation can be solved in isolation. That is, the system must be treated as a whole.

Check Your Understanding.

Checkpoint 286. πŸ€”πŸ’­ Understanding Partially Coupled Systems.
(a) πŸ€”πŸ’­ Identify the Coupling.
Consider the system:
\begin{align*} \frac{dx}{dt} \amp = -x \\ \frac{dy}{dt} \amp = -3y + x \end{align*}
Which of the following statements are true?
  • \(x\) evolves independently of \(y\text{.}\)
  • \(y\) is influenced by \(x\text{.}\)
  • Both variables depend on each other equally.
  • This is a partially coupled system.
  • You must solve the \(dy/dt\) equation first.
(b) πŸ€”πŸ’­ Solving Strategy.
What is the best way to solve a partially coupled system?
  • Solve the independent equation first, then substitute into the other equation.
  • Solve both equations at the same time using elimination.
  • You could do that, but it’s much simpler to solve the independent one first.
  • Write the system as one second-order DE before solving.
  • That’s unnecessary for partially coupled systems β€” keep it simple.
(c) πŸ€”πŸ’­ Spot the Coupling.
Consider the system:
\begin{align*} \frac{dx}{dt} \amp = 3x + y\\ \frac{dy}{dt} \amp = 4x - 2y \end{align*}
Which of these statements are true?
  • \(x\) depends on \(y\text{.}\)
  • \(y\) depends on \(x\text{.}\)
  • This is a partially coupled system.
  • Not quite β€” both equations are entangled, which makes it fully coupled.
  • You can solve for \(x\) first, then plug into \(y\text{.}\)
  • Because \(x\) depends on \(y\) as well, you can’t isolate them that easily.
You have attempted of activities on this page.