Differential Equations: Concepts, Models, and Methods

Foundations: Functions, Derivatives, and Conditions

A differential equation relates an unknown function to one or more of its derivatives. For example, the equation

specifies the rate of change of an unknown function $y=f(x)$. A function is a solution when substitution of the function and its derivatives makes the equation true. Since

every function of the form $y=x^3+C$ is a solution. The constant $C$ produces a family of solutions rather than one unique function.

An Initial-value problem adds a condition such as $y(0)=2$. That condition identifies one particular member of the solution family.

Differential equations are especially useful for modeling quantities that change. If $P(t)$ denotes a population, temperature, amount of radioactive material, or concentration, then $dP/dt$ represents its instantaneous rate of change.

Takeaway: A differential equation describes how a quantity changes, while an initial condition helps determine which solution describes the particular situation.

Building Differential-Equation Models

The main modeling task is to translate a verbal statement about change into a derivative equation. A reliable sequence is:

1. Define each variable and include its units.

2. Translate the rate statement into a derivative.

3. Add an initial condition when one is given.

4. Solve the equation exactly or approximate its solution.

5. Interpret the result in context and check its units and behavior.

For example, if a population $P$ increases at a rate proportional to its current size, the statement becomes

Here, $k$ is a constant of proportionality. More generally, a model can be written as

meaning that the rate can depend on time, the current amount, or both.

A differential equation gives the slope or rate of change at each point; it does not directly give the value of the quantity. This distinction explains why graphical tools such as slope fields and numerical procedures such as Euler’s method are useful.

Takeaway: Modeling begins with a precise translation from words about rates into derivatives, followed by solving or approximating and then interpreting the result.

Solving Separable Equations

A Separable differential equation can be written in the form

When $g(y)$ is nonzero, rearrange the equation so that the variables are separated:

Then integrate both sides:

For an initial-value problem

use the following procedure:

1. Separate the variables.

2. Integrate both sides and include a constant of integration.

3. Apply the initial condition.

4. Solve explicitly when useful, or retain an implicit equation.

5. Differentiate and substitute back to check the result.

When dividing by an expression involving $y$, check whether that expression can equal zero. A constant solution may be lost during the division.

For example, solve

Separating and integrating gives

The solution can be written as $y=Ce^{x^2}$. Applying $y(0)=3$ gives $C=3$, so

A check confirms that

Takeaway: Separation converts one differential equation into two integrals, but possible equilibrium or constant solutions must be checked before accepting the final answer.

Visualizing Solution Behavior

A Slope field represents the equation

by placing a short line segment with slope $f(x,y)$ at each point $(x,y)$. A solution curve passing through a point must be tangent to the segment at that point.

To construct a field for

calculate the slope at selected grid points. At $(1,2)$, for instance,

Draw a short segment with slope $-1$ there, and repeat the process across the grid.

Use the field to identify behavior:

• Positive slopes indicate that solution curves rise as $x$ increases.

• Negative slopes indicate that solution curves fall as $x$ increases.

• Zero slopes indicate horizontal tangents.

• Repeated horizontal levels can reveal constant or equilibrium solutions.

• The arrangement of segments can show whether curves approach or move away from a level.

For an autonomous equation such as $dP/dt=f(P)$, the slope depends only on $P$, so the pattern repeats horizontally across the field. A solution curve is not one individual segment; it is a smooth path that follows the local directions throughout the plane.

Takeaway: A slope field gives a qualitative picture of solution behavior without requiring an exact formula.

Approximating Solutions Numerically

When an exact solution is inconvenient or unavailable, Euler’s method estimates the solution by taking successive tangent-line steps. For

and step size $h$, the update rules are

To approximate $y(0.3)$ for

with $h=0.1$, begin with $(x_0,y_0)=(0,1)$. The successive calculations are:

1. At $(0,1)$, the slope is $1$, so $y_1=1+0.1(1)=1.1$.

2. At $(0.1,1.1)$, the slope is $1.2$, so $y_2=1.1+0.1(1.2)=1.22$.

3. At $(0.2,1.22)$, the slope is $1.42$, so $y_3=1.22+0.1(1.42)=1.362$.

Therefore,

A smaller step size generally improves the approximation because each tangent-line step covers a shorter interval, although it requires more calculations. The method can overestimate or underestimate depending on the curvature of the actual solution. Compare the numerical path with the slope field when checking the result.

Takeaway: Euler’s method converts local slope information into a sequence of approximate values.

Exponential Growth and Decay

Exponential growth and decay assume that the rate of change is proportional to the current amount:

Separating and integrating produces

where $P_0=P(0)$. If $k>0$, the quantity grows; if $k<0$, it decays; and if $k=0$, it remains constant.

For a culture beginning with 500 bacteria and having proportionality constant $0.2$ per hour,

After 4 hours,

For growth, the doubling time is

For decay, the half-life is

which is positive because $k<0$.

The exponential model assumes unlimited resources and a constant proportional growth rate. It may describe a short interval well, but it predicts unbounded growth when $k>0$, which is often unrealistic over a long period.

Takeaway: Exponential models are appropriate when the current amount controls the rate and limiting resources are not yet important.

Logistic Growth and Limited Resources

A Logistic model modifies exponential growth by including a Carrying capacity, the sustainable level represented by $K$:

where $r>0$ is the intrinsic growth rate. The factor $1-P/K$ reduces growth as the population approaches $K$.

The sign of the derivative explains the behavior:

• If $0<P<K$, then $dP/dt>0$, so the population increases.

• If $P>K$, then $dP/dt<0$, so the population decreases.

• If $P=0$ or $P=K$, then the rate is zero, giving Equilibrium solution levels.

For a population starting between zero and $K$, the curve is typically sigmoidal. Growth begins slowly, becomes fastest near $P=K/2$, and then slows as the population approaches $K$. At the midpoint,

The general solution is

where the initial condition determines $A$. If $P(0)=P_0$, then

and therefore

For $K=1000$, $P_0=100$, and $r=0.4$ per year, $A=9$, so

As $t$ increases, $e^{-0.4t}$ approaches zero and $P(t)$ approaches 1000.

Takeaway: Logistic growth incorporates resource limits: the population grows rapidly near the middle of its range but approaches its carrying capacity over time.

Connecting Models and Representations

The main models differ in the assumptions they make about resources and long-term behavior.

• The exponential equation is $P'=kP$. It assumes unlimited resources and produces unbounded growth when $k>0$.

• The logistic equation is $P'=rP(1-P/K)$. It includes limited resources and approaches the carrying capacity $K$.

• Exponential growth has no finite maximum growth rate within the model.

• Logistic growth is fastest at $P=K/2$.

• The logistic model has equilibrium levels $P=0$ and $P=K$, while the exponential model has the constant solution $P=0$.

A strong solution process moves among several representations:

1. Verbal: describe how a quantity changes.

2. Symbolic: write a differential equation such as $dy/dx=x+y$.

3. Graphical: inspect a slope field.

4. Numerical: use Euler’s method to generate approximations.

5. Algebraic: find an exact solution when possible.

6. Contextual: interpret the result, including its units and limitations.

For example, the verbal statement that a population grows proportionally to its current size becomes $P'=kP$, which leads to an exponential model. If limited resources matter, the statement must be modified to produce a logistic model instead.

Final takeaway: Differential equations connect rates, formulas, graphs, numerical approximations, and real-world interpretations. The most appropriate model depends on what controls the rate of change and on whether the assumptions remain reasonable over the time interval being considered.