Proper Rational Functions: The Mathematical Language of Physical Systems

In the quest to understand and engineer the world around us, mathematics provides a powerful language. We model everything from electrical circuits to robotic arms using equations, but how do we ensure these abstract models are faithful to physical reality? The universe imposes fundamental constraints; for instance, systems possess inertia and cannot respond instantaneously. A crucial challenge in engineering and physics is embedding these physical laws into our mathematical descriptions. This article addresses this very problem by exploring the concept of a ​​proper rational function​​.

We will uncover how this simple algebraic property of a system's transfer function serves as a gatekeeper for physical realizability. The journey begins in the first chapter, ​​Principles and Mechanisms​​, where we will define proper, improper, and strictly proper functions and connect them to core physical limitations like finite gain and non-instantaneous response. We will see how polynomial degrees dictate a system's behavior at extreme frequencies and initial moments. Following this, the second chapter, ​​Applications and Interdisciplinary Connections​​, will demonstrate how these principles are not just theoretical curiosities but are the bedrock of modern control engineering, signal processing, and system analysis, enabling us to design stable, effective, and predictable systems.

Have you ever wondered why nothing in our world seems to happen instantaneously? When you flip a light switch, the filament in the bulb takes a moment to heat up and glow. When you press the accelerator in a car, it takes time to build up speed. This isn't just a quirk of engineering; it's a fundamental principle of the physical universe. Systems have inertia. They cannot change their state from one value to another in zero time. There's a sort of cosmic speed limit on how fast things can respond.

In the world of signals and systems, we have a wonderfully elegant way to describe how a system behaves: the ​​transfer function​​, often denoted as $H(s)$. Think of it as the system's mathematical identity card. It tells us how the system will react to any possible input, from a simple nudge to a complex vibration. If we want our mathematical models to reflect reality, they must respect this cosmic speed limit. The question is, how does this profound physical constraint—the impossibility of instantaneous reaction—manifest itself in the simple equation of a transfer function?

The answer, as we'll see, is surprisingly simple and beautiful, hidden in the basic algebra of polynomials.

For a vast number of physical systems, from electrical circuits to mechanical suspensions, the transfer function takes the form of a ​​rational function​​—that is, a fraction with a polynomial in the numerator, $N(s)$, and a polynomial in the denominator, $D(s)$:

$H(s) = \frac{N(s)}{D(s)}$

The secret to physical realizability lies in comparing the "power" or ​​degree​​ of these two polynomials. The degree is simply the highest exponent of the variable $s$ in the polynomial. Based on this comparison, we can sort all rational transfer functions into three fundamental categories:

• ​​Improper​​: A transfer function is improper if the degree of the numerator is strictly greater than the degree of the denominator ($\deg(N) > \deg(D)$). As we'll see, these are the mathematical outlaws, representing systems that violate our cosmic speed limit.

• ​​Proper​​: A transfer function is proper if the degree of the numerator is less than or equal to the degree of the denominator ($\deg(N) \le \deg(D)$). This is the family of "physically realizable" systems. They are well-behaved and respect the laws of physics.

• ​​Strictly Proper​​: A special, very common subclass of proper functions where the degree of the numerator is strictly less than the degree of the denominator ($\deg(N) \deg(D)$).

• ​​Biproper​​: This is the case where the degrees are exactly equal ($\deg(N) = \deg(D)$). It's a borderline case, proper but not strictly so.

This simple classification based on polynomial degrees is the first key to understanding the deep connection between abstract mathematics and concrete physical behavior.

So what's actually wrong with an improper system? Why do we label it "unphysical"? Let's play the role of an engineer and try to build one.

Consider a transfer function like $H(s) = \frac{s^2}{s+b}$. Notice that the degree of the numerator (2) is greater than the degree of the denominator (1), so this is an improper function. Using simple polynomial long division, we can rewrite it as:

$H(s) = s - b + \frac{b^2}{s+b}$

Now, let's translate this back into the time domain. The term $\frac{b^2}{s+b}$ is proper and corresponds to a well-behaved exponential decay, $b^2 \exp(-bt)$. But what about the other terms? In the language of the Laplace transform, multiplying a signal's transform by $s$ is equivalent to taking the derivative of the signal in the time domain. This means our "improper" system is trying to compute the derivative of its input signal!

This is the "crime" of improper systems. An ideal differentiator is a theoretical fiction. To see why, think about its frequency response. For a differentiator $H(s)=s$, the response to a sinusoidal input of frequency $\omega$ is $H(j\omega) = j\omega$. The magnitude of this response is $|H(j\omega)| = \omega$. This means as the input frequency gets higher and higher, the output gets larger and larger, without any bound. Any real-world signal has some high-frequency noise. An ideal differentiator would amplify this noise to infinite levels, completely swamping the actual signal. No physical device can supply infinite energy or have infinite gain.

Therefore, for a system to be considered physically realizable without resorting to these impossible ideal differentiators, its transfer function must be ​​proper​​. This elegant mathematical condition, $\deg(N) \le \deg(D)$, is the direct embodiment of our physical intuition that systems can't react infinitely fast or have infinite gain. It's a cornerstone of system theory, and it holds true for complex multi-input, multi-output (MIMO) systems as well. The fundamental theorem of realization states that a system can be described by the standard state-space equations if and only if its transfer function matrix is both rational and proper.

If an improper system's gain shoots off to infinity, what do proper systems do at very high frequencies? Let's find out by looking at the limit of $H(s)$ as $s$ becomes very large ($|s| \to \infty$). This limit is called the ​​high-frequency gain​​.

For a ​​strictly proper​​ function ($\deg(N) \deg(D)$), the denominator grows faster than the numerator, so the fraction must go to zero:

$\lim_{|s|\to\infty} H(s) = 0 \quad (\text{Strictly Proper})$

These systems act like low-pass filters; they effectively block signals that oscillate extremely rapidly.

For a ​​biproper​​ function ($\deg(N) = \deg(D)$), the highest powers of $s$ in the numerator and denominator balance each other out, and the limit is a finite, non-zero constant determined by the leading coefficients:

$\lim_{|s|\to\infty} H(s) = \frac{a_n}{b_n} = K \neq 0 \quad (\text{Biproper})$

These systems allow very high-frequency signals to pass through with a certain fixed gain.

This behavior has a beautiful interpretation when we look at the system's internal structure using the ​​state-space representation​​. In this view, a system is described by matrices $(A, B, C, D)$, and its transfer function is given by $H(s) = C(sI - A)^{-1}B + D$. The term $C(sI - A)^{-1}B$ is always strictly proper. Therefore, when we take the limit as $|s| \to \infty$, this term vanishes, and we are left with a stunningly simple result:

$\lim_{|s|\to\infty} H(s) = D$

The high-frequency gain is nothing more than the ​​direct feedthrough matrix​​, $D$! A strictly proper system ($D=0$) means the input signal must pass through the system's internal dynamics (represented by matrix $A$) before influencing the output. A biproper system ($D \neq 0$) has a direct, instantaneous path from input to output. This gives a physical meaning to our classification: strict properness implies a dynamic delay, while biproperness implies an instantaneous (but finite!) connection.

We can be even more precise about how a system's gain fades at high frequencies. The key is the ​​relative degree​​, $r$, defined as the difference between the denominator and numerator degrees: $r = \deg(D) - \deg(N)$. For any proper system, $r \ge 0$.

Let's take an example: $H(s) = \frac{s^2+3s+2}{s^3+4s^2+5s+2}$. Here, $\deg(D)=3$ and $\deg(N)=2$, so the relative degree is $r=1$. For very large $s$, this function behaves like $\frac{s^2}{s^3} = \frac{1}{s}$. Its magnitude dies off proportionally to $1/|s|$. If the relative degree were 2, it would behave like $1/s^2$ and die off much faster.

Engineers have a favorite way to visualize this: the ​​Bode magnitude plot​​, which graphs the gain in ​​decibels (dB)​​ against frequency on a logarithmic scale. A key feature of this plot is its slope at high frequencies. And here lies another moment of beautiful unity. The high-frequency asymptotic slope of the Bode plot is given by a simple, elegant formula:

$\text{Slope} = -20r \text{ dB per decade}$

A "decade" is a tenfold increase in frequency. So, for a system with relative degree $r=1$, the gain drops by 20 dB every time the frequency increases by a factor of 10. For $r=2$, it drops by 40 dB per decade. For $r=3$, it's -60 dB/decade, and so on. A simple integer, the relative degree, which you can find just by inspecting the transfer function, tells you the exact angle of a line on a graph that characterizes the system's physical response to high-frequency vibrations.

The relative degree doesn't just tell us about the far future (high frequencies); it also reveals secrets about the very first instant of time, $t=0^+$. The ​​Initial Value Theorem​​ of the Laplace transform connects the behavior of a function at $t=0^+$ to the behavior of its transform as $s \to \infty$.

Consider an impulse hitting a system. What is the output value, $y(t)$, at the very moment after the impulse, $t=0^+$? If the system is biproper ($r=0$), the output will jump to a finite value determined by the matrix $D$. If the system is strictly proper with $r=1$ (e.g., $H(s) = 1/(s+1)$), the output also jumps instantaneously to a non-zero value. The response is $y(t) = \exp(-t)$, and $y(0^+) = 1$.

But what if we are designing a mechanical system, and we cannot tolerate any instantaneous jumps in position or velocity? We need a response that starts at zero and then smoothly begins to change. To guarantee this, the system must have a relative degree of at least 2.

If $r \ge 2$, then the transform $H(s)$ dies off at least as fast as $1/s^2$ at high frequencies. The Initial Value Theorem then guarantees that the impulse response $y(0^+)$ will be zero. This subtle distinction within the "strictly proper" family is crucial in many engineering designs. A relative degree of 1 ensures physical realizability, but a relative degree of 2 ensures a "soft start." This same principle applies in discrete-time systems, where a signal's Z-transform is strictly proper if and only if the signal's initial value is zero.

We have seen that the degrees of the polynomials in $H(s)$ tell us about the system's behavior at the extremes of frequency and time. But what about the entire response? The answer lies not in the degrees, but in the roots of the denominator polynomial, $D(s)$. These roots are called the ​​poles​​ of the system.

Let's look at a stable, strictly proper system like $G(s) = \frac{s+4}{(s+1)(s+2)(s+3)}$. The poles are at $s=-1, s=-2,$ and $s=-3$. Using a technique called ​​partial fraction expansion​​, we can break this complicated function into a sum of simpler terms:

$G(s) = \frac{3/2}{s+1} - \frac{2}{s+2} + \frac{1/2}{s+3}$

Each of these simple terms has a well-known inverse Laplace transform: it's an exponential function. The term $\frac{1}{s+a}$ corresponds to $\exp(-at)$ in the time domain. Thus, the total impulse response of our system is simply a weighted sum of these fundamental exponential "modes":

$h(t) = \left(\frac{3}{2}\exp(-t) - 2\exp(-2t) + \frac{1}{2}\exp(-3t)\right) u(t)$

Here we see the full picture. The properness of a rational function is a simple algebraic check that ensures a system is physically plausible. Its relative degree quantifies its high-frequency behavior and its initial response. And finally, the roots of its denominator dictate the characteristic exponential building blocks of its behavior over all time. What begins as a simple question about physical limits unfolds into a rich and interconnected theory, where the elementary properties of polynomials provide a deep and powerful language for describing the real world.

Now that we have explored the principles and mechanisms of proper rational functions, you might be thinking, "This is elegant mathematics, but what is it for?" This is where the real fun begins. It turns out that this seemingly simple mathematical idea is not just a curiosity; it is the natural language used to describe a staggering variety of phenomena in the physical world. From the hum of your refrigerator to the stability of a soaring aircraft, the ghost of the proper rational function is there, quietly dictating the rules of the game. Let us embark on a journey to see where these ideas come alive, moving from the concrete world of engineering to the abstract realms of mathematics itself.

Imagine you are an engineer tasked with building a device. Before you even solder a single component, there's a fundamental question you must answer: is your design physically possible? Nature has certain non-negotiable laws, and one of them can be stated in the language of our new friend, the rational function.

Consider the task of building a perfect "differentiator," a device whose output is the rate of change of its input. In the language of Laplace transforms, this ideal device has a transfer function $G(s) = K s$. Notice something? The degree of the numerator (1) is greater than the degree of the denominator (0). This function is improper. What does nature say about this? If you analyze its frequency response, you find that its gain—its amplification factor—grows infinitely large as the frequency of the input signal increases. Any real-world signal is contaminated with at least a tiny amount of high-frequency noise. A device with this transfer function would act like a megaphone for this noise, amplifying it to an unmanageable, potentially infinite level. The device would be completely overwhelmed, its output saturated and meaningless. Nature, in its wisdom, forbids infinite energy, and thus forbids improper systems.

This leads to a profound conclusion: for a system to be physically realizable, its transfer function must be ​​proper​​. It must not amplify signals infinitely at high frequencies. This simple mathematical constraint acts as a fundamental gatekeeper, separating the blueprints of possible machines from the fantasies of impossible ones.

Now, let's flip the coin. What about an ideal "integrator," a system whose output is the accumulated sum of its input over time? Its transfer function is $H(s) = \frac{1}{s}$. Here, the degree of the numerator (0) is less than the degree of the denominator (1). This is a strictly proper rational function. It is causal, has memory (as it must, to remember the past input it's integrating), and, most importantly, it is physically realizable. Its gain decreases with frequency, meaning it naturally suppresses high-frequency noise. This is the kind of well-behaved system that nature permits. The property of properness is the mathematical signature of physical possibility.

So, a proper rational function is a valid blueprint for a physical system. But what does the blueprint tell us? It turns out that every detail of the function's structure corresponds to a specific characteristic of the system's behavior.

The most important features are the poles of the function—the roots of the denominator polynomial. These poles are like the system's genetic code. They determine the "natural modes" of the system's response when left to its own devices. When you analyze a linear time-invariant system, its output's Laplace transform is often a proper rational function. By breaking this function down using partial fraction expansion, we can see that the time-domain signal is a sum of simple terms, each corresponding to a pole. A real pole at $s = -a$ corresponds to an exponential decay $e^{-at}$. A pair of complex conjugate poles at $s = -a \pm jb$ corresponds to a damped oscillation, $e^{-at}\cos(bt)$ and $e^{-at}\sin(bt)$. By simply looking at the poles of the transfer function, we can immediately predict whether the system will oscillate, decay quickly, or decay slowly. The same logic applies beautifully to the digital world of signals and systems, where the poles of a Z-transform function tell a similar story about the behavior of a discrete-time sequence.

This predictive power extends to how a system responds to different frequencies. Suppose you're designing a low-pass filter to block out high-frequency noise from a sensor reading. Your design specification might demand that the filter's gain drops off very quickly for high frequencies, say at a rate of at least -100 decibels per decade on a Bode plot. How do you achieve this? The answer lies in the relative degree of the filter's transfer function, $H(s) = \frac{N(s)}{D(s)}$, which is the difference between the degree of the denominator and the numerator, $n-m$. For high frequencies, the gain of the filter rolls off at a rate of $-20 \times (n-m)$ dB/decade. To achieve a -100 dB/decade slope, you need a relative degree of at least 5. This gives engineers a direct, quantitative tool: to make a filter more aggressive, you simply need to make its transfer function "more proper" by increasing the relative degree.

The theory of proper rational functions is not just for analyzing systems; it is a powerful toolkit for synthesis—for building and controlling them.

Perhaps the crowning achievement in this arena is the theory of feedback control. Imagine trying to keep a rocket upright or maintain a constant temperature in a chemical reactor. These are inherently unstable or sluggish processes. The solution is feedback: measure the output, compare it to the desired value, and use the error to adjust the input. But be careful! Poorly designed feedback can make things worse, causing violent oscillations. The system can become unstable.

How can we guarantee stability? This is where a beautiful piece of 19th-century complex analysis, the Argument Principle, comes to the rescue in the form of the ​​Nyquist Stability Criterion​​. By treating the system's "loop transfer function" $L(s)$—a proper rational function—as a mapping in the complex plane, we can determine the stability of the entire closed-loop system. The criterion states that the number of unstable poles in the final system ($Z$) is equal to the number of unstable poles you started with ($P$) plus the number of times the Nyquist plot of $L(s)$ encircles the critical point $(-1, 0)$ ($N$). This gives us the famous equation $Z = P + N$. It feels like magic: by tracing a path in a mathematical space, we can predict whether a real-world machine will be stable or tear itself apart. We design controllers—themselves described by proper rational functions—to shape the Nyquist plot and steer it clear of the dreaded -1 point, thereby engineering stability.

This design philosophy is not confined to the analog world. Most modern controllers are digital, implemented on microprocessors. Here too, proper rational functions are central. A continuous-time controller, like a lead compensator $C(s) = K \frac{T s + 1}{\alpha T s + 1}$, can be systematically translated into a discrete-time algorithm that a computer can execute. A common method is the bilinear transform, which essentially replaces the continuous operator $s$ with a discrete-time equivalent. This process transforms the proper rational function in $s$ into a new proper rational function in the discrete variable $z$. This allows the entire powerful framework of control design to be ported into the digital domain, forming the bedrock of modern automation.

What about systems that aren't so well-behaved? Many real-world processes involve pure time delays. Think of the time it takes for hot water to travel from the heater to your showerhead. In the Laplace domain, this delay corresponds to a factor of $e^{-s\tau}$, which is a transcendental function, not a rational one. A system with a delay is technically infinite-dimensional and doesn't fit our neat framework.

Does this mean our beautiful theory breaks down? Not at all. This is where we see its true power and flexibility. If we cannot analyze the exact system, we can create a finite-dimensional approximation of it that is a proper rational function. The ​​Padé approximation​​ is a brilliant technique for finding a rational function that mimics the behavior of a transcendental function like $e^{-s\tau}$ remarkably well. By replacing the delay term with its Padé approximant, we get an overall transfer function that is a high-order, but perfectly standard, proper rational function. We can then apply all our familiar tools—pole-zero analysis, stability tests, controller design—to this approximate model. This allows us to bring the unruly, infinite-dimensional reality of delays into the tractable, finite-dimensional world of rational functions.

Finally, let's take a step back and appreciate the deep mathematical structure underlying all these applications. Consider all the possible voltage responses of a passive LTI circuit as it settles down from some initial state. If we know that the Laplace transform of any such response is a strictly proper rational function with a fixed denominator polynomial of degree $n$, say $P(s) = s^n + \alpha_{n-1}s^{n-1} + \dots + \alpha_0$, what can we say about the set of these time-domain voltage functions? It turns out this set forms an $n$-dimensional ​​vector space​​. This is a stunning revelation. The complex dynamics of the circuit are governed by the same simple rules of linear algebra that describe vectors in space. The degree of the denominator polynomial tells you the dimension of this abstract space of behaviors! Furthermore, deep properties of the system are encoded in the coefficients of the polynomial. For instance, the sum of all the system's natural frequencies (the poles $\lambda_i$) is simply given by the negative of the coefficient of the second-highest power term, $-\alpha_{n-1}$. This is a direct consequence of Vieta's formulas, a result from high school algebra, now seen to govern the physics of complex circuits.

From ensuring a circuit won't burn out to guaranteeing a rocket flies straight, and from designing digital filters to revealing the hidden algebraic structure of physical laws, the proper rational function is an indispensable tool. It is a testament to the "unreasonable effectiveness of mathematics," where a simple constraint on the degrees of two polynomials unfolds into a rich, powerful, and beautiful framework for understanding and shaping the world around us.