I was talking to someone about the deposit we can get from soda cans and bottles when we returned them to the store for recycling and she remarked that the containers do a good job telling us that they are worth some money, because on the cans are written the words "ME 5¢". Sometimes the messages are quite friendly; on one bottle it says "HI ME 5¢".
Wednesday, March 19, 2014
Sunday, March 16, 2014
Multiplication tricks
When I was a young boy I was shown the following trick to square certain numbers very quickly. In particular, given a 2-digit number where the unit digit is 5, the square of this number can be easily computed with the following rule:
If the tens digit is $a$ and the unit digit is $5$, the square can be found by computing $a(a+1)$ and appending $25$ to the result.
For instance $75^2$ is computed by multiplying $7$ and $8$ ($= 56$)and adding $25$ at the end to get $5625$. This also work with $3$-digit numbers (and longer numbers). For instance, to compute $115^2$, we compute $11\times 12 = 132$ and get $13225$.
To show why this is true is a simple exercise. A number with tens digit $a$ and unit digit $5$ is equal to $10a+5$. Squaring this number results in $(10a+5)^2 = 100a^2 + 100a + 25 = 100a(a+1) + 25$ which corresponds to the rule above.
This same argument is true when multiplying two 2-digit numbers where the tens digits are equal and the unit digits add up to 10. In this case we are multiplying $10a+b$ and $10a+c$ where $b+c =10$. The product $(10a+b)(10a+c)$ is equal to $100a^2+100a+bc = 100a(a+1)+bc$. So the corresponding rule is:
Multiply $a$ with $a+1$ and append $bc$ at the end.
The only caveat is that when $b=1$ and $c = 9$, you append $09$ (rather than just $9$) at the end. For instance to compute $83\times 87$, we get $8\times 9 = 72$ and append $3\times 7$ to get $7221$. Similarly for $124\times 126$, we compute $12\times 13 = 156$ and append $24$ to get $15624$.
You can use a similar trick to multiply two 2-digit number where the tens digits add up to 10 and the unit digits are the same, i.e. multiplying $10b+a$ with $10c+a$ where $b+c = 10$. The product is
$$(10b+a)(10c+a) = 100bc + 100a + a^2 = 100(bc+a) + a^2$$
Thus the corresponding rule is:
Multiply $b$ with $c$, add $a$ and then append $a^2$ to the result.
Again, if $a^2$ has only one digit, prefix it with a $0$ before appending it to $bc+a$. For example, to compute $43\times 63$, we compute $4\times 6 + 3 = 27$ and append $3^2 = 09$ to obtain $2709$.
If the tens digit is $a$ and the unit digit is $5$, the square can be found by computing $a(a+1)$ and appending $25$ to the result.
For instance $75^2$ is computed by multiplying $7$ and $8$ ($= 56$)and adding $25$ at the end to get $5625$. This also work with $3$-digit numbers (and longer numbers). For instance, to compute $115^2$, we compute $11\times 12 = 132$ and get $13225$.
To show why this is true is a simple exercise. A number with tens digit $a$ and unit digit $5$ is equal to $10a+5$. Squaring this number results in $(10a+5)^2 = 100a^2 + 100a + 25 = 100a(a+1) + 25$ which corresponds to the rule above.
This same argument is true when multiplying two 2-digit numbers where the tens digits are equal and the unit digits add up to 10. In this case we are multiplying $10a+b$ and $10a+c$ where $b+c =10$. The product $(10a+b)(10a+c)$ is equal to $100a^2+100a+bc = 100a(a+1)+bc$. So the corresponding rule is:
Multiply $a$ with $a+1$ and append $bc$ at the end.
The only caveat is that when $b=1$ and $c = 9$, you append $09$ (rather than just $9$) at the end. For instance to compute $83\times 87$, we get $8\times 9 = 72$ and append $3\times 7$ to get $7221$. Similarly for $124\times 126$, we compute $12\times 13 = 156$ and append $24$ to get $15624$.
You can use a similar trick to multiply two 2-digit number where the tens digits add up to 10 and the unit digits are the same, i.e. multiplying $10b+a$ with $10c+a$ where $b+c = 10$. The product is
$$(10b+a)(10c+a) = 100bc + 100a + a^2 = 100(bc+a) + a^2$$
Thus the corresponding rule is:
Multiply $b$ with $c$, add $a$ and then append $a^2$ to the result.
Again, if $a^2$ has only one digit, prefix it with a $0$ before appending it to $bc+a$. For example, to compute $43\times 63$, we compute $4\times 6 + 3 = 27$ and append $3^2 = 09$ to obtain $2709$.
Sunday, March 9, 2014
Beyond superlatives
My son and I went skiing the other day and he asked me why double black diamond trails which are billed as "Extremely difficult" are more difficult than black diamond trails which are described as "Most difficult". He said, "How could there be something more difficult than 'most difficult'?" Is "most difficult" more difficult or less difficult than "extremely difficult"? According to the New Oxford American Dictionary, "most" is defined as "greatest in amount or degree", whereas "extreme" is defined as "reaching a high or the highest degree". So there is a sense where "extremely difficult" is the same as "most difficult", but there is an interpretation of the definitions where "extremely difficult" is less difficult than "most difficult". Perhaps the wording in the ski resorts should be changed to more clearly indicate which trails are more difficult.
Wednesday, March 5, 2014
Fractals in a fairy tale
We were all watching the Oscars last Sunday and my son and I noticed that the song "Let it go" from the movie Frozen mentions fractals. Fractal is a mathematical term and we were quite surprised but delighted to see it used in a children's song. One of the most famous fractals is the Mandelbrot set, named after Benoit Mandelbrot, the father of fractals. When I got my Commodore VIC-20 in the 1980's, my brother and I (and I am sure countless other kids playing with their personal computers) wrote a program to compute the Mandelbrot set, and zoom in and display various sections of it, (not knowing that 15 years later, I will work at the same place where Mandelbrot worked and have a chance to have lunch with him on several occasions). At the time, it takes hours to compute a picture of the Mandelbrot set.
The definition of the Mandelbrot set is quite simple. Points on the plane correspond to complex numbers and each point is used as a parameter $c$ in the equation $x_{n+1}= x_n^2 + c$. The iterates $x_i$ starting from $x_0 = 0$ remain bounded if and only if $c$ belongs to the Mandelbrot set. Since it may take many iterations before the iterates start to diverge, practically the computer program assumes the iterates are bounded if it hasn't diverged after a large number of iterations.
The definition of the Mandelbrot set is quite simple. Points on the plane correspond to complex numbers and each point is used as a parameter $c$ in the equation $x_{n+1}= x_n^2 + c$. The iterates $x_i$ starting from $x_0 = 0$ remain bounded if and only if $c$ belongs to the Mandelbrot set. Since it may take many iterations before the iterates start to diverge, practically the computer program assumes the iterates are bounded if it hasn't diverged after a large number of iterations.
Sunday, December 1, 2013
The perils of dividing by zero
In a middle school math competition a few years ago, the following question was asked: given an n by m matrix of numbers whose average row sum is R and average column sum is C, what is the ratio R/C. Since R = T/n and C = T/m, where T is the sum of all the entries in the matrix, it would appear that R/C = m/n. This is the expected answer for this question and is correct in almost all cases. But one must be careful when dividing by a number to ensure that the divisor is not zero. We assume the matrix is nondegenerate, i.e. m > 0 and n > 0, so dividing by n and m is okay and R, C are well-defined. However, it is possible that C = 0, in which case R/C is not well-defined. This cases occurs when T = 0, in which case R is also 0. A simple example is when the matrix consists of all zeros. The use of undefined operations such as dividing by 0 can lead to an alleged "proof" of 1 = 2 which is not good as we do not want an inconsistent theory.
Thursday, November 28, 2013
Opposing views
The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel in 2013 was awarded to Eugene Fama, Lars Peter Hansen and Robert Schiller. In an article in the New York Times, Professor Schiller commented on how he disagrees with the other 2 winners of the Prize even though they are all cited for their work on "empirical analysis of asset prices". This reminds me that in virtually all sciences, even in hard sciences such as physics, there are disagreements on the validity of various theories. One notable exception is in mathematics, where a formal proof of a theorem from antiquity is still accepted today. Of course there are disagreements in mathematics, such as whether the proof of the four color theorem is valid, since it was only checkable using a computer program, but that is a debate whether there is a mistake in the proof, not whether the proof methodology is invalid. Mathematical logic is a branch of mathematics that contemplates questions about consistency and completeness of formal mathematical proofs. For instance, is it possible to provide a formal proof of a statement and its negation? Consistent theories means that you cannot prove both a statement and its negation. A complete theory means that for every statement you can prove the statement or its negation. Most practicing mathematicians are probably not worried much about consistency and completeness of the mathematical framework they are working in, they just assume that it is consistent and complete. It was quite a shock when Kurt Gödel proved in 1931 that various models of arithmetic are incomplete, i.e. there are true statements about natural numbers that cannot be proven with these axiomatic theories.
Sunday, October 27, 2013
Orange and orange
In a previous post, I looked at the sound for the word tea and the letter T in various European languages and found that they are the same in several languages. Recently I noticed that the fruit "orange" and the color "orange" are described by the same word in English. In what other languages are they described by the same word?
This is true in Chinese (both Mandarin and Cantonese), Spanish and French, but not Dutch. In the Dutch language, the color is called oranje (coincidentally orange is the national color of the Netherlands because it is also the name of the Dutch Royal family, derived from the name of a French principality belonging to their ancestor Willem de Zwijger), while the fruit is called sinaasappel or appelsien. I wonder what other languages have different words for these two concepts as well?
This is true in Chinese (both Mandarin and Cantonese), Spanish and French, but not Dutch. In the Dutch language, the color is called oranje (coincidentally orange is the national color of the Netherlands because it is also the name of the Dutch Royal family, derived from the name of a French principality belonging to their ancestor Willem de Zwijger), while the fruit is called sinaasappel or appelsien. I wonder what other languages have different words for these two concepts as well?
Saturday, October 26, 2013
Theorem of the alternative and logical equivalence
In the theory of linear equalities and inequalities, several fundamental results are formulated as a theorem of the alternative: two alternatives are given, and exactly one of the alternatives is satisfied at any one time. A well known example is Farkas' lemma (1894):
For a real matrix $A$ and real vector $b$, exactly one of the following conditions is true:
where we say a real vector is (non)negative if all its entries are (non)negative. Consider a theorem of the alternative given in the following form: given two alternatives $P$ or $Q$, only one of them is true at any time, no more and no less. The truth table for this conclusion (denoted as $Z$) is then:
This means that $Z$ is equal to $P\oplus Q$ which is equivalent to $P\Leftrightarrow \neg Q$ which is equivalent to $\neg P\Leftrightarrow Q$.
Thus the theorem
Exactly one of the two statements $P$ and $Q$ is true
can also be restated as:
P is true if and only if Q is false
or as:
P is false if and only if Q is true
Setting the 2 statements in Farkas' Lemma above as $P$ and $Q$, then
$P\Leftrightarrow \neg Q$ is written as
Let $A$ be a real matrix and let $b$ be a real vector. Then there exists a nonnegative real vector $x$ such that $Ax=b$ if and only if for each row vector $y$ either $yA$ is not a nonnegative vector or $yb$ is nonnegative.
Given that $P\Rightarrow Q$ is logically equivalent to $\neg P \vee Q$, this can be rewritten as
Let $A$ be a real matrix and let $b$ be a real vector. Then there exists a nonnegative real vector $x$ such that $Ax=b$ if and only if $yb$ is nonnegative for each row vector $y$ such that $yA$ is a nonnegative vector.
which is the form of Farkas' Lemma as stated in A. Schrijver, Theory of Linear and Integer Programming, Wiley, 1998.
For a real matrix $A$ and real vector $b$, exactly one of the following conditions is true:
- There exists a real nonnegative vector $x$ such that $Ax = b$
- There exists a real row vector y such that $yA$ is nonnegative and $yb$ is negative.
where we say a real vector is (non)negative if all its entries are (non)negative. Consider a theorem of the alternative given in the following form: given two alternatives $P$ or $Q$, only one of them is true at any time, no more and no less. The truth table for this conclusion (denoted as $Z$) is then:
| P | Q | Z |
|---|---|---|
| 0 | 0 | 0 |
| 1 | 0 | 1 |
| 0 | 1 | 1 |
| 1 | 1 | 0 |
This means that $Z$ is equal to $P\oplus Q$ which is equivalent to $P\Leftrightarrow \neg Q$ which is equivalent to $\neg P\Leftrightarrow Q$.
Thus the theorem
Exactly one of the two statements $P$ and $Q$ is true
can also be restated as:
P is true if and only if Q is false
or as:
P is false if and only if Q is true
Setting the 2 statements in Farkas' Lemma above as $P$ and $Q$, then
$P\Leftrightarrow \neg Q$ is written as
Let $A$ be a real matrix and let $b$ be a real vector. Then there exists a nonnegative real vector $x$ such that $Ax=b$ if and only if for each row vector $y$ either $yA$ is not a nonnegative vector or $yb$ is nonnegative.
Given that $P\Rightarrow Q$ is logically equivalent to $\neg P \vee Q$, this can be rewritten as
Let $A$ be a real matrix and let $b$ be a real vector. Then there exists a nonnegative real vector $x$ such that $Ax=b$ if and only if $yb$ is nonnegative for each row vector $y$ such that $yA$ is a nonnegative vector.
which is the form of Farkas' Lemma as stated in A. Schrijver, Theory of Linear and Integer Programming, Wiley, 1998.
Wednesday, October 16, 2013
Gravity
Over the weekend, my family and I went to see the movie "Gravity" in IMAX 3D after hearing all the buzz about it. Admittedly there are some scientific errors in the movie (my son was whispering to me during the movie why Clooney needs to detach himself from Bullock to save her), the movie was a thrilling ride and made you feel what it is like to be adrift in space. I have always been fascinated by spaceflight ever since I read the books "The right stuff" and "2001: a space odyssey" (I have also seen the movies, of course) as a boy. The movie also reminds me of a common misconception we have when we see footage of astronauts in space: that there is no gravity in space (hence the term "zero gravity"). We see that as soon as the spacecraft escapes Earth's atmosphere, things like pens start to float around in the capsule. While is it true that in deep space, far away from other celestial bodies, the effect of their gravitational pull can be minuscule, all manned spaceflights are still relatively close to Earth and will feel its gravitational pull.
Consider astronauts at the International Space Station (ISS). They experience a gravitational pull similar in strength to what we experience on the ground. Recall Sir Isaac Newton's celebrated equation describing the forces between two bodies:
$$ F = \frac{Gm_1m_2}{r^2}$$
where $G\approx 6.67384 \times 10^{-14}m^3g^{-1}s^{-2}$.
Using Newton's second law of motion $F=ma$, the acceleration an object is subjected to due to another object with mass $m$ and a distance $r$ away is $\frac{Gm}{r^2}$. To calculate the acceleration induced by the gravitational force you feel from the Earth, if we assume the Earth is a sphere with a spherically symmetric density, then $r$ is the distance between you and the center of the Earth, approximately 6.371 Mm. The mass of the Earth $M_e$ is approximately $5.972\times 10^{27}g$, resulting in a constant $ \frac{G M_e}{r^2} \approx 9.82 ms^{-2}$.
The ISS is about 0.37 Mm above the ground, increasing $r$ to be about 6.731 Mm. The resulting constant is then approximately $8.8$, a 10% reduction in weight for everyone on board. However, the astronauts do not feel this force as they are orbiting the Earth, as they are essentially in continuous free fall. In other words, without gravity, the spacecraft would move in a straight line, and the speed of the spacecraft is such that the acceleration needed to change the trajectory to an orbit around the Earth is exactly equal to the acceleration induced by the gravitational force. It is a nice exercise in calculus (which by the way was also invented by Isaac Newton through his introduction of Fluxions and Fluents) to show this.
Since everything is relative, we must also note that we (along with the Earth) are orbiting the Sun as well at a speed of over 100Mm/hr. The distance from the Earth to the Sun is approximately $1.496\times 10^{11}m$ and the mass of the Sun is approximately $1.989\times 10^{33}g$. Applying Newton's formula above, we get $ \frac{G M_s}{r^2} \approx 0.00593$, which is much smaller than the constant for Earth. Since we and the Earth are in the same orbit (and thus in free fall) around the Sun, we do not notice the effects of the Sun's gravitational pull.
When I told my wife that we are moving around the Sun at high speed, she quipped "So we don't have a sedentary lifestyle then."
Consider astronauts at the International Space Station (ISS). They experience a gravitational pull similar in strength to what we experience on the ground. Recall Sir Isaac Newton's celebrated equation describing the forces between two bodies:
$$ F = \frac{Gm_1m_2}{r^2}$$
where $G\approx 6.67384 \times 10^{-14}m^3g^{-1}s^{-2}$.
Using Newton's second law of motion $F=ma$, the acceleration an object is subjected to due to another object with mass $m$ and a distance $r$ away is $\frac{Gm}{r^2}$. To calculate the acceleration induced by the gravitational force you feel from the Earth, if we assume the Earth is a sphere with a spherically symmetric density, then $r$ is the distance between you and the center of the Earth, approximately 6.371 Mm. The mass of the Earth $M_e$ is approximately $5.972\times 10^{27}g$, resulting in a constant $ \frac{G M_e}{r^2} \approx 9.82 ms^{-2}$.
The ISS is about 0.37 Mm above the ground, increasing $r$ to be about 6.731 Mm. The resulting constant is then approximately $8.8$, a 10% reduction in weight for everyone on board. However, the astronauts do not feel this force as they are orbiting the Earth, as they are essentially in continuous free fall. In other words, without gravity, the spacecraft would move in a straight line, and the speed of the spacecraft is such that the acceleration needed to change the trajectory to an orbit around the Earth is exactly equal to the acceleration induced by the gravitational force. It is a nice exercise in calculus (which by the way was also invented by Isaac Newton through his introduction of Fluxions and Fluents) to show this.
Since everything is relative, we must also note that we (along with the Earth) are orbiting the Sun as well at a speed of over 100Mm/hr. The distance from the Earth to the Sun is approximately $1.496\times 10^{11}m$ and the mass of the Sun is approximately $1.989\times 10^{33}g$. Applying Newton's formula above, we get $ \frac{G M_s}{r^2} \approx 0.00593$, which is much smaller than the constant for Earth. Since we and the Earth are in the same orbit (and thus in free fall) around the Sun, we do not notice the effects of the Sun's gravitational pull.
When I told my wife that we are moving around the Sun at high speed, she quipped "So we don't have a sedentary lifestyle then."
Saturday, September 28, 2013
A short tutorial of nonlinear circuit theory - part 2
In a previous post, we look at Kirchoff's laws and device constituency relations to derive state equations of an electrical circuit. Today we will look at some simple circuits and their corresponding state equations. The simplest dynamic circuit must have at least one capacitor (or inductor). Let us consider a circuit with one linear capacitor and one nonlinear resistor.
The state equations can then be written as
\begin{equation} \frac{dq}{dt} = -g(v) = -g(q/C) \end{equation}
The trajectories can be visualized on the dynamic route, where we plot $\frac{dq}{dt}$ against $q$ and the arrows indicate how the trajectory will evolve:
How do we resolve such nonphysical situations? By adding parasitic elements. Suppose we add a parasitic inductor L in series to the nonlinear resistor.
The resulting state equations will be given by: \[ \begin{array}{lcl}\frac{dq}{dt} & = & i_L = \frac{\phi}{L} \\ \frac{d\phi}{dt} & = & v_L = v_R - v_C = f(i_R) - q/c = f(\phi/L) - q/c \end{array} \]
By introducing the variable $i_C = -i_L = -\phi/L$ we get the state equations: \[ \begin{array}{lcl}\frac{dq}{dt} & = & i_C \\ \frac{di_C}{dt} & = & -\frac{1}{L}(q/C - f(-i_C)) \end{array} \] By drawing the phase portrait, we see that at the curve $q = Cf(-i_C)$, $di_C/dt = 0$. In fact, we have an oscillator.
For the nonlinear resistor with a constituency relation as shown above, the existence of the limit cycle can be proved more rigorously by using the Poincaré -Bendixson theorem and the fact that the system has a bounded trapping region and a single equilibrium point that is unstable.
For $f$ an odd function, a second order system in this form is a Liénard equation [5] and the Levinson-Smith theorem [6] allows us to prove the existence of a stable limit cycle when the nonlinear resistor is of a certain form:
Theorem [Levinson-Smith 1942]: Consider the second order equations
\[ \begin{array}{lcl} dx/dt & = & y - F(x) \\ dy/dt & = & -g(x) \end{array} \] Suppose that
Note that our circuit above satisfies the hypothesis of the Levinson-Smith theorem if our nonlinear resistor has the form $v = f(i)$ where $f$ is a cubic polynomial of the form $x^3-ax$ for $a>0$. In particular, we get exactly the Van der Pol equation $x'' + \alpha (x^2-1)x' + x = 0$ [8], an equation used to model oscillations occurring in a circuit involving vacuum tubes.
Another condition for guaranteeing the existence of a unique periodic solution is the following result [7]:
Theorem: Consider the second order equations
\[ \begin{array}{lcl} dx/dt & = & y - F(x) \\ dy/dt & = & -g(x) \end{array} \] Suppose that
References:
[1] L. O. Chua, Introduction to nonlinear network theory, McGraw-Hill, 1969.
[2] L. O. Chua, C. A. Desoer, E. S. Kuh, Linear and nonlinear circuits, McGraw-Hill, 1987.
[3] L. O. Chua, Memristor-The missing circuit element, IEEE Transactions on Circuit Theory, Sept. 1971, vol. 18, no. 5, pp. 507-519.
[4] D. B. Strukov, G. S. Snider, D. R. Stewart1 & R. S. Williams, The missing memristor found, Nature 453, 1 May 2008, pp. 80-83.
[5] A. Liénard, Étude des oscillations entretenues, Revue génér. de l'électr. 23, (1928), 901 - 902; 906 - 954.
[6] N. Levinson, O. Smith, A general equation for relaxation oscillations Duke Math. J. 9, (1942), 382 - 403.
[7] D. W. Jordan and P. Smith, Nonlinear Ordinary Differential Equations, Oxford University Press, 1977.
[8] Van der Pol, B. and Van der Mark, J., Frequency demultiplication, Nature, vol. 120, pp. 363-364, 1927.
Case 1: Assume that the resistor is voltage-controlled ($i = g(v)$)
The state equations can then be written as
\begin{equation} \frac{dq}{dt} = -g(v) = -g(q/C) \end{equation}
The trajectories can be visualized on the dynamic route, where we plot $\frac{dq}{dt}$ against $q$ and the arrows indicate how the trajectory will evolve:
Case 2: assume that the resistor is current controlled ($v = f(i)$).
We have $v = f(i) = f(-dq/dt)$. Since $v = q/C$ we get $q = Cf(-dq/dt)$. Using the dynamic route interpretation and depending on the shape of $f$, it is possible to find impasse points, where the trajectory does not know what to do.How do we resolve such nonphysical situations? By adding parasitic elements. Suppose we add a parasitic inductor L in series to the nonlinear resistor.
The resulting state equations will be given by: \[ \begin{array}{lcl}\frac{dq}{dt} & = & i_L = \frac{\phi}{L} \\ \frac{d\phi}{dt} & = & v_L = v_R - v_C = f(i_R) - q/c = f(\phi/L) - q/c \end{array} \]
By introducing the variable $i_C = -i_L = -\phi/L$ we get the state equations: \[ \begin{array}{lcl}\frac{dq}{dt} & = & i_C \\ \frac{di_C}{dt} & = & -\frac{1}{L}(q/C - f(-i_C)) \end{array} \] By drawing the phase portrait, we see that at the curve $q = Cf(-i_C)$, $di_C/dt = 0$. In fact, we have an oscillator.
For the nonlinear resistor with a constituency relation as shown above, the existence of the limit cycle can be proved more rigorously by using the Poincaré -Bendixson theorem and the fact that the system has a bounded trapping region and a single equilibrium point that is unstable.
For $f$ an odd function, a second order system in this form is a Liénard equation [5] and the Levinson-Smith theorem [6] allows us to prove the existence of a stable limit cycle when the nonlinear resistor is of a certain form:
Theorem [Levinson-Smith 1942]: Consider the second order equations
\[ \begin{array}{lcl} dx/dt & = & y - F(x) \\ dy/dt & = & -g(x) \end{array} \] Suppose that
- $F$ is an odd $C^1$ function (i.e. $F(-x) = -F(x)$),
- $g$ is an odd $C^0$ function with $xg(x)>0$ for all $x\neq 0$,
- There is a constant $a>0$ such that $F(x) < 0$ on $0 < x < a$, $F(x) > 0$ on $x > a$, and $F'(x) > 0$ on $x>a$,
- $\int_0^x g(s)ds \rightarrow\infty$ as $|x|\rightarrow \infty$ and $F(x) \rightarrow\infty$ as $x\rightarrow \infty$.
Note that our circuit above satisfies the hypothesis of the Levinson-Smith theorem if our nonlinear resistor has the form $v = f(i)$ where $f$ is a cubic polynomial of the form $x^3-ax$ for $a>0$. In particular, we get exactly the Van der Pol equation $x'' + \alpha (x^2-1)x' + x = 0$ [8], an equation used to model oscillations occurring in a circuit involving vacuum tubes.
Another condition for guaranteeing the existence of a unique periodic solution is the following result [7]:
Theorem: Consider the second order equations
\[ \begin{array}{lcl} dx/dt & = & y - F(x) \\ dy/dt & = & -g(x) \end{array} \] Suppose that
- $F$ is an odd $C^1$ function and is zero only at $x=0$,$x=a$ and $x=-a$ for some $a>0$,
- $F(x) \rightarrow \infty$ monotonically as $x\rightarrow \infty$ for $x >a$.
- $g$ is an odd $C^0$ function with $xg(x)>0$ for all $x\neq 0$,
References:
[1] L. O. Chua, Introduction to nonlinear network theory, McGraw-Hill, 1969.
[2] L. O. Chua, C. A. Desoer, E. S. Kuh, Linear and nonlinear circuits, McGraw-Hill, 1987.
[3] L. O. Chua, Memristor-The missing circuit element, IEEE Transactions on Circuit Theory, Sept. 1971, vol. 18, no. 5, pp. 507-519.
[4] D. B. Strukov, G. S. Snider, D. R. Stewart1 & R. S. Williams, The missing memristor found, Nature 453, 1 May 2008, pp. 80-83.
[5] A. Liénard, Étude des oscillations entretenues, Revue génér. de l'électr. 23, (1928), 901 - 902; 906 - 954.
[6] N. Levinson, O. Smith, A general equation for relaxation oscillations Duke Math. J. 9, (1942), 382 - 403.
[7] D. W. Jordan and P. Smith, Nonlinear Ordinary Differential Equations, Oxford University Press, 1977.
[8] Van der Pol, B. and Van der Mark, J., Frequency demultiplication, Nature, vol. 120, pp. 363-364, 1927.
Subscribe to:
Posts (Atom)





