web statisticsRealtime website tracking software

Of Bernoullied Euler and Eulered Lagrange and Other Gianted Giants - Part III

Posted on July 22, 2021

Keywords: Euler-Lagrange equation, Optimization, Genius

We can now finally unveil the Euler-Lagrange equations of motion with just a little bit more calculation. Recall that kinetic energy of a system of \(k\) particles is given by

\[K=\sum\limits_{i=1}^{k}\frac{1}{2}m_i\boldsymbol{v}_i^T\boldsymbol{v}_i\]

Now the following computations should be straightforward to understand

\[\frac{\partial K}{\partial q_j}=\sum\limits_{i=1}^{k}\frac{1}{2}m_i\left(\frac{\partial\boldsymbol{v}_i}{\partial q_j}\right)^T\boldsymbol{v}_i+\sum\limits_{i=1}^{k}\frac{1}{2}m_i\boldsymbol{v}_i^T\frac{\partial\boldsymbol{v}_i}{\partial q_j}\]

\[=\sum\limits_{i=1}^{k}m_i\boldsymbol{v}_i^T\frac{\partial\boldsymbol{v}_i}{\partial q_j}=\sum\limits_{i=1}^{k}m_i\boldsymbol{v}_i^T\frac{d}{d t}\left(\frac{\partial\boldsymbol{r}_i}{\partial q_j}\right)\]

\[\frac{\partial K}{\partial\dot{q_j}}=\sum\limits_{i=1}^{k}\frac{1}{2}m_i\left(\frac{\partial\boldsymbol{v}_i}{\partial \dot{q_j}}\right)^T\boldsymbol{v}_i+\sum\limits_{i=1}^{k}\frac{1}{2}m_i\boldsymbol{v}_i^T\frac{\partial\boldsymbol{v}_i}{\partial \dot{q_j}}\]

\[=\sum\limits_{i=1}^{k}m_i\boldsymbol{v}_i^T\frac{\partial\boldsymbol{v}_i}{\partial \dot{q_j}}=\sum\limits_{i=1}^{k}m_i\boldsymbol{v}_i^T\frac{\partial\boldsymbol{r}_i}{\partial q_j}\]

Now recall from the previous article that

\[\sum\limits_{i=1}^{k} m_{i} \ddot{\boldsymbol{r}}_{i}^{T} \frac{\partial \boldsymbol{r}_{i}}{\partial q_{j}}=\sum\limits_{i=1}^{k}\left\{\frac{d}{d t}\left[m_{i} \dot{\boldsymbol{r}}_{i}^{T} \frac{\partial \boldsymbol{r}_{i}}{\partial q_{j}}\right]-m_{i} \dot{\boldsymbol{r}}_{i}^{T} \frac{d}{d t}\left[\frac{\partial \boldsymbol{r}_{i}}{\partial q_{j}}\right]\right\}\]

Substituting \(\dot{\boldsymbol{r}}_{i}=\boldsymbol{v}_i\) above, we obtain

\[\sum\limits_{i=1}^{k} m_{i} \ddot{\boldsymbol{r}}_{i}^{T} \frac{\partial \boldsymbol{r}_{i}}{\partial q_{j}}=\sum\limits_{i=1}^{k}\left\{\frac{d}{d t}\left[m_{i} \boldsymbol{v}_i^{T} \frac{\partial \boldsymbol{r}_{i}}{\partial q_{j}}\right]-m_{i} \boldsymbol{v}_i^{T} \frac{d}{d t}\left[\frac{\partial \boldsymbol{r}_{i}}{\partial q_{j}}\right]\right\}\]

And using the tricks from the previous article, we obtain

\[\sum\limits_{i=1}^{k} m_{i} \ddot{\boldsymbol{r}}_{i}^{T} \frac{\partial \boldsymbol{r}_{i}}{\partial q_{j}}=\sum\limits_{i=1}^{k}\left\{\frac{d}{d t}\left[m_{i} \boldsymbol{v}_i^{T} \frac{\partial\boldsymbol{r}_i}{\partial\dot{q_j}}\right]-m_{i} \boldsymbol{v}_i^{T} \frac{\partial\boldsymbol{v}_i}{\partial q_j}\right\}=\sum\limits_{i=1}^{k}\frac{d}{d t}\left[m_{i} \boldsymbol{v}_i^{T} \frac{\partial\boldsymbol{r}_i}{\partial\dot{q_j}}\right]-\sum\limits_{i=1}^{k}m_{i} \boldsymbol{v}_i^{T} \frac{\partial\boldsymbol{v}_i}{\partial q_j}\]

The above expression can be simplified in terms of the kinetic energy as follows

\[\sum\limits_{i=1}^{k} m_{i} \ddot{\boldsymbol{r}}_{i}^{T} \frac{\partial \boldsymbol{r}_{i}}{\partial q_{j}}=\frac{d}{d t}\left[\frac{\partial K}{\partial\dot{q_j}}\right]-\frac{\partial K}{\partial q_j}\]

Making use of the above expression, we can obtain the following

\[\sum\limits_{i=1}^{k} \dot{p}_{i}^{T} \delta \boldsymbol{r}_{i}=\sum\limits_{j=1}^{n} \left\{\frac{d}{d t} \frac{\partial K}{\partial \dot{q}_{j}}-\frac{\partial K}{\partial q_{j}}\right\} \delta q_{j}\]

To obtain the above, we simply need to recall from the previous article that

\[\sum\limits_{i=1}^{k} \dot{p}_{i}^{T} \delta \boldsymbol{r}_{i}=\sum\limits_{j=1}^{n} \left(\sum\limits_{i=1}^{k} m_{i} \ddot{\boldsymbol{r}}_{i}^{T} \frac{\partial \boldsymbol{r}_{i}}{\partial q_{j}}\right) \delta q_{j}\]

Now since

\[\sum\limits_{i=1}^{k} \boldsymbol{f}_{i}^{T} \delta \boldsymbol{r}_{i}-\sum\limits_{i=1}^{k} \dot{\boldsymbol{p}}_{i}^{T} \delta \boldsymbol{r}_{i}=0\quad \mathrm{and} \quad \sum\limits_{i=1}^{k} \boldsymbol{f}_{i}^{T} \delta \boldsymbol{r}_{i}=\sum\limits_{j=1}^{n} \psi_{j} \delta q_{j}\]

then it is obvious that

\[\sum\limits_{j=1}^{n}\left\{\frac{d}{d t} \frac{\partial K}{\partial \dot{q}_{j}}-\frac{\partial K}{\partial q_{j}}-\psi_{j}\right\} \delta q_{j}=0\]

From the independence of the generalized displacements, it follows that

\[\frac{d}{d t} \frac{\partial K}{\partial \dot{q}_{j}}-\frac{\partial K}{\partial q_{j}}=\psi_{j}\quad\mathrm{for}\quad j=1,\cdots,n\]

The generalized forces \(\psi_j\) acting on each particle may arise from the potential energy fields (electrostatic, gravitational or magnetic) or they may simply represent an externally applied force denoted by \(\tau_j\). Let the net potential energy field acting on particle be denoted by \(P(q)\). From classical mechanics, we have

\[\psi_j=-\frac{\partial P}{\partial q_j}+\tau_j\]

I have frequently thought about the classical fields (whether gravitational or electric) as denoting both a Force and an Energy field. Both the Energy and Force fields are then dual manifestations of a single object called Field. This is in a similar vein to the cumulative distribution function and probability density function being two manifestations of the same probability distribution for some particular random variable. The Energy part aims to quantify the motion and vibrations undergone by a particle in the past while the Force part aims to quantify the motion and vibrations to be experienced by that particle in the future. Thus, Field basically aims to track the trajectory of a particle through spacetime, both its past and future; and why worry about the present because that is already there! At any moment, Energy sums up the effects caused by the Field on the particle in the past and Force is the information that decides what effects the Field will cause to the particle in future. However, there is no unique way to interpret these concepts of science.

We also note that \(\partial P/\partial \dot{q}_j=0\) because potential energy does not depend on the velocity of the particle. Also define Lagrangian of the system as \(\mathcal{L}=K-P\). Then it is not difficult to obtain the equations given below which are referred to as the Euler-Lagrange Equations of Motion.

\[\frac{d}{d t} \frac{\partial \mathcal{L}}{\partial \dot{q}_{j}}-\frac{\partial \mathcal{L}}{\partial q_{j}}=\tau_{j}\quad\mathrm{for}\quad j=1,\cdots,n\]

As long as constraint forces do no work and only holonomic constraints are involved, we can apply this equation. Otherwise when nonholonomic constraints are involved or the frictional forces must be taken into account, the only straightforward route to take is that of falling back to the Newtonian approach which itself is quite enthralling in its own right.1

The historical account behind the discovery of Euler-Lagrange equations is much richer than can be indicated here. The above theory is an equal mix of mechanics and calculus. If you want to dig up the interesting stories from the Physics point-of-view, the best route is to find more mechanics material to read in connection with the Brachistochrone problem. This is a pivotal problem in Euler-Lagrange theory which due to scope-beyondness issues I will skip although it is unfair to discuss the mathematics of Euler-Lagrange equations without discussing the mathematics behind the Brachistochrone problem. And for the calculus point-of-view, need you a better starting point than Dunham!2

Of S.E.C.T. in S.T.E.M.

Shouldn't STEM be regarded as the most beautiful abbreviation but well, beauty lies in the eyes of the beholder3 as they say. Talking about Genius generally comes to be associated with STEM fields though other fields may be equally worth exploring.

To motivate what follows, we need to briefly touch upon certain historical accounts4 surrounding the Brachistochrone problem which was proposed by Johann Bernoulli to the \(``\)most brilliant mathematicians in the world.\("\) While solving the isoperimetric problem (these two problems along with the least resistance aerodynamical problem must definitely form the Holy Trinity in the context of computing trajectories that optimized functionals, the so-called optimal curves), Euler had discovered a geometric approach to these equations much earlier. Lagrange discovered an analytic approach to these equations to which Euler said: \(``\)Your solution to the isoperimetric problem leaves nothing to be desired.\("\) Inspired by Jakob Bernoulli's solution to the Brachistochrone problem, Euler was able to use Lagrange's ideas to rigorously develop the theory behind calculus of variations. At the moment when he was about to publish his detailed report, he resisted the desire to publish his own ideas about this equation (a rare moment in academia, indeed!). He must probably have felt that Lagrange's ideas were fresh and more insightful, and could be applied to many complicated problems in mechanics in a rather elegant way, so he wrote to Lagrange: \(``\)the importance of the matter has led me to outline, with the aid of your light, an analytical solution to which I will give no publicity until you yourself have published the whole of your research, so that I do not take away any part of the glory that is due to you.\("\)

In fact, thanks largely to Lagrange's efforts (who undoubtedly received a huge assistance from Euler but still), Analytical mechanics rose and prospered while being distinctly different from Newtonian mechanics. Though both types of mechanics combine to form classical mechanics, we know that analytical mechanics receives much deeper appreciation from the scientific community today (at least when considering idealized problems which are free from real world complications and where both approaches are equally applicable). Moreover, knowing Euler's hunger to publish,5 one has to appreciate how difficult it must have been for him to resist the desire but at the same time his commitment was stronger towards growth of human knowledge and he believed that Lagrange had developed the theory better. Another profound Brachistochrone episode6 occurred when Newton famously said "I do not love to be dunned and teased by foreigners about mathematical things." How can we forget to talk about Newton here!

In fact, the most vital feature of Newton is excellently identified by Richard Westfall7: "In his age of celebrity, Newton was asked how he had discovered the law of universal gravitation. By thinking on it continually, was the reply. No better characterization of the man can be given, not only in its delineation of a life whose central adventure lay in the world of thought rather than action, but also in its description of his mode of work."

Tikhomirov8 says: "Newton solved the aerodynamical problem completely. Why was Newton's curve relegated to the role of the unfortunate Cinderella for 300 years? Why was Newton's idea not fully understood for so long? Newton's problem belongs to the category of problems of optimal control (which was not rigorously understood back then). Within the framework of this theory Newton's problem has a natural and standard solution. On the other hand, this problem has no natural and standard solution within the framework of the calculus of variations (which was the only sophisticated tool available to attack such problems in those days). Thus this problem put Newton 300 years ahead of his time! Newton's problem is certainly a remarkable mathematical event. For 250 years it seemed likely that it had no physical basis and its solution was absurd. But the \(``\)mistake\("\) of a genius turned out to have been an insight. In a word, hasty judgments are sometimes just that. It can happen that the thought of a genius, which we regard as a mistake carries within it the imprint of truth — a truth clear to him but hidden as yet from us."

Newton's greatest achievement may be the rigorous initiation of human thought process vis-à-vis the abstract concept of Force which he made concrete by allowing us to measure it. But soon Euler, Lagrange and others unleashed their Genius to replace Force with Energy as the standard way of thinking, and then finally Force and Energy got relegated to antiquity by Field when Maxwell, Faraday and others unleashed their Genius.

Genius is always identified only in hindsight and it is always very hard to deconstruct and reconstruct it. But nevertheless a Giant is almost always Gianted by another Giant. The human story has it that cerebral traditions have existed in various forms throughout the human history. Isn't the previously mentioned intellectual circle set up in the thoughtosphere among the Bernoullis, Newton, Euler, Lagrange, and others typifying some sort of a cerebral communal tradition? This circle's activities are summarizable only as fascinating and even fantastic. I choose to call such an event a Singular Event in Cerebral Traditions (S.E.C.T.). One could also whimsically refer to \(``\)Cerebral Traditions\("\) as Cereberations but that would destroy the proposed abbreviation which sounds better.

There is another terrific and more recent academic tradition of this kind but it may probably pale in comparison to the former. I am referring to the mathematicians at RAND Corporation who had triggered a similar wave of excitement. I call the former group "mechanics optimizers" and the latter "programming optimizers". The latter, in my opinion, get dwarfed by the former's intellectual heft (though this position is debatable and contradictable) but the quality of academic tradition and significance of achievements are similar for both. Vasek Chvatal and William Cook say9 that "The RAND Corporation in the early 1950s contained what may have been the most remarkable group of mathematicians working on optimization ever assembled: Arrow, Bellman, Dantzig, Flood, Ford, Fulkerson, Gale, Johnson, Nash, Orchard-Hays, Robinson, Shapley, Simon, Wagner, and other household names. Groups like this need their challenges."

Indeed, they do!

Brachistochrone problem (among others) was the former group's challenge around which they rallied. Travelling salesman problem (among others) was the latter's.10 Without a rallying point, they would not have been able to reinforce each other's thought processes; they would have remained in isolation and struggled with their academic pursuit in their individual capacities, probably failing to accomplish anything of much note and fading into oblivion like any layperson. An academic \(``\)sect\("\) has to form around such a problem.

Now there is a chicken and egg problem. Did the challenge problem bring the Giants together or did they come together and thereafter create a problem to exercise their brains? Given their capabilities, they must surely have intuited beforehand that the rallying problem, if solved satisfactorily, would lead to a paradigm shift and disrupt the status quo in their surrounding academic traditions in a good way. Anyway SECTs happen very rarely and when they occur, they come with an expiry date.

A third striking example that has been well recorded in historical texts seems to be the school of Greek geometers but discussing their legacy goes well beyond the present scope. Would it be fair to say that this SECT reached its E-L summit with Euclid's Elements? But of course, of course! And when Euclid's parallel postulate was picked upon by Gauss, Lobachevsky, and others for further nitpicking, it ushered in the era of non-Euclidean geometry. This derived SECT, in turn, reached its E-L summit via the progress made by Levi-Civita and Riemann. I could also mention the quantum circle but that seems to be just too obvious.

And what about the early thinkers who laid down the first rules of language! Who were they who constructed the rules of Sanskrit language (like Pāṇini and others) and which SECT's genius (like Saptarishi and others) contributed to the compilation of the Rig Veda? How did early societies come up with language, or did a few Geniuses invent language that led to the formation of human societies? Language is probably the only medium for intellectual thinking. There may surely have been other languages before Sanskrit, but Sanskrit seems to stand out among others for some reason. Vedic Sanskrit must have played a very important role in our early intellectual traditions. And though I can only present this as an intuition right now, it seems to be powerful enough to pay attention towards. Imagine the cerebral traditions that must have been required to create a writing system as pivotal as Brahmi and a natural language as sophisticated as Sanskrit! Of the origins of Vedic Sanskrit, I have not found a more insightful description than Sri Aurobindo's ideas below11

\(``\)...words, like plants, like animals, are in no sense artificial products, but growths — living growths of sound with certain seed-sounds as their basis [...] In their beginnings language-sounds were not used to express what we should call ideas; they were rather the vocal equivalents of certain general sensations and emotion-values. It was the nerves and not the intellect which created speech...\("\)

\(``\)...language expresses at first a remarkably small stock of ideas and these are the most general notions possible and generally the most concrete, such as light, motion, touch, substance, extension, force, speed, etc. Afterwards there is a gradual increase in variety of idea and precision of idea. The progression is from the general to the particular, from the vague to the precise...\("\)

\(``\)...one remarkable feature of language in its inception is the enormous number of different meanings of which a single word was capable and also the enormous number of words which could be used to represent a single idea. Afterwards this, tropical luxuriance came to be cut down. The intellect intervened with its growing need of precision, its growing sense of economy...\("\)

Even the most trenchant western academicians will realize the importance of Vedic Sanskrit. Doesn't NASA, building up on Rick Briggs' work, believe Sanskrit to be one of the most suitable natural languages to guide the development of an AI programming language? Even Donald Knuth who may be the most impactful computer scientist the world has ever known, took to learning Sanskrit while developing his formal language and parsing theory. Now if one accepts the importance of Vedic Sanskrit, then one should not fail to realize the importance of the Rig Veda, probably the first poetic composition to worship nature. Its exact date of origin can never be known because of its oral traditions. Acceding importance to Vedic Sanskrit and not according an equal (or more) importance to the Rig Veda is like being an ardent follower of Newtonian achievements but not having the slightest urge to browse over the unabridged version of Philosophiae Naturalis Principia Mathematica. Newton is Newton to us due to his Principia.

A superficial skimming of the Rig Veda may persuade someone not well trained in interpreting ancient poetic compositions, let alone the surrounding sociocultural nuances (like how was the society thinking in those days, or what was the zeitgeist of those times?), to label it as an inconsequential collection of rituals or just praising nature in the anticipation of building such rituals. But that is a blatant mistake and Sri Aurobindo brilliantly challenges such claims. His message is unmistakably scientific and intellectual. If only he used simpler English to communicate, his ideas would not have been that difficult to get access to. Consider, for example, that immortal line by William Wordsworth, \(``\)The Child is father of the Man.\("\) What did he mean by it? What is the deepest interpretation possible? He probably meant that those things that evoked the strongest emotions in him towards nature's beauty during his childhood will stay with him for his entire life. A more pragmatic interpretation would be that an adult is formed as a result of the habits and sensibilities that he develops during his childhood. So, the Child basically refers to childhood of the person and the Man refers to adulthood of the same person. Our entire adult life is spent paying homage to our childhood dreams, aspirations and experiences. Thus, a poem can have several interpretations. And the more you think about it, you come up with interpretations and incisive insights which even the poet may not have been aware of. And that may be the goal of the most sublime poems which is to allude to the subtlest and deepest truths — and a higher truth. Sri Aurobindo shows that something similar is true for the Rig Veda because there is a deeper interpretation possible for this epic poetry. However, it is difficult to recreate its exact hidden meanings because we live in very different times now and have got very different sensibilities today. That is the price we pay for forgetting all about a SECT, or even worse, not noticing a SECT as it sweeps by us.

Such SECTs seem to show up whenever a new language of sufficient intellectual calibre gets created. Although it is not absolutely accepted among experts, I regard mathematics too as a language, even in the strictest sense. Sure, its grammar is quite rigorous and abstract which makes it a rather peculiar language differing drastically from the natural languages. But think about it, mathematicians keep discovering theorems that are more and more abstract while inventing proofs for those theorems that keep getting more and more rigorous. It is their form of novel writing. Maybe more precisely, Mathematical Logic is a meta-language or a language template so that all the major sub-branches of mathematics such as probability and geometry instead are closer to our definition of language, in the same way that (Vedic) Sanskrit is seen to be a language template from which most Indian languages are derived. I also regard Music as a language expressing human emotions. Typically, a certain type of creative person turns to music to express himself because of the inadequacy of such inherent expressiveness in natural languages. Then the vocal cord of the creative or an instrument in the adept hands of the creative produces music. Repeated practice of that music creates in his mind some suitable lyrics. Applying the lyrics to the music produces a song. And finally, repeated singing by the creative may motivate him to devise a dance to accompany the song so that he could express himself better. This is most commonly the total lifespan of such a creative\(/\)artistic process, which in itself is a unique kind of intellectual\(/\)cereberal endeavor and definitely deserves much appreciation. I would, if I could, do a Noam Chomsky to justify both these intuitions but let them be hand-waved for now. However, more importantly, where and how do we find the other SECTs!

When a SECT is identified, the academic community impulsively feels the need to sustain the newly created traditions. In fact, if the goal of academic pursuit is to discover new knowledge which it is, then it feels almost ironical that the urge to sustain the newfound cerebral traditions is generally higher than the urge to search for new knowledge. But it is not actually so because new knowledge cannot simply be created by craving for it. New knowledge appears when the Homo sapiens species is ready for it but to be ready for it, we must cherish and preserve the spirit and practice revealed to us by the SECTs. It is due to this power of the Giants to install their traditions in society in such a strong way that sustaining those traditions in itself becomes the goal of academic pursuit. SECTs inspire many circles to repeat the same feat and such circles may, more frequently than expected, match or even exceed the level of achievement secured by their inspirational figures but will generally fail to generate the same level of amazement in society as the Pioneers. Just like pandemics sweep the planet once in a century or two, so do such SECTs with an unknown (albeit probably lower and more uncertain) frequency. These Genius-types appear to exist in a parallel, nonmaterial, Platonic-spiritual world, and just like during pandemics, a very few of them get infected on some rare occasion by a highly potent and virulent strain; then they infect the rest of the society but in a good way. In fact like Hilbert said, no one shall expel us from the paradise which these very few great men of the past have created for us. I am referring here to those very few who built the world of thoughts. This is not to say that those clever and enterprising doers who excelled in the world of actions thereby lifting us from physical misery and bringing us all into the paradise of material comfort are any smaller; only that this article is not about the doers but about the thinkers both of whom are remarkable coordinates albeit living in different yet related manifolds.

Can we now get back to the original question that was posed, viz. what is a Genius?

And the answer is...

A Genius is what a Genius does.

Okay, Okay! I know that is a non-answer.

But then asking for a definition of Genius is like asking for a definition of God. It's pretty much redundant to do so. You will fail to satisfy yourself with any kind of answer, let alone others. The vital characteristic of a Genius is that his thinking is directed towards Nature instead of people, towards science rather than money, that he professes cerebral thinking (PFC!) over thinking originating from the limbic system. Geniuses can be heroes or titans. Heroes expand a field. Titans expand the field and also understand it with so much clarity that they can explain its intricate concepts lucidly even to laypeople. For example, Faraday was a Hero and Weinberg a Titan.

But much about Genius as with any complicated or complex notion is hazy and fuzzy. Instead of talking about Genius, talking about SECT makes more sense. So how is a SECT created? Go figure!

The best way to understand Genius in a deep sense is by spending time reading the original manuscripts written by Geniuses otherwise it is strongly advisable to steer clear of them and move on with the immediate urgencies of daily lives.12 Now in what spirit should we end this three-part discussion?

Newton had climactically announced in Principia: I feign no hypotheses.

And to paraphrase what Descartes famously said in First Meditation: I bring in for questioning all those things that may be called into doubt.

Well, let that linger on...

References

  1. Friedrich Pfeiffer, and Christoph Glocker, Multibody dynamics with unilateral contacts. John Wiley & Sons, 1996.(back)

  2. William Dunham, The calculus gallery. Princeton University Press, 2018.(back)

  3. Come to think of it, for an observation established several centuries ago, it is still quite a profound statement even by the standards of modern neuroscience.(back)

  4. I cannot summarize this episode better than by quoting from Wikipedia: "In an attempt to outdo his brother, Jakob Bernoulli created a harder version of the brachistochrone problem. In solving it, he developed new methods that were refined by Leonhard Euler into what the latter called (in 1766) the calculus of variations. Joseph-Louis Lagrange did further work that resulted in modern infinitesimal calculus." What started as sibling rivalry ended up becoming something so profound! The incident itself is very revealing about certain aspects of human dynamics.(back)

  5. The Euler Archive. He published an unbelievable 886 papers and books. On a slightly unrelated note, it has long been realized in academia that excellent ideas take a long time to make an impact on society, and those who provide these ideas are mostly recognized much later, even posthumously in some unfortunate cases. So there is this overwhelming sentimental urge to motivate researchers to sell and market their work from early on to avoid such a tragedy. But the fact that gets overlooked is that every player in the field then starts getting into this business, and the very few who will come up with excellent ideas will then anyway get ignored again under the barrage of rolling advertisements. There is no cure for human mediocrity to fail any initiative, however novel the intentions during its conception. What must be realized is that these very few producers of excellent ideas operate on a different level. What they care about foremost is doing their work and recording it for others. When they have published their life's genius, then they may start caring about prestige, status, fame, and so on. But the process of insertion of their ideas into the public domain remains unaffected whether or not there are mechanisms in society to recognize excellent ideas and raise its giver to an exalted pedestal immediately. Also, any idea requires a gestation period to be fully assimilated into society. But with the trend that grips academia (or for that matter any economic enterprise) today, it is in one's best interest to market their work to the hilt (but presentation quality needs to be high — the iron rule of the times we live in); it's no more just publish or perish, it's also publicize or perish!(back)

  6. Again to provide context I must quote from Wikipedia: "Newton stayed up all night to solve it and mailed the solution anonymously by the next post. Upon reading the solution, Bernoulli immediately recognized its author, exclaiming that he 'recognizes a lion from his claw mark.' This story gives some idea of Newton's power, since Johann Bernoulli took two weeks to solve it."(back)

  7. Richard S. Westfall, and Samuel Devons, Never at rest: A biography of Isaac Newton. Cambridge University Press, 1981.(back)

  8. Vladimir Mikhailovich Tikhomirov, Stories about maxima and minima. Universities Press, 1990.(back)

  9. Michael Junger et al., 50 Years of integer programming 1958-2008: From the early years to the state-of-the-art. Springer Science & Business Media, 2009.(back)

  10. To fill in some context, let me plug in some fast-food-style history of Optimization. The field of Optimization is rightfully the bloodline of all problem-solving endeavors.

    One thread of Optimization developing since Newton's times focused on Physics and can be summarized as, unconstrained optimization\(\rightarrow\)Lagrange multipliers\(\rightarrow\)KKT conditions\(\rightarrow\)calculus of variations\(\rightarrow\)optimal control (Pontryagin's maximum principle & Hamilton-Jacobi-Bellman equation)\(\rightarrow\)dynamic programming (Bellman equation).

    Another thread due to Dantzig focusing more on theoretical insights for discrete problems and less on Physics-type insights can be summarized as, linear programming\(\rightarrow\)stochastic programming\(\rightarrow\)ILP\(\rightarrow\)MILP\(\rightarrow\)quadratic programming\(\rightarrow\)convex programming\(\rightarrow\)set-theoretical combinatorial optimization\(\rightarrow\)graph-theoretical combinatorial optimization.

    But this is mostly a structural view of ontologizing the field of optimization, one can also look at it from a functional perspective. Start with optimizing functions of finitely many variables. If the function is unconstrained, Fermat takes care of your problem; if equality constrained, then Lagrange multipliers are at your service; if inequality constrained, then KKT it is! Now let us shift to optimizing functionals. If unconstrained, then apply the Euler equation; if equality constrained, use the Euler-Lagrange equation; if inequality constrained, formulate your problem in terms of controls and specify boundary conditions. Then you have a dynamic optimization problem for which you must resort to the method of optimal control.

    Now, what if we switch from the continuous to the dreadful discrete? When the golden age of algorithmic research dawned upon humanity, this shift was its ultimate beacon. The discrete version of dynamic optimization is just an MDP. For small problem sizes, you have dynamic programming and for large problem sizes, reinforcement learning dominates. Here we must mention multi-armed bandits (MABs) even though it is purely a learning problem but it seamlessly connects here. Naively you could solve it with backward induction (DP) but Gittin's forward induction is better. And then there is this whole field of numerical nonlinear optimization with its convergence proofs and whatnots. Plus a relatively new entrant with a staggering heft, viz. metaheuristic evolutionary optimization algorithms.

    But still, we know who is the reigning king of the discrete world, of course it is mathematical programming — a place where both continuous and discrete optimization truly seem to merge! Dynamical systems in continuous domains are described using differential equations but in the discrete world, we are forced to adopt matrix equations whereby we enter the arena of programming theory. On top of that, the optimization theory today is mostly seen as an algebraic or analytic endeavor but the intuitive geometric beginnings cannot be denied. Linear programming builds a direct connection with geometry via polytopes, hence the importance of these types of optimization problems over all others is understandable. If you have to solve an LP, use Dantzig's simplex algorithm. Khachiyan analyzed the ellipsoid method to rid LP of exponential complexity but polynomial complexity truly felt as having arrived only with Karmarkar's interior-point method. By then Gomory had already created a lot of excitement around cutting planes for ILP and MILP. Initially, the OR community aimed to solve combinatorial optimization problems using mathematical programming techniques like LP, ILP, and MILP which was all good (Dantzig did indeed succeed in solving TSP, an ILP, using his simplex algorithm which was developed for LP formulations) but when mathematicians well-versed in graph theory having exposure to Coxeter-style polytope geometry took a fresh look at these combinatorial optimization problems, then they created this whole trend of searching for efficient graph algorithms. No more branch and bound business or that cutting plane obsession! This may well have been the dawn of the golden age of algorithmic research — a shift from hitherto theoretical style of research wherein computers had impressed an indelible mark upon academic research methodologies for the posterity to abide by.

    If Euler-Lagrange equations marked the climax for the former SECT, then the cutting plane method, and maybe more specifically Concorde's method, marks the culmination of the efforts made by the latter SECT. Or one could also say, its culmination lies in the transformation of programming optimizers into combinatorial optimizers a.k.a. graph optimizers a.k.a. network optimizers who wanted to solve graph-theoretical optimization problems using graph algorithms and this derived SECT reached its E-L summit with Edmonds' Blossom algorithm. To conclude, as was said about Euler, so can also be said about the field of Optimization, thus \(``\)Read Optimization, read Optimization, it is the master of them all.\("\)(back)

  11. Sri Aurobindo, The secret of the Veda: the philological method of the Veda. Lotus Press, 2018.(back)

  12. The approach taken in this article to broach the subject of Geniuses was surface-level and will also appear antiquated. There are more modern and fulsome takes on the aspect of Genius. To an engineering or a technology crowd, understandably Claude Shannon and Alan Turing will appeal more than Newton or Einstein. On that note, I highly recommend reading Walter Isaacson's The Innovators: How a Group of Hackers, Geniuses, and Geeks Created the Digital Revolution, or better still try reading one of these books: Jimmy Soni's A Mind at Play: How Claude Shannon Invented the Information Age, or Andrew Hodges' Alan Turing: The Enigma. But maybe analyzing cases where talent fails to achieve excellence will be more useful to society because by figuring out how so many factors need to come together for that to happen, we may be able to uproot the organic production of Genius and supplant it by the industrial production of Genius. Well, of course, I do not say this in all seriousness.(back)