# Are the solutions in radicals of cubic and quartic of any use in physics?

We all know that there are analytic formulae to solve quadratic, cubic and quartic polynomial equations. But it seems to me that the only solution that widely used is physics is the solution of quadratic we all learnt at school.

Is it true that when physicist faces quartic or even cubic equation -- he turns to numerical methods to solve them? Have anyone seen the analytic analysis of some real problems in physics, where these formulae were used?

• I prefer the expression 'solutions in radicals' than 'analytic formulae'. Although it is clear from the context it carries different connotations and associations.
– MBN
Commented Mar 10, 2011 at 15:44

It was a lot of fun to derive the solution for the cubic and quartic equation 22 years ago, when I was a high school kid, but I haven't really used it afterwards. I think the same is true for most physicists I met: they didn't use the complicated formulae, especially not the quartic one.

Much more generally, the situations in which it would be useful are extremely rare for several reasons:

• the situations in which the functions are truncated to a cubic or quartic polynomial are extremely rare

• known corrections ultimately modify the cubic or quartic polynomial to a more general function

• the position of roots is usually not the key interesting quantity

• the truly interesting and invariant quantities are functions of the roots that can be expressed more easily in terms of the coefficients

• quite typically, physical problems wouldn't lead to a cubic or quartic equation for one variable but to a set of such equations for many variables which can't be solved in radicals

• in the interesting cases when a cubic or quartic equation has to be solved, it reduces to a linear or quadratic equation, anyway.

To give some examples, one may consider the most general classical potential for a scalar field which produces a renormalizable theory. This is given by at most a 4th order polynomial of the scalar field. For a general potential, the zeros of such potential would be roots of the polynomial.

However, zeros of a potential are not too interesting: much more interesting are its stationary points, i.e. extrema, which are the solutions of the cubic equation. Moreover, the coefficients in the polynomial remain more important than the positions of the minima - and observable quantities may be expressed as their functions.

Also, the canonical shapes of the potential have a $Z_2$ or a related symmetry that effectively reduces the degree of the equation to half of it. Moreover, generally, the theories may include several fields which produces a set of higher-order equations that usually can't be solved in radicals. Equally importantly, quantum corrections add non-polynomial corrections to the potential, anyway.

So the set of physical situations in which the problem could be solved using the nontrivial formulae is limited for all the reasons above, and even when the problem may be solved, we usually don't learn anything "qualitative" about the system. In this sense, the existence of the solution for cubic and quartic solutions is a mathematical curiosity that doesn't affect a physicist's life.

• Lubos did you get anything for doing that in school except for the fun of course? I am asking because Mark Kac did it too when he was a schoolboy, for some competition I guess, and that brought his family five zloti. I don't know how valuable that was at that time.
– MBN
Commented Mar 11, 2011 at 0:08
• Dear @MBN, I didn't do those things for any school. It has always been my completely individual research in full solitude. At elementary school, I would be rewarded for winning the Czechoslovak math olympiads etc. - things like a backpack for \$30 haha. In fact, I didn't realize at that time that the trip to Moscow to the international math olympiad was much more valuable, and if I remember, it was paid for haha. Commented Mar 12, 2011 at 17:53

The problem is that the actual analytic solutions to them are extremely ugly, and it's difficult to see how variation in one of the parameters in your cubic or quartic will affect the location of the roots. It's much easier to solve the problem numerically, and then plot the location of the roots as a function of the parameters you're varying than it is to carefully analyze the actual analytic solution.

If you google around you can find applications. For example, applications of the quartic equation:

Nonlinear Landau-Zener Tunnelling in Coupled Waveguide Arrays http://arxiv.org/abs/1008.1358

But they are rare.

If you wanna find the time it takes to travel some distance, and your acceleration is increasing linearly in time (constant jerk), then you'd have to solve a cubic.

The jerk-equation is also used in electric circuits, jerk circuits.

FWIIW, coincidentally last month I ran into an algebraic cubic equation when finding the spherically symmetric spacetime equivalent to the flat matter-only FLRW model (the Einstein-de Sitter model). Specifically when trying to express comoving time in terms of non-comoving time and radial coordinate. The cubic equation was of the kind with three real roots that must be found by way of intermediate complex numbers, so that none was reducible to a real algebraic expression.

I did learn something "qualitative" from that fact: that expansion of space is the only possible conceptual framework to work with, except for the Milne and de Sitter cases. I had become interested in the topic after reading an exchange of papers mainly between Abramowicz and Chodorowski a few years ago, which had left the issue unsettled.