I'm wondering if there exists a closed-form or analytic expression for the roots of an equation of the form

$ax^2 + bx + c\log x=0.$

considering the natural $\log$. Wolfram alpha is leading me to expressions involving the Lambert W (product log) function when I include either the quadratic term or the linear term (but not both) and analytic approximations when I supply real values for the coefficients.

This is OK, but does a more general solution exist in terms of the coefficients?



Any equation which can write $$A+B x+C \log(D+Ex)=0$$ has a solution which can be expressed in terms of Lambert function. $$x=\frac{C }{B}W(F)-\frac{D}{E} $$ where $$F=\frac{B }{C E} e^{\frac{B D-A E}{C E}}$$This is why Wolfram Alpha gave you solutions for $$ax^2 + bx + c\log x=0$$ in the case where $a=0$ or $b=0$.

However, for the general case you are considering $(a \neq 0,b\neq 0)$, I do not think that exists any analytical solution and only numerical methods (such as Newton) would solve the problem and this should be quite easy.

  • $\begingroup$ I suspected as much, thanks for the input! $\endgroup$ – zzk Dec 31 '14 at 19:51
  • $\begingroup$ You are very welcome ! **Happy New Year ** !!! $\endgroup$ – Claude Leibovici Jan 1 '15 at 6:51

The equation $ax^2 + bx + c\log x=0$ can be rewritten as (supposing a not null): $$\frac{a}{c}(x+\frac{b}{2a })^2-\frac{b^2}{4a c}+ \log(x)=0$$ Taking the exponential of both members: $$e^{-\frac{a}{c}(x+\frac{b}{2a })^2+\frac{b^2}{4a c}}=x$$ Remembering Lagrange series: $$x=\sum_{n=1} \frac{D^{n-1} \left. e^{-n\frac{a}{c}(x+\frac{b}{2a })^2 + \frac{nb^2}{4a c}}\right|_{x=0}}{n!}$$

$$x=\sum_{n=1} e^{n\frac{b^2}{4a c}} \frac{D^{n-1} \left. e^{-n\frac{a}{c}(x+\frac{b}{2a })^2}\right|_{x=0}}{n!}$$ Supposing $a/c > 0$ and remembering Rodrigues formula of Hermite polynomials: $$H_n(x)=(-1)^n e^{x^2}\frac{d^n}{dx^n}e^{-x^2}$$ we have: $$x=\sum_{n=1}^{\infty} \frac{ \left( -\sqrt\frac{na}{c} \right)^{n-1} H_{n-1}\left( \sqrt\frac{na}{c} \frac{b}{2a}\right) }{n!}$$


"Generalization of Lambert $W$ function, Bessel polynomials and transcendental equations", Giorgio Mugnaini, http://arxiv.org/pdf/1501.00138


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.