Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Consider the following equation $$\frac{dx}{dt}=-\lambda x +\epsilon x(t-a), \quad x(0)=1,\quad |\epsilon|\ll1$$ where $a$ and $\lambda$ are positive constants and $x(t-a)$ means the function $x(t)$ evaluated at $t-a$. The task is to obtain a solution as a series in powers of $\epsilon$. The book suggests to show that $x(t)=e^{-\lambda t} w(t)$ satisfies $$\frac{dw}{dt}=\delta w(t-a)\quad w(0)=1,\quad \delta=\epsilon e^{\lambda a}$$ and show that the perturbation expansion is given by the following expression: $$w(t)=1+t\sum_{n=1}^{\infty}\frac{\delta^n}{n!}\left(t-na\right)^{n-1}$$ hence $$x(t)=e^{-\lambda t}\left(1+t\sum_{n=1}^{\infty}\frac{\delta^n}{n!}\left(t-na\right)^{n-1}\right)$$ Since the method given in the solution is basically verification by substitution and proof by induction, I have been looking for a more systematic approach. The presence of time delay in RHS suggests that Laplace transform method could be appropriate here. So I proceed as follows. Apply Laplace transform to both sides of the original equation observing the initial conditions: $$x(t)\fallingdotseq X(s)$$ $$sX(s)-x(0)=-\lambda X(s) +\epsilon e^{-sa}X(s)$$ Solve for $X(s)$: $$X(s)=\frac{x(0)}{s+\lambda-\epsilon e^{-sa}}=\frac{1}{s+\lambda-\epsilon e^{-sa}}$$ Expand in powers of $\epsilon$: $$X(s)=\frac{1}{s+\lambda}\left(\frac{1}{1-\epsilon\frac{e^{-sa}}{s+\lambda}}\right)=\frac{1}{s+\lambda}\left(1+\sum_{n=1}^{\infty}\frac{\epsilon^n e^{-nas}}{(s+\lambda)^n}\right)=\frac{1}{s+\lambda}+\sum_{n=1}^{\infty}\frac{\epsilon^n e^{-nas}}{(s+\lambda)^{n+1}}$$ Taking inverse transforms $$\frac{1}{s+\lambda} \risingdotseq e^{-\lambda t}$$ $$\frac{e^{-nas}}{(s+\lambda)^{n+1}} \risingdotseq \frac{(t-na)^n e^{-\lambda(t-na)}}{n!}$$ But then I lose the factor of $t$ before the sum as compared to the book solution. Where am I making the slip? Thanks in advance.

share|cite|improve this question
What is the \fallingdotseq $\fallingdotseq$ for? I've never seen this notation. – Patrick Da Silva May 29 '12 at 22:44
@PatrickDaSilva For noting it is an inverse transform equality, if I'm not reading wrongly. – Pedro Tamaroff May 29 '12 at 22:44
@Peter Tamaroff : I understand that on one side you have a function and on the other its Laplace Transform, but I wanted to know the precise meaning of the symbol. – Patrick Da Silva May 29 '12 at 22:46
I was just telling anon, sometimes people choose a simbol to denote something. Such as $e^{ta} \sqsubseteq \frac{1}{s-a}$ – Pedro Tamaroff May 29 '12 at 22:47
I realised that this is a possibly outdated symbol specific to old Russian textbooks for example this one. Could not find any I am aware of available in preview – Valentin May 29 '12 at 23:02
up vote 2 down vote accepted

This might be a slightly different approach, but here is how I would tackle this.

Let $D=\frac{\mathrm{d}}{\mathrm{d}t}$, then we can rewrite your equation as $$ (D+\lambda-\epsilon e^{-aD})x=0\tag{1} $$ To invert the operator, consider the expansion: $$ \begin{align} \frac{1}{\lambda+x-\epsilon e^{-ax}} &=\frac{1}{\lambda+x}\frac{1}{1-\epsilon\frac{e^{-ax}}{\lambda+x}}\\ &=\frac{1}{\lambda+x}+\epsilon\frac{e^{-ax}}{(\lambda+x)^2}+\epsilon^2\frac{e^{-2ax}}{(\lambda+x)^3}+\epsilon^3\frac{e^{-3ax}}{(\lambda+x)^4}+\dots\tag{2} \end{align} $$ Using integrating factors we get that $$ (\lambda+D)^{-1}f=e^{-\lambda t}\int e^{\lambda t}f\,\mathrm{d}t\tag{3} $$ which leads us to $$ (\lambda+D)^{-k-1}0=e^{-\lambda t}\sum_{j=0}^kc_j\frac{t^{k-j}}{(k-j)!}\tag{4} $$ and therefore the coefficient of $\epsilon^k$ would be $$ e^{-kaD}(\lambda+D)^{-k-1}0=e^{-\lambda (t-ka)}\sum_{j=0}^kc_j\frac{(t-ka)^{k-j}}{(k-j)!}\tag{5} $$

share|cite|improve this answer
Thank you, I will give it a close look. Surprisingly, your approach links to my other question here – Valentin May 29 '12 at 23:33

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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