Let $X_t$ be a stochastic process on the state space {Working, Broken}. Let $U$ be the cumulative sojourn Working during an interval $[0,\tau]$ (the process's uptime). It is well-known [1] that if the process has a finite mean time to failure (MTTF) and mean time to repair (MTTR), then $$ \begin{align*} P\left\{ X_t = \text{Working}\ | \ X_s = \text{Working} \right\} &= p + (1-p) \cdot z(t;s) \\ P\left\{ X_t = \text{Broken}\ | \ X_s = \text{Broken} \right\} &= 1-p + p \cdot z(t;s) \end{align*} $$ for some function $z$, where $p = \frac{\text{MTTF}}{\text{MTTF}+\text{MTTR}}$.

Now further suppose that the times to failure have finite variance, and so too for the times to repair. Then in a classic result [2], Takács showed $U$ is asymptotically normal as $ \tau \rightarrow \infty $, and calculated the distribution's mean $\mu_U$ and variance $\sigma^2_U$.

My situation

I can prove that $U$ is asymptotically normal and calculate $\mu_U$ and $\sigma^2_U$, but with two caveats:

  1. I need to make an assumption about $z$. Thus my result is less general than Takács's.

  2. Calculating $\sigma^2_U$ requires calculating $\lim_{\tau\rightarrow\infty} \int_{0}^{\tau} z(t; 0) \ dt$. Thus my result is less useful than Takács's, as solving $z$ for arbitrary processes is an open question.

However, the result is (perhaps?) of interest as an elementary application of Billingsley's central limit theorem for dependent variables under strong mixing [3, Theorem 27.4]. (And yes: What happened is that I learned about Billingsley's result through Wikipedia when first reviewing literature, found a proof, and then I found Takács result. There must be Murphy's Law variant for this.)

If Prof Billingsley was still alive, I'd probably write to him and offer the result as a homework exercise for a new edition of his book. However, failing that option ...

My question

... what journal might be good for my result?

(I know I can self-publish to arXiv or equivalent with my employer. I am interested in alternatives.)

Many thanks.


[1] Kishor Shridharbhai Trivedi, 2002, Probability and statistics with reliability, queuing, and computer science applications, Second ed., John Wiley & Sons, New York.

[2] Lajos Takács, 1959, On a sojourn time problem in the theory of stochastic processes, Transactions of the American Mathematical Society, 93(3), 531--540.

[3] Patrick Billingsley, 2008, Probability and measure, John Wiley & Sons, New York.


Real Analysis Exchange has an Inroads section that specifically covers the case of "a clever new proof of an important theorem".


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.