I am wondering if there is a uniform bound $C$ (independent of $\lambda>10$): $$\sum_{k=-\infty}^{-1}\Big|\int_{2^k}^{2^{1+k}}\frac{\sin(\lambda t^3)}{t}dt\Big|\le C.$$

Remark: (1) An easy upper bound is $C\log\lambda$, since $$\Big|\int_{2^k}^{2^{1+k}}\frac{\sin(\lambda t^3)}{t}dt\Big|\lesssim \min\{1,\lambda 2^{3k}\},$$ by $|\sin(\lambda t^3)|\le \min\{1,\lambda t^3\}.$

(2) I can verify this uniform bound by Matlab, but I am not able to find a proof.


You are not exploiting that the integrands are oscillatory for $\lambda t^3\gtrsim 1$, and the usual way to do this is by integration by parts. Let me change notations slightly and use $n=-k$. $$ \int_{2^{-n-1}}^{2^{-n}} e^{i\lambda t^3} \, \frac{dt}{t} =\int_{2^{-3n-3}}^{2^{-3n}} e^{i\lambda s}\, \frac{ds}{3s} = O(2^{3n}/\lambda) - i \int_{2^{-3n-3}}^{2^{-3n}} e^{i\lambda s}\frac{ds}{3\lambda s^2} = O(2^{3n}/\lambda) , $$ so summing over those $n$ for which $\lambda 2^{-3n}\ge 1$ makes an overall contribution $\lesssim 1$.

This is also true for the $n$'s with $\lambda 2^{-3n}\le 1$, by the argument you already gave, so your claim follows.


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.