# Is the following asymptotic relation correct?

Assume $\epsilon$ is a small positive parameter, $0 < \epsilon <<1$, $k$ is another parameter (the size of which we don't know), and $B_{1}$ and $B_{2}>1$ are constants. Suppose I have a quantity $\Delta t$ satisfying the inequalities $B_{2} \epsilon \leq \Delta t \leq B_{1}k$.

Now assume I have some other quantity, call it $Q$, such that $Q \sim O((\Delta t)^{2} + \epsilon \Delta t)$. Do the above inequalities imply that I can write

$$O((\Delta t)^{2} + \epsilon \Delta t) \sim O(k^{2})$$

?

• Should $Q$ have something to do with $k$? I don't see where $k^2$ came into the picture. – Ian Jan 18 '17 at 23:54
• Well, I can see that since $\Delta t \leq B_{1}k$, we can say $(\Delta t)^{2} \leq B_{1}^{2}k^{2}$, is that what you mean? @Ian – Alex Jan 18 '17 at 23:57
• Sure, $Q\sim O(k^2)$. You can prove this directly with algebraic manipulation. – Matt Samuel Jan 18 '17 at 23:59
• Thanks! I'd be grateful if you could provide a few more details though @MattSamuel – Alex Jan 19 '17 at 0:00

$$(\Delta t)^2\leq B_1^2k^2$$ $$\epsilon\leq\frac{B_1}{B_2}k$$ $$\epsilon\Delta t\leq \frac{ B_1^2}{B_2}k^2$$ Sum the first and the third inequalities to finish the proof.