Apply the epsilon delta definition of limit to the multivariable function

75 Views Asked by At

The rigorous definition of limit for the function of two variable is given as following:

Give a function of two variables $f : D → \mathbb{R}, D ⊆ \mathbb{R}^2$ such that $D$ contains points arbitrarily close to a point $(a, b)$, we say that the limit of $f (x, y)$ as $(x, y)$ approaches $(a, b)$ exists and has value $L$, if and only if for every real number $ε > 0$ there exists a real number $δ > 0$ such that

$$|f (x, y) − L| < ε$$

whenever $$ 0 <\sqrt{(x − a)^2 + (y − b)^2}< δ.$$ we then write $$ \lim_{(x,y)\to(a,b)} f(x,y)=L.$$

I want to apply this definition to the function $ f(x,y)=e^{-x^2y},$ but unlike the single variable cases, I could not find a way to substitute the $δ$ in to the inequality, and then choose a minimum value to show that the $ε$ is bounded. Could some one help me on this problem? Thanks.

1

There are 1 best solutions below

2
On BEST ANSWER

Warning: Using only first principles, this is extremely tedious.

Let $(x_0,y_0)$ be given and fixed. First choose $\epsilon>0$ sufficiently small ($\epsilon<e^{-x_0^2y_0}$) so that \begin{align} 0<e^{-x_0^2y_0}-\epsilon<e^{-x^2y}<e^{-x_0^2y_0}+\epsilon \\ \implies \ln(e^{-x_0^2y_0}-\epsilon)<-x^2y<\ln(e^{-x_0^2y_0}+\epsilon) \\ \implies \color{red}{\ln(e^{-x_0^2y_0}-\epsilon)+x_0^2 y_0}<\color{blue}{-x^2y+x_0^2 y_0}<\color{green}{\ln(e^{-x_0^2y_0}+\epsilon)+x_0^2 y_0}. \end{align} Note that the red is negative and the green is positive. By adding and subtracting $x^2y_0$ respectively, the middle term can be written as \begin{align} (x_0^2-x^2)y_0+x^2(y_0-y)=(x_0-x)(x_0+x)y_0+x^2(y_0-y). \label{*}\tag{*} \end{align} First start with $\delta_1=1$. Then $-1<x-x_0<1\implies -1+2x_0<x+x_0<1+2x_0$ so $|x+x_0|\leq C_1(x_0):=\max\{|-1+2x_0|,|1+2x_0|\}$.

Similarly, $-1<x-x_0<1\implies x_0-1<x<x_0+1\implies |x|<C_2(x_0):=\max\{|x_0-1|,|x_0+1|\}.$ Therefore, \eqref{*} is bounded above by $$ |x_0-x||x_0+x||y_0|+|x|^2|y_0-y|\leq C_1|y_0|\delta_2+C_2^2\delta_2=(C_1|y_0|+C_2^2)\delta_2 $$ with $\delta_2$ to be chosen. Choose $\delta_2=\frac{\ln(e^{-x_0^2y_0}+\epsilon)+x_0^2 y_0}{C_1|y_0|+C_2^2}$. This part takes care of showing the blue is less than the green above.

Similar techniques show \eqref{*} is bounded below by $$ -|x_0-x||x_0+x||y_0|-|x|^2|y-y_0|\geq -(C_1|y_0|+C_2^2)\delta_3. $$ Choose $\delta_3=-\frac{\ln(e^{-x_0^2y_0}-\epsilon)+x_0^2 y_0}{C_1|y_0|+C_2^2}$. This will take care of showing that the blue is greater than the red.

In conclusion, let $\delta=\min\{\delta_1,\delta_2,\delta_3\}$. Then when $\sqrt{(x-x_0^2)+(y-y_0)^2}<\delta$, we will have red is less than blue is less than green. Working backwards shows $|e^{-x^2y}-e^{-x_0^2 y_0}|<\epsilon$.

If $\epsilon$ is not small, find $\epsilon_1<\epsilon$ that is and show everything is less than $\epsilon_1$ (using above). Then everything will be less than $\epsilon$.