User Tools

Site Tools


math104-s21:s:victoriatuck

Victoria's Page

Course Notes

Sequences

Limsup and Liminf

Metrics and Topology

Series

Continuity

Convergence

Mean Value Theorem

Taylor Expansions

Differentiation

Integration

Questions

1. Question: (2,2)Q(-\sqrt{2}, \sqrt{2}) \cap \mathbb{Q} = [2,2]Q[-\sqrt{2}, \sqrt{2}] \cap \mathbb{Q} is both closed and open on Q\mathbb{Q}. However, for a given space E, Rudin Theorem 2.27 states that if XX is a metric space and EXE \subset X then E=EˉE = \bar{E} if and only if E is closed. [Eˉ=EE\bar{E} = E \cup E', EE' denotes the set of all limit points of E in X]. How does this work with the idea that a sequence of rationals can converge to a number outside of the rationals?

Answer: Let X=QX = \mathbb{Q}. E=(2,2)QQE = (-\sqrt{2}, \sqrt{2}) \cap \mathbb{Q} \subset \mathbb{Q}. The set of all limit points E' must be in X, so it can only contain rationals. We can prove that the limit points of E must be inside the set E by contradiction. Assume there is a point xQ:xEx \in \mathbb{Q}: x \notin E. This may be a limit point of E because it is a rational. Given an ϵ>0\epsilon > 0, in order for x to be a limit point, ϵ\forall \epsilon q,qx\exists q, q \neq x st qEq \in E. Let ϵ=min(x2,x2)2\epsilon = \frac{min(x - \sqrt{2}, -x - \sqrt{2})}{2}. Given this value, we can not have a point qEq \in E where qx<ϵ|q - x| < \epsilon. Therefore, no rational number outside of E can be a limit point of E, and E contains all of its limit points.

The confusion with this question for me lay with the use of the limit point argument for situations such as E=(0,1)QE = (0, 1) \cap \mathbb{Q}, which is not closed. However, the difference here is that 1 and 0 are rational numbers.

2. Question: Is the set E=(0,2)QE = (0,2) \cap \mathbb{Q} connected on Q\mathbb{Q}? [With explanation]

Answer: No, this set is disconnected. QR\mathbb{Q} \subset \mathbb{R}. Therefore, in general, we can pick an element outside of our set E but inside the metric space R\mathbb{R} i.e., xRQx \in \mathbb{R} \setminus \mathbb{Q} to “split” our set E into two open subsets U1=(0,x)QU_1 = (0, x) \cap \mathbb{Q}, U2=(x,2)QU_2 = (x, 2) \cap \mathbb{Q} where (EU1)(EU2)=(E \cap U_1) \cap (E \cap U_2) = \varnothing and (EU1)(EU2)=E(E \cap U_1) \cup (E \cap U_2) = E where EU1E \cap U_1 \neq \varnothing and EU2E \cap U_2 \neq \varnothing. Thus, we can use the definition of a disconnected set to show that E is a disconnected set.

3. Question: Let f:XYf: X → Y be a continuous mapping with AXA \subset X and BYB \subset Y. I know that if B is open/closed then f1(B)f^{-1}(B) is open/closed, respectively, but why is this true?

Answer: For an open set B, this is by definition of a continuous mapping, i.e., a mapping f:XYf: X → Y is continuous iff BY\forall B \subset Y, f1(B)f^{-1}(B) is open in X.

4. Question: Exam 2, Problem 5. Let f:QRf: \mathbb{Q} → \mathbb{R} be a continuous map. Is it true that one can always find a continuous map g:RRg: \mathbb{R} → \mathbb{R} extending ff, namely, g(x)=f(x)g(x) = f(x) for any xQx \in \mathbb{Q}? Prove or find a counterexample.

Answer:

5. Assume f is a continuous map f:XYf: X → Y with AXA \subset X and BYB \subset Y

a) Question: What is a counterexample to “If AA is closed, then f(A)f(A) is closed”.

Answer: Let f:=arctan(x)f := arctan(x). Then (,)(π2,π2)(-\infty, \infty) → (-\frac{\pi}{2}, \frac{\pi}{2}).

b) Question: What is a counterexample to “If BB is connected, then f1(B)f^{-1}(B) is connected.”?

Answer: Let f:=x2f := x^2. Then [1,)(,1][1,)[1, \infty) → (-\infty, -1] \cup [1, \infty), which is disconnected.

6. Question: Given an open cover U\mathcal{U} of f(E)f(E) where f is a continuous function and E is a compact set, why is {f1(U):UU}\{f^{-1}(U) : U \in \mathcal{U}\} a cover of E? [Question from Ross Theorem 21.4]

Answer: https://math.stackexchange.com/questions/605195/detail-on-a-theorem-of-continuity-and-compactness

7. Question: For non 1-to-1 functions (like f=x2f = x^2), how do we define the inverse f1f^{-1} of a set in the range of that function? E.g., for x2x^2, what would f1([1,inf))f^{-1}([1, inf)) be?

Answer: In Ross 168, we define the inverse f1(U)f^{-1}(U) to include any value ss st f(s)Uf(s) \in U. So f1([1,inf))=(inf,1][1,inf)f^{-1}([1, inf)) = (-inf, -1] \cup [1, inf).

8. Question:

math104-s21/s/victoriatuck.txt · Last modified: 2022/01/11 10:57 by pzhou