× Linear Transformation Multiplicity Dual Space Minimal Polynomial \(T\)-Complement Cyclic Operator Indecomposable Operator Characteristic Polynomial Diagonalizable Smith Normal Form Normal Subgroup Isomorphism Center Centralizer Normalizer Stabilizer Orbit \(p\)-group Sylow \(p\)-Subgroup Sylow's Theorem Simple Group Solvable Group Field Integral Domain Ascending Chain Condition Principal Ideal Domain Unique Factorization Domain Polynomial Ring Division Algorithm in SageMath Prime and Irreducible Adjoining Element Splitting Field Minimal Polynomial Field Extension Separable Algebraic Closure Galois Group Galois Theory KU 2015 (January) KU 2015 (August) KU 2016 (January) KU 2016 (August) KU 2017 (January) KU 2017 (August) KU 2018 (January) KU 2018 (August) KU 2019 (January) KU 2019 (August) KU 2020 (January) KU 2020 (August) KU 2021 (January) KU 2021 (August) KU 2022 (January) KU 2022 (August) KU 2023 (January) KU 2023 (August) KU 2024 (January)
☰ Menu

KU 2019 (August) ALGEBRA QUALIFYING EXAM

My name is Hanzhang Yin, and I have developed this website as a resource to facilitate the review of key concepts in abstract algebra, and it is not fully completed. Feel free to email me at hanyin@ku.edu if there are any errors or suggestions for improvement.

Problem 2. Let \(V\) be a finite dimensional vector space over the complex numbers and \(T\) a linear operator on \(V\) such that the subspaces of \(V\) generated by all the eigenvectors is one dimensional. Prove that the minimal polynomial of \(T\) is \((x − \lambda)^n\) where \(n\) is the dimension of \(V\) .

Proof. Given that the subspaces of \(V\) generated by all the eigenvectors is one dimensional, we can know that the basis of the subspace is \(\{v\}\), where \(v\) is the eigenvector of \(T\). Then we have \(T(v) = \lambda v\), where \(\lambda\) is the correspondent eigenvalue of \(v\). Hence, if \(v'\) is an eigenvector of \(T\). Then, we can know that \(v' = cv\) for some \(c\in\mathbb{C}\). Hence, we have \[ T\cdot v' = T\cdot cv = cT\cdot v = c\lambda v = \lambda cv = \lambda v'. \] Thus, we show that there is only one eigenvalue of \(T\), which is \(\lambda\). Then, we can know that the characteristic polynomial of \(T\) is \((x - \lambda)^n\), where \(n\) is the dimension of \(V\). In that case, the minimal polynomial of \(T\) is \((x - \lambda)^m\), where \(m\leq n\) (to be continued...)

5. Consider the ring \(R = \mathbb{Q}[x, y, z]/I\), where \(I = (x^2y - z^5)\).

(i) Prove that \(R\) is not a field. Then decide whether \(R\) is an integral domain, and justify your answer.

(ii) Given \(f\in \mathbb{Q}[x, y, z]\), let \(\overline{f}\) denote the element \(f + I\) of \(R\). Is the ideal \(J = (\bar x, \bar y)\) a prime ideal of \(R\)? Prove your answer.

(iii) Show that \(\overline{z^5} \in J^2\), but that \(\overline{z^4}\notin J^2\).

Proof (i). We know that \(\mathbb{Q}[x, y, z]/I\) is a field if and only if \(I\) is a maximal ideal. We know that \((x^2y - z^5)\subset (x^2y - z^5, z)\) since \(x^2y - z^5 = 1\cdot (x^2y - z^5) + 0\cdot z\). Now, I want to show that \((x^2y - z^5, z)\) is a proper ideal of \(\mathbb{Q}[x, y, z]\). Firstly, for any \(f\neq 0\in \mathbb{Q}[x, y, z]\), we can know that \(f\cdot (x^2y - z^5)\) does not contain a term of constant. Similarly, we can know that \(g\cdot z\) is a polynomial without constant term for any \(g\neq 0\in \mathbb{Q}[x, y, z]\). Hence, we can know that if \(f\cdot (x^2y - z^5) + g\cdot z\neq 0\), then \(f\cdot (x^2y - z^5) + g\cdot z\notin \mathbb{Q}\). In other words, if \(i\in (x^2y - z^5, z)\), then \(i\) is not a unit. Thus, we have \((x^2y - z^5, z)\neq \mathbb{Q}[x, y, z]\), which implies that \((x^2y - z^5, z)\) is a proper ideal. Then, we can conclude that \((x^2y - z^5)\) is not a maximal ideal. Therefore, we can know that \(R\) is not a field. \(\blacksquare\)

Now, we want to show that \(R\) is an integral domain.

Proof (i). We know that \(\mathbb{Q}[x, y, z]/I\) is an integral domain if \(I\) is a prime ideal. Since \(\mathbb{Q}[x, y, z] = \mathbb{Q}[x, z][y]\), we can know that \(x^2y - z^5\) is a degree \(1\) polynomial of \(\mathbb{Q}[x, z][y]\). Hence, \(x^2y - z^5\) is irreducible, which implies that \(x^2y - z^5\) is prime. Thus, we can know that \(I = (x^2y - z^5)\) is a prime ideal. Therefore, we have \(R\) is an integral domain. \(\blacksquare\)

Proof (ii). We want to show that \(J = (\bar x, \bar y)\) is a prime ideal of \(R\). If \(\bar r\in (\bar x, \bar y)\), we can know that \(\bar r = \bar x\bar f + \bar y\bar g\) for some \(\bar f, \bar g\in R\). Then, we we can know that \(\bar r\) does not contain a term of \(a_nz^n\) for any \(n\in \mathbb{N}\) and \(a_n\in\mathbb{Q}\) (including the case that \(n = 0\). In the other direction, if \(\bar r\) is an element of \(R\) without a term of \(a_nz^n\) for any \(n\in \mathbb{N}\) and \(a_n\in\mathbb{Q}\), then we can know that \(\bar r = \bar x\bar f + \bar y\bar g\) for some \(\bar f, \bar g\in R\). Hence, if \(\bar f, \bar g\in R\) do not contain a term of \(a_nz^n\) for any \(n\in \mathbb{N}\) and \(a_n\in\mathbb{Q}\) (i.e. \(\bar f\notin J, \bar g\notin J\)), then we can know that \(\bar f\cdot \bar g\) does not contain a term of \(a_nz^n\) for any \(n\in \mathbb{N}\) and \(a_n\in\mathbb{Q}\) (i.e. \bar f\bar g\notin J). Therefore, we can know that \(J\) is a prime ideal of \(R\). \(\blacksquare\)

Proof (iii). Firstly, we want to show that \(J^2 = ({\bar x}^2, \bar x\bar y, {\bar y}^2)\). Let \(\bar p, \bar q\in J\). Then, we can know that \(\bar p = \bar x\bar f + \bar y\bar g\) and \(\bar q = \bar x\bar h + \bar y\bar k\) for some \(\bar f, \bar g, \bar h, \bar k\in R\). Then, we can know that \(\bar p\bar q = \bar x\bar x\bar f\bar h + \bar x\bar y\bar f\bar k + \bar x\bar y\bar g\bar h + \bar y\bar y\bar g\bar k\). Hence, we can know that \(\bar p\bar q\in ({\bar x}^2, \bar x\bar y, {\bar y}^2)\), which implies that \(J^2\subseteq ({\bar x}^2, \bar x\bar y, {\bar y}^2)\). In the other direction, if \(\bar r\in ({\bar x}^2, \bar x\bar y, {\bar y}^2)\), then we can know that \(\bar r = \bar x\bar x\bar f + \bar x\bar y\bar g + \bar y\bar y\bar h\) for some \(\bar f, \bar g, \bar h\in R\).

4. Recall that if \( G \) is a group and \( x, y \in G \), \( x \) and \( y \) are conjugate if \( y = gxg^{-1} \) for some \( g \in G \), and that conjugacy is an equivalence relation on \( G \). Let \( [x] \) denote the conjugacy class of an element \( x \in G \); i.e., \( [x] = \{gxg^{-1} \mid g \in G\} \).

(i) Prove that for \( x \in G \), \( [x] \) has exactly one element if and only if \( x \in Z(G) \), where \( Z(G) \) denotes the center of \( G \).

(ii) More generally, prove that the number of elements in \( [x] \) equals \( |G : C_G(x)| \), where \( C_G(x) \) is the centralizer of \( x \) in \( G \), i.e., \( C_G(x) = \{g \in G \mid gxg^{-1} = x\} \). Hint: When is \( gxg^{-1} \) equal to \( hxh^{-1} \)?

(iii) Suppose that \( G \) is a finite group of odd order, and suppose that \( N \) is a normal subgroup of \( G \) of order 3. Prove that \( N \leq Z(G) \). Hint: What does the normality of \( N \) say about the conjugacy classes of elements \( x \in N \)? How can one use this to partition \( N \)?

Proof (i). Firstly, we can know that \(x\in [x]\) since \(exe^{-1} = exe = x\). Thus, if \([x]\) has exactly one element, then we can know that \(gxg^{-1} = x\) for all \(g\in G\). Hence, we have \(x = gxg^{-1}\) for all \(g\in G\), which implies that \(x\in Z(G)\). If \(x\in Z(G)\), then we have that \(gxg^{-1} = x\) for all \(g\in G\), which implies that \(\{x\} = [x]\). Thus, we can know that \([x]\) has exactly one element. \(\blacksquare\)

Proof (ii). Since the number of elements in \([x]\) is the number of elements in the orbits. We define the map \(\varphi: [x]\to G/C_G(x))\) by \(\varphi(gxg^{-1}) = gC_G(x)\). Firstly, we can know that the map is surjective since for any \(gC_G(x)\), we have \(gxg^{-1}\in [x]\) such that \[ \varphi(gxg^{-1}) = gC_G(x). \] Now, suppose that \(gC_G(x) = hC_G(x)\). Then, we have \(h^{-1}g\in C_G(x)\), which implies that \(h^{-1}gxg^{-1}h = x\). Hence, we have \[ \begin{align} h^{-1}gxg^{-1}h &= x\\ h(h^{-1}gxg^{-1}h)h^{-1} &= hxh^{-1}\\ gxg^{-1} &= hxh^{-1}. \end{align} \] Thus, we can know that the map is injective. Hence, we can know that the number of elements in \([x]\) equals \(|G/C_G(x)| = |G:C_G(x)|\). \(\blacksquare\)

Proof (iii).We define a map \(\varphi: G\to \text{Aut}(N)\) by \(\varphi(g) = gng^{-1}\). The reason that \(\varphi(g)\) is a group automorphism of \(N\) is \(N\) is normal in \(G\). We can see that \(\varphi\) is well-defined since for any \(g = h\in G\), we have \(gng^{-1} = hnh^{-1}\). Then, we can know that \(\varphi\) is a group homomorphism since for any \(g, h\in G\), we have \[ \varphi(g)\varphi(h) = \varphi(g)\circ \varphi(h) = g h n h^{-1}g^{-1} = (gh)n(gh)^{-1} = \varphi(gh). \] Hence, we can see that the image of \(\varphi\), \(\varphi(G)\), is a subgroup of \(\text{Aut}(N)\). Given that \(N\) has order of \(3\), we have \(N\cong \mathbb{Z}/3\mathbb{Z}\), which implies that \(N\) is cyclic. Hence, any non-trivial element is a generator of \(N\). Thus, we have \(N = \{e, x, x^2\}\), where \(x, x^2\) are the generators of \(N\). Once we know where \(x\) is mapped to, we can know where \(x^2\) is mapped to since \(\varphi\) is a group homomorphism. There are only two options, either \(x\to x\) or \(x\to x^2\). Hence, we can see that \(|\text{Aut}(N)| = 2\), which implies that \(\mathbb{Z}/2\mathbb{Z}\cong \text{Aut}(N)\). Since \(\varphi(G)\) is a subgroup of \(\text{Aut}(N)\), we have \(|\varphi(G)| = 1\) or \(2\). According to the \(\textbf{First Isomorphism Theorem}\), we have \(G/\ker(\varphi) \cong \varphi(G)\)\). Since \(|\varphi(G)| = |G/\ker(\varphi)| = [G:\ker(\varphi)]\), we have \([G:\ker(\varphi)] = 1\) or \(2\). Given that \([G:\ker(\varphi)]\mid |G|\) and \(|G|\) is odd, we have \([G:\ker(\varphi)] = 1\). Thus, we can see that \(\varphi(G) = \{e\}\), which implies that \(\ker(\varphi) = G\). And \(\ker(\varphi) = \{g\in G\mid \varphi(g) = e\}\), which implies that \(gng^{-1} = n\) for all \(n\in N\). Therefore, we can get that \(N\leq Z(G)\). \(\blacksquare\)

6. Consider the polynomial \( p(x) = x^3 - 2 \in \mathbb{Q}[x] \). Find \( \alpha, \beta \in \mathbb{C} \) for which \( F = \mathbb{Q}(\alpha, \beta) \) is a splitting field for \( p(x) \), and prove that this is the case. Then find, with justification, the degree of the field extension \( \mathbb{Q} \subseteq F \).

Solution (6). Firstly, we want to find the all the roots of \(p(x)\). Suppose that \(a\) is the root of \(p(x)\). Then we have \(p(a) = a^3 - 2 = 0\) and \(a^3 = 2\). Hence, we can know that \(a^3\) is a cube of \(2\). Thus, we can know that \(a = \sqrt[3]{2}\), which is a real number. Now, if \(a^3 = 2\cos(2\pi) + 2\sin(2\pi)i = 2(\cos(2\pi) + sin(2\pi)i) = 2e^{2\pi i}\), then we have \(a = \sqrt[3]{2}e^{2\pi i/3}\). Hence, \(a = \sqrt[3]{2}(\cos(2\pi/3) + \sin(2\pi/3)i) = \sqrt[3]{2}(-1/2 + \sqrt{3}/2i)\). Similarly, we can have \(a^3 = 2\cos(-2\pi) + 2\sin(-2\pi)i = 2e^{-2\pi i}\), which implies that \(a = \sqrt[3]{2}e^{-2\pi i/3}\). Hence, we have \(a = \sqrt[3]{2}(\cos(-2\pi/3) + \sin(-2\pi/3)i) = \sqrt[3]{2}(-1/2 - (\sqrt{3}/2)i)\). Then, we have the three distinct roots of \(p(x)\), which are \[ a_1 = \sqrt[3]{2},\quad a_2 = \sqrt[3]{2}(-1/2 + (\sqrt{3}/2)i),\quad a_3 = \sqrt[3]{2}(-1/2 - (\sqrt{3}/2)i). \] Thus, we can know that the splitting field of \(p(x)\) is \(F = \mathbb{Q}(a_1, a_2, a_3)\). Since \(a_2 - a_3 = \sqrt[3]{2}\cdot\sqrt{3}i\), we have \[ \dfrac{\sqrt[3]{2}\cdot\sqrt{3}i}{\sqrt[3]{2}} = \sqrt{3}i. \] Hence, \(\mathbb{Q}(\sqrt[3]{2}, \sqrt{3}i)\subset \mathbb{Q}(a_1, a_2, a_3)\). It is not hard to see that \(\mathbb{Q}(a_1, a_2, a_3)\subset \mathbb{Q}(\sqrt[3]{2}, \sqrt{3}i)\). Since \(\sqrt{3}i\) is a root of \(x^2 + 3 = 0\). Since there is no rational roots of \(x^2 + 3 = 0\), we can know that \(x^2 + 3\) is irreducible over \(\mathbb{Q}\). Thus, we can know that \[ \mathbb{Q}(\sqrt{3}i\) \cong \mathbb{Q}[x]/(x^2 + 3). \] Hence, \(\mathbb{Q}(\sqrt{3}i\)\) has degree \(2\) since the minimal polynomial of \(\sqrt{3}i\) has degree \(2\). Then, we want to show that \(x^3 - 2\) is irreducible over \(\mathbb{Q}(\sqrt{3}i)\) by contradiction. If \(x^3 - 2\) is a reducible polynomial over \(\mathbb{Q}(\sqrt{3}i)\), then we have a root of \(x^3 - 2\) in \(\mathbb{Q}(\sqrt{3}i)\). Since \(\mathbb{Q}(\sqrt{3}i\) = \{a + b\sqrt{3}i\mid a, b\in \mathbb{Q}\), we can know that the root of \(x^3 - 2\) is \(a + b\sqrt{3}i\) for some \(a, b\in \mathbb{Q}\). Then suppose that \(a + b\sqrt{3}i) is a root of \(p(x)\), \[ \begin{align} (a + b\sqrt{3}i\)^3 &= 2, \\ (a^2 + 2ab\sqrt{3}i - 3b^2)(a + b\sqrt{3}i) = 2, \\ (a + b\sqrt{3}i\)^3&= a^3 - 9 a b^2 + 3 i \sqrt{3} b (a^2 - b^2) = 2 \end{align} \]\Hence, we can know that \(3b(a^2 - b^2)\) =0\), which implies that \(b = 0\) or \(a-b = 0\) or \(a + b = 0\). Given that \(b\neq 0\), or \(a + b\sqrt{3}i = a\) is a root of \(p(x)\), which is a contradiction. Suppose that \(a = b\), then we can have \(a^3 - 9a^3 = -8a^3 = 2\), which implies that \(a^3 = -1/4\), which is a contradiction. Suppose that \(a = -b\), then we can have \(a^3 - 9a^3 = -10a^3 = 2\), which implies that \(a^3 = -1/4\), which is a contradiction. Hence, there exists no roots in \(p(x)\) and \(p(x)\) is irreducible. Then \(\mathbb{Q}(\sqrt{3}i, \sqrt[3]{2})\cong \mathbb{Q}(\sqrt{3}i)[x]/(x^3 - 2)\). Therefore, we have \[ [\mathbb{Q}(\sqrt{3}i, \sqrt[3]{2}): \mathbb{Q}(\sqrt{3}i)][\mathbb{Q}(\sqrt{3}i): \mathbb{Q}] = 3\cdot 2 = 6. \]