美文网首页大学生数学竞赛
高代——数学竞赛初赛答案

高代——数学竞赛初赛答案

作者: 抄书侠 | 来源:发表于2019-02-18 01:22 被阅读0次

    初赛试题

    第一届

    二、【证明】(1)的证明:记
    A=\left(\alpha_{1}, \alpha_{2}, \cdots, \alpha_{n}\right), \quad M=a_{n 1} F^{n-1}+a_{n-11} F^{n-2}+\cdots+a_{21} F+a_{11} E
    要证明M=A,只需证明AM的各个列向量对应相等即可.若以e_i记第i个基本单位列向量.于是,只需证明:对每个iM e_{i}=A e_{i}\left(=\alpha_{i}\right)
    \beta=\left(-a_{n},-a_{n-1}, \cdots,-a_{1}\right)^{T}F=\left(e_{2}, e_{3}, \cdots, e_{n}, \beta\right).注意到F e_{1}=e_{2}, F^{2} e_{1}=F e_{2}=e_{3}, \cdots, F^{n-1} e_{1}=F\left(F^{n-2} e_{1}\right)=F e_{n-1}=e_{n}\left(^{*}\right)

    \begin{aligned} M e_{1} &=\left(a_{n 1} F^{n-1}+a_{n-11} F^{n-2}+\cdots+a_{21} F+a_{11} E\right) e_{1} \\ &=a_{n 1} F^{n-1} e_{1}+a_{n-11} F^{n-2} e_{1}+\cdots+a_{21} F e_{1}+a_{11} E e_{1} \\ &=a_{n 1} e_{n}+a_{n-11} e_{n-1}+\cdots+a_{21} e_{2}+a_{11} e_{1} \\ &=\alpha_{1}=A e_{1} \end{aligned}
    M e_{2}=M F^{} e_{1}=F^{} M e_{1}=F^{} A e_{1}=A F^{} e_{1}=A e_{2}
    M e_{3}=M F^{2} e_{1}=F^{2} M e_{1}=F^{2} A e_{1}=A F^{2} e_{1}=A e_{3}
    \ldots \ldots \ldots
    M e_{n}=M F^{n-1} e_{1}=F^{n-1} M e_{1}=F^{n-1} A e_{1}=A F^{n-1} e_{1}=A e_{n},所以,M=A.
    (2)解:由(1),C(F)=\operatorname{span}\left\{E, F, F^{2}, \cdots, F^{n-1}\right\},设x_{0} E+x_{1} F+x_{2} F^{2}+\cdots+x_{n-1} F^{n-1}=O,等式两边同右乘e_1,利用(*)
    \begin{aligned} \theta &=O e_{1}=\left(x_{0} E+x_{1} F+x_{2} F^{2}+\cdots+x_{n-1} F^{n-1}\right) e_{1} \\ &=x_{0} E e_{1}+x_{1} F e_{1}+x_{2} F^{2} e_{1}+\cdots+x_{n-1} F^{n-1} e_{1} \\ &=x_{0} e_{1}+x_{1} e_{2}+x_{2} e_{3}+\cdots+x_{n-1} e_{n} \end{aligned}e_{1}, e_{2}, e_{3}, \cdots, e_{n}线性无关,故,x_{0}=x_{1}=x_{2}=\dots=x_{n-1}=0,所以,E, F, F^{2}, \cdots, F^{n-1}线性无关.因此,E, F, F^{2}, \cdots, F^{n-1}C(F)的基,特别地,dimC(F)=n.
    三、【证明】:假设\lambda_0f的特征值,W是相应的特征子空间,即W=\{\eta \in V | f(\eta)=\lambda_{0} \eta\}.于是,Wf下是不变的.
    下面先证明,\lambda=0.任取非零\eta \in W,记m为使得\eta, g(\eta), g^{2}(\eta), \cdots, g^{m}(\eta)线性相关的最小的非负整数,于是,当0\leq i\leq m-1时,\eta, g(\eta), g^{2}(\eta), \cdots, g^{i}(\eta)线性无关。
    0\leq i\leq m-1时令W_{i}=\operatorname{span}\left\{\eta, g(\eta), g^{2}(\eta), \cdots, g^{i-1}(\eta)\right\},其中,W_{0}=\{\theta\}.因此,\operatorname{dim} W_{i}=i(1 \leq i \leq m),并且,W_{m}=W_{m+1}=W_{m+2}=显然,g\left(W_{i}\right) \subseteq W_{i+1},特别地,W_mg下是不变的.
    下面证明,W_mf下也是不变的.事实上,由f(\eta)=\lambda_{0} \eta,知\begin{aligned} f g(\eta) &=g f(\eta)+f(\eta)=\lambda_{0} g(\eta)+\lambda_{0} \eta_{0} \\ f g^{2}(\eta) &=g f g(\eta)+f g(\eta) \\ &=g\left(\lambda_{0} g(\eta)+\lambda_{0} \eta\right)+\left(\lambda_{0} g(\eta)+\lambda_{0} \eta\right) \\ &=\lambda_{0} g^{2}(\eta)+2 \lambda_{0} g(\eta)+\lambda_{0} \eta \end{aligned}
    根据\begin{aligned} f g^{k}(\eta) &=g f g^{k-1}(\eta)+f g^{k-1}(\eta) \\ &=g\left(f g^{k-1}\right)(\eta)+f g^{k-1}(\eta) \end{aligned}
    用归纳法不难证明,fg^k(\eta)一定可以表示成\eta, g(\eta), g^{2}(\eta), \cdots, g^{k}(\eta)的线性组合,且表达式中g^k(\eta)前的系数为\lambda_0,因而,这一限制的迹为m\lambda_0
    由于fg-gf=fW_m上仍然成立,而fg-gf的迹一定为零,故m\lambda_0=0,即\lambda_0=0
    任取\eta\in W由于
    f(\eta)=\theta, f g(\eta)=g f(\eta)+f(\eta)=g(\theta)+f(\eta)=\theta
    所以g(\eta)\in W因此,Wg下是不变的。从而,在W中存在g的特征向量,这也是f,g的公共特征向量。

    第二届

    二、【证明】:反证法。设方程有解,即存在复矩阵A使得A^2=B.
    注意到B的特征根为0,且其代数重根为3
    \lambdaA的一个特征根,则\lambda^2B的特征根,所以\lambda=0。从而A的特征根为0。于是AJordan 标准型只可能为
    J_{1}=\left( \begin{array}{lll}{0} & {0} & {0} \\ {0} & {0} & {0} \\ {0} & {0} & {0}\end{array}\right), J_{2}=\left( \begin{array}{lll}{0} & {1} & {0} \\ {0} & {0} & {0} \\ {0} & {0} & {0}\end{array}\right), J_{3}=\left( \begin{array}{ccc}{0} & {1} & {0} \\ {0} & {0} & {1} \\ {0} & {0} & {0}\end{array}\right)
    从而A^2Jordan 标准型只能为,J_{1}=J_{1}^{2}=J_{2}^{2}J_{2}=J_{3}^{2}。因此A^2的秩不大于1B=A^2的秩为2矛盾。所以X^2=B无解。
    六、【证明】:取任意实数r,由题设知(v+r \beta) A(v+r \beta)^{T} \geq 0,即v A v^{T}+r v A \beta^{T}+r \beta A v^{T}+r^{2} \beta A \beta^{T} \geq 0。亦即v A v^{T}+r\left(v A \beta^{T}+\beta A v^{T}\right)+r^{2} \beta A \beta^{T} \geq 0。若v A \beta^{T} \neq 0,则有v A \beta^{T}+\beta A v^{T} \neq 0,因此可取适当的实数r使得v A v^{T}+r\left(v A \beta^{T}+\beta A v^{T}\right)+r^{2} \beta A \beta^{T}<0矛盾。

    第三届

    三、【证明】:设\sigmaF^n的标准基\varepsilon_{1}, \cdots, \varepsilon_{n}下的矩阵为B,则\sigma(\alpha)=B \alpha,\left(\forall \alpha \in F^{n}\right)
    由条件:\forall A \in M_{n}(F), \sigma(A \alpha)=A \sigma(\alpha),\left(\forall \alpha \in F^{n}\right),有B A \alpha=A B \alpha, \forall \alpha \in F^{n}B A=A B\left(\forall A \in M_{n}(F)\right).
    B=\left(b_{i j}\right),取A=\operatorname{diag}(1, \cdots, 1, c, 1, \cdots, 1),其中c \neq 0,1,则由AB=BA可得b_{i j}=0, \forall i \neq j.又取A=I_{n}-E_{i i}-E_{j j}+E_{i j}+E_{j i},这里E_{st},是(s,t)-位置为1,其他位置为0的矩阵,则由AB=BA可得a_{i i}=a_{i j}(\forall i, j)。取\lambda=a_{11},故B=\lambda I_{n}。,从而\sigma=\lambda \cdot \mathrm{id}_{F^{n}}
    六、【证明】:设VFn维线性空间,\sigmaV上的线性变换,它在V的一组基下的矩阵为A。下面证明存在\sigma-不变子空间,V_1,V_2满足V=V_{1} \oplus V_{2},且\left.\sigma\right|_{V_{1}}是同构,\left.\sigma\right|_{V_{2}}是幂零变换。
    首先有子空间升链:\operatorname{Ker} \sigma \subseteq \operatorname{Ker} \sigma^{2} \subseteq \cdots \subseteq \operatorname{Ker} \sigma^{k} \subseteq从而存在正整数m使得\operatorname{Ker} \sigma^{m}=\operatorname{Ker} \sigma^{m+i}(i=1,2, \cdots)。进而有K e r \sigma^{m}=K e r \sigma^{2 m}.
    下面证明V=K e r \sigma^{m} \oplus \operatorname{Im} \sigma^{m}.
    \forall \alpha \in K e r \sigma^{m} \cap \operatorname{Im} \sigma^{m},由\alpha \in \operatorname{Im} \sigma^{m},存在\beta \in V使得\alpha=\sigma^{m}(\beta)。由此0=\sigma^{m}(\alpha)=\sigma^{2 m}(\beta),所以\beta \in K e r \sigma^{2 m}.
    从而\beta \in K e r \sigma^{m}=K e r \sigma^{2 m}.故\alpha=\sigma^{m}(\beta)=0 \operatorname{Ker} \sigma^{m} \cap \operatorname{Im} \sigma^{m},=(0)从而V=K e r \sigma^{m} \oplus \operatorname{Im} \sigma^{m}
    \sigma\left(\operatorname{Ker} \sigma^{m}\right) \subseteq K e r \sigma^{m}, \sigma\left(\operatorname{Im} \sigma^{m}\right) \subseteq \operatorname{Im} \sigma^{m},知\operatorname{Ker} \sigma^{m}, \operatorname{Im} \sigma^{m}\sigma-不变子空间。又由\sigma^{m}\left(K e r \sigma^{m}\right)=(0)\left.\sigma\right|_{K e r \sigma^{m}}是幂零变换。由\sigma\left(\operatorname{Im} \sigma^{m}\right) \subseteq Im \sigma^{m}\left.\sigma\right|_{\operatorname{lm} \sigma^{m}}是满线性变换,从而可逆。
    V_{1}=\operatorname{Im} \sigma^{m}, V_{2}=\mathrm{ker} \sigma^{m}中各找一组基\alpha_{1}, \cdots, \alpha_{s} ; \beta_{1}, \cdots, \beta_{t},合并成V的一组基,\sigma在此基下的矩阵为\left( \begin{array}{ll}{B} & {0} \\ {0} & {C}\end{array}\right),其中B\left.\sigma\right|_{V_{1}}在基\alpha_{1}, \cdots, \alpha_{s}下的矩阵,从而可逆;C\left.\sigma\right|_{V_{2}}在基\beta_{1}, \cdots, \beta_{t}下的矩阵,是幂零矩阵。从而A
    相似于\left( \begin{array}{ll}{B} & {0} \\ {0} & {C}\end{array}\right),其中B是可逆矩阵,C是幂零矩阵。【注】如果视F为复数域直接用若当标准型证明,证明正确可给10分:
    存在可逆矩阵P,使得
    P^{-1} A P=\operatorname{diag}\left(J\left(\lambda_{1}, n_{1}\right), \cdots, J\left(\lambda_{s}, n_{s}\right), J\left(0, m_{1}\right), \cdots, J\left(0, m_{t}\right)\right)
    其中J\left(\lambda_{i}, n_{i}\right)是特征值为\lambda_i的阶为n_i,的若当块\lambda_{i} \neq 0J\left(0, m_{j}\right)是特征值为0的阶为m,的若当块。令\begin{aligned} B &=\operatorname{diag}\left(J\left(\lambda_{1}, n_{1}\right), \cdots, J\left(\lambda_{s}, n_{s}\right)\right) \\ C &=\operatorname{diag}\left(J\left(0, m_{1}\right), \cdots, J\left(0, m_{t}\right)\right) \end{aligned}
    B为可逆矩阵,C为幂零矩阵,A相似于\left( \begin{array}{ll}{B} & {0} \\ {0} & {C}\end{array}\right)

    第四届

    4.【解】:设\lambdaf(t)的根,则有detP(t)=0.从而P(t)n个列线性相关。于是存在\alpha\not=0,使得P(\lambda) \alpha=0,进而\alpha^{*} P(\lambda) \alpha=0.
    具体地,\alpha^{*} A \alpha \lambda^{2}+\alpha^{*} B \alpha \lambda+\alpha^{*} C \alpha=0.令a=\alpha^{*} A \alpha, b=\alpha^{*} B \alpha, c=\alpha^{*} C \alpha,则A,B,C皆为正定矩阵知a>0,b>0,c>0,且\lambda=\frac{-b \pm \sqrt{b^{2}-4 a c}}{2 a}
    注意到,当b^{2}-4 a c \geq 0时,\sqrt{b^{2}-4 a c}<b,从而有\operatorname{Re} \lambda=\frac{-b \pm \sqrt{b^{2}-4 a c}}{2 a}<0.
    b^2-4ac<0时,\sqrt{b^{2}-4 a c}=i \sqrt{4 a c-b^{2}},从而有\operatorname{Re} \lambda=\frac{-b}{2 a}<0

    【解】:(1)矩阵方程AX=B有解等价于B的列向量可由A的列向量线性表示。BY=A无解等价于A的某个列向量不能由B的列向量线性表示。对(A,B)作初等行变换:
    \left( \begin{array}{llll}{2} & {2} & {4} & {b} \\ {2} & {a} & {3} & {1}\end{array}\right) \rightarrow \left( \begin{array}{cccc}{2} & {2} & {4} & {b} \\ {0} & {a-2} & {-1} & {1-b}\end{array}\right)
    可知,B的列向量组可由A的列向量线性表示当且仅当a \neq 2.对矩阵(B,A)做初等行变换:
    (B, A)=\left( \begin{array}{llll}{4} & {b} & {2} & {2} \\ {3} & {1} & {2} & {a}\end{array}\right) \rightarrow \left( \begin{array}{cccc}{4} & {b} & {2} & {2} \\ {0} & {1-3 b / 4} & {1 / 2} & {a-3 / 2}\end{array}\right)
    由此可知A的列向量组不能由B的列向量或性表示的重要条件是b=\frac{4}{3}所以矩阵方程AX=B有解但BY=A无解的充要条件是a \neq 2, b=\frac{4}{3}
    (2)若A,B相似,则有trA=trB|A|=|B|,故有a=3,b=\frac{2}{3}.反之,若a=3, b=\frac{2}{3},则有A=\left( \begin{array}{ll}{2} & {2} \\ {2} & {3}\end{array}\right), B=\left( \begin{array}{ll}{4} & {2 / 3} \\ {3} & {1}\end{array}\right)
    AB的特征多项式均为\lambda^{2}-5 \lambda+2。由于\lambda^{2}-5 \lambda+2有两个不同的根,从而AB都可以相似于同一对角阵,所以AB相似。
    (3)由于A为对称阵,若AB合同,则B也是对称阵,故b=3。矩阵B对应的二次型为
    g\left(x_{1}, x_{2}\right)=4 x_{1}^{2}+6 x_{1} x_{2}+x_{2}^{2}=\left(3 x_{1}+x_{2}\right)^{2}-5 x_{1}^{2}
    在可逆线性变换,y_{1}=3 x_{1}+x_{2}, y_{2}=x_{1}下,g(x_1,x_2)变成标准型:y_{1}^{2}-5 y_{2}^{2}。由此,B的正、负惯性指数为1。类似地,A的对应二次型为
    f\left(x_{1}, x_{2}\right)=2 x_{1}^{2}+4 x_{1} x_{2}+a x_{2}^{2}=2\left(x_{1}+x_{2}\right)^{2}+(a-2) x_{2}^{2}
    在可逆线性变换z_{1}=x_{1}+x_{2}, z_{2}=x_{2}下,f(x_1,x_2)变成标准型:2 z_{1}^{2}+(a-2) z_{2}^{2}AB合同的充要条件是它们有相同的正、负惯性指数,故AB合同的充要件是a<2,b=3.

    第五届

    二、【证明】:设B(t)的第i列为B_i(t),i=1,2…,n。断言:t-t_0d(t), d_{1}(t), \cdots, d_{n}(t)的公因式。反证。
    不失一般性,设d_{1}\left(t_{0}\right) \neq 0,于是
    \left[B\left(t_{0}\right), b\left(t_{0}\right)\right]=n,因为d_{1}\left(t_{0}\right) \neq 0.
    注意到秩B\left(t_{0}\right) \leq n-1,结果
    增广阵\left[B\left(t_{0}\right), b\left(t_{0}\right)\right]的秩不等于B\left(t_{0}\right)的秩,从而B\left(t_{0}\right) X=b\left(t_{0}\right)不相容。矛盾。
    六、
    【证明】:(1)\forall A, B \in \Gamma_{r},表明A可以表示为A=PBQ,其中P,Q可逆。结果\phi(A)=\phi(P) \phi(B) \phi(Q),从而秩\phi(A)\leq\phi(B);对称地有,秩\phi(B)\leq\phi(A);即有秩\phi(A)=秩\phi(B)成立。
    (2)考察矩阵集合\left\{\phi\left(E_{i j}\right) | i, j=1,2, \cdots, n\right\}。考察\phi\left(E_{11}\right) \cdots, \phi\left(E_{m n}\right).由(1)知\phi\left(E_{i j}\right)为非零阵,特别地,\phi\left(E_{i i}\right)为非零幂等阵,故存在单位特征向量w_i使得\phi\left(E_{i i}\right) w_{i}=w_{i}, i=1,2, \cdots, n
    从而得向量值w_{1}, w_{2}, \cdots, w_{n}
    此向量组有如下性质:
    a) \phi\left(E_{i i}\right) w_{k}=\left\{\begin{array}{l}{\phi\left(E_{i i}\right) \phi\left(E_{k k}\right) w_{k}=\phi\left(E_{i i} E_{k k}\right) w_{k}=0, k \neq i} \\ {w_{i}, k=i}\end{array}\right.
    b)W_{1}, W_{2}, \cdots, W_{n}线性无关,从而构成R^n的基,矩阵W=\left[w_{1}, w_{2}, \cdots, w_{n}\right]
    为可逆矩阵。事实上,x_{1} w_{1}+x_{2} w_{2}+\dots+x_{n} w_{n}=0,则在两边用\phi\left(E_{i i}\right)作用之,得x_{i}=0, i=1,2, \cdots, n
    c)当k\not=j时,\phi\left(E_{i j}\right) w_{k}=\phi\left(E_{i j}\right) \phi\left(E_{k k}\right) w_{k}=\phi\left(E_{i j} E_{\mu k}\right) w_{k}=0
    k=j时,\phi\left(E_{i j}\right) w_{k}=b_{1 j} w_{1}+\cdots+b_{i j} w_{i}+\cdots+b_{n j} w_{n}两边分别用\phi\left(E_{11}\right), \cdots, \phi\left(E_{i-1, i-1}\right), \phi\left(E_{i+1, j+1}\right), \cdots, \phi\left(E_{m n}\right)作用,得
    0=\phi\left(E_{11} E_{i j}\right) w_{j}=\phi\left(E_{11}\right) \phi\left(E_{i j}\right) w_{k}=b_{1 j} w_{1}, \cdots
    0=\phi\left(E_{m} E_{i j}\right) w_{j}=\phi\left(E_{m}\right)\left(b_{1 j} w_{1}+\cdots+b_{i j} w_{i}+\cdots+b_{n j} w_{n}\right)=b_{n j} w_{n}
    即有b_{1 j}=\cdots=b_{i-1, j}=b_{i+1, j}=\cdots=b_{n j}=0从而\phi\left(E_{i j}\right) w_{j}=b_{i j} w_{i}
    进一步,b_{ij}\not=0,否则有
    \phi\left(E_{i j}\right)\left[w_{1}, w_{2}, \cdots, w_{n}\right]=0
    导致\phi(E_{ij})为零阵,不可能,这样通过计算\phi\left(E_{i j}\right) w_{j}, i, j=1,2, \cdots, n我们得到n^2个非零的实数:\begin{array}{ccc}{b_{11}} & {\cdots} & {b_{1 n}} \\ {\vdots} & {\vdots} & {\vdots} \\ {b_{n 1}} & {\cdots} & {b_{m}}\end{array}
    注意到E_{m r} E_{r s}=E_{m s},从而有\begin{aligned} b_{m s} w_{m} &=\phi\left(E_{m s}\right) w_{s}=\phi\left(E_{m r}\right) \phi\left(E_{r s}\right) w_{s} \\ &=\phi\left(E_{m r}\right) b_{r s} w_{r}=b_{r s} b_{m r} w_{m} \end{aligned}
    因此有b_{m r} b_{r s}=b_{m s}
    最后,令v_{i}=b_{i 1} w_{i}, i=1,2, \cdots, n,则有
    \phi\left(E_{i j}\right) v_{k}=\left\{\begin{array}{l}{\phi\left(E_{i j}\right) b_{j 1} w_{j}=b_{j 1} b_{i j} w_{i}=b_{i 1} w_{i}=v_{i}, k=j} \\ {0, k \neq j}\end{array}\right.R=\left[v_{1}, \cdots, v_{n}\right]R=\left[w_{1}, \cdots, w_{n}\right] \left( \begin{array}{cccc}{b_{11}} & {{ } & {{ } \\ {{ } & {\ddots} & {{ } \\ {{ } & {{ } & {{ } & {b_{n 1}}\end{array}\right)为可逆矩阵,且
    \begin{aligned} \phi\left(E_{i j}\right) R &=\phi\left(E_{i j}\right)\left[v_{1}, \cdots, v_{n}\right] \\ &=\left[0, \cdots, 0, v_{i}, 0, \cdots, 0\right]=\left[v_{1}, \cdots, v_{n}\right] E_{i j} \end{aligned}

    \phi\left(E_{i j}\right)=R E_{i j} R^{-1}

    第六届

    三、【证明】:2)\Rightarrow1).
    考虑方程\lambda_{1} f_{1}+\dots+\lambda_{n} f_{n}=0.将a_{1}, \cdots, a_{n}分别代入,得
    \left\{\begin{array}{l}{\lambda_{1} f_{1}\left(a_{1}\right)+\cdots+\lambda_{n} f_{n}\left(a_{1}\right)=0} \\ {\dots \cdots} \\ {\lambda_{1} f_{1}\left(a_{n}\right)+\cdots+\lambda_{n} f_{n}\left(a_{n}\right)=0}\end{array}\right.
    注意到上述方程组的系数矩阵为\left(f_{i}\left(a_{j}\right)\right)^{T},因此\operatorname{det}\left[f_{i}\left(a_{j}\right)\right] \neq 0直接知道\lambda_{1}=\cdots=\lambda_{n}=0
    1)\Rightarrow2).用归纳法。首先,n=1时显然成立;其次,设n=k时结论成立,则n=k+1时,由f_{1}, \cdots, f_{k+1}线性无关知,f_{1}, \cdots, f_{k}线性无关。因此\exists a_{1}, \cdots, a_{k} \in[0,1]使得\operatorname{det}\left[f_{i}\left(a_{j}\right)\right]_{k \times k} \neq 0.观察函数

    image.png

    按最后一列展开得F(x)=\lambda_{1} f_{1}(x)+\cdots+\lambda_{k} f_{k}(x)+\lambda_{k-1} f_{k-1}(x)
    其中\lambda_{1}, \cdots, \lambda_{k+1}均为常量。注意到\lambda_{k+1} \neq 0,由f_{1}, \cdots, f_{k+1}线性无关知F(x)不恒为0,从而\exists a_{k+1} \in[0,1]使得F\left(a_{k+1}\right) \neq 0亦即a_{1}, \cdots, a_{k+1} \in[0,1], \operatorname{det}\left[f_{i}\left(a_{j}\right)\right] \neq 0。证毕。
    五、【证明】:(1)令

    image.png

    则所求的方程变为X^{n}+X^{l}=2 I+2 H+3 H^{2}+\cdots+m H^{m-1}
    (2)考察形如\left( \begin{array}{cccccc}{1} & {0} & {0} & {\cdots} & {0} & {0} \\ {a_{1}} & {1} & {0} & {\cdots} & {0} & {0} \\ {a_{2}} & {a_{1}} & {1} & {\cdots} & {0} & {0} \\ {\vdots} & {\vdots} & {\vdots} & {\ddots} & {\vdots} & {\vdots} \\ {a_{m-1}} & {a_{m-2}} & {a_{m-3}} & {\cdots} & {1} & {0} \\ {a_{m}} & {a_{m-1}} & {a_{m-2}} & {\cdots} & {a_{1}} & {1}\end{array}\right)的矩阵X,则有X=I+a_{1} H+a_{2} H^{2}+\cdots+a_{m} H^{m-1}.结果\begin{aligned} X^{n}=&\left(I+a_{1} H+a_{2} H^{2}+\cdots+a_{m} H^{m-1}\right)^{n} \\ &=I+\left(n a_{1}\right) H+\left(n a_{2}+f_{1}\left(a_{1}\right)\right) H^{2}+\cdots \\ &+\left(n a_{m}+f_{m-1}\left(a_{1}, \cdots, a_{m-1}\right)\right) H^{m-1} \end{aligned},其中f_{1}\left(a_{1}\right)a_1,确定\cdots, f_{m-1}\left(a_{1}, \cdots, a_{m-1}\right)a_{1}, \cdots, a_{m-1}确定。
    类似地,有X^{\prime}=I+\left(l a_{1}\right) H+\left(l a_{2}+g_{1}\left(a_{1}\right)\right) H^{2}+\cdots
    +\left(l a_{m}+g_{m-1}\left(a_{1}, \cdots, a_{m-1}\right)\right) H^{m-1}
    (3)观察下列方程组
    \left\{\begin{array}{l}{(n+l) a_{1}=2} \\ {(n+l) a_{2}+\left(f_{1}\left(a_{1}\right)+g_{1}\left(a_{1}\right)\right)=3} \\ {\dots \cdots} \\ {(n+l) a_{m}+\left(f_{m-1}\left(a_{1}, \cdots, a_{m-1}\right)+g_{m-1}\left(a_{1}, \cdots, a_{m-1}\right)\right)=m}\end{array}\right.
    直接可以看出该方程组有解,命题得证。

    第七届

    二、【解】:|A|=\frac{1}{24}
    过程如下:
    首先,记A的4各特征值为,\lambda_{1}, \lambda_{2}, \lambda_{3}, \lambda_{4}A的特征多项式为p(\lambda)=\lambda^{4}+a_{3} \lambda^{3}+a_{2} \lambda^{2}+a_{1} \lambda+a_{0},则由p(\lambda)=\left(\lambda-\lambda_{1}\right)\left(\lambda-\lambda_{2}\right)\left(\lambda-\lambda_{3}\right)\left(\lambda-\lambda_{4}\right)可知
    \left\{\begin{array}{l}{a_{3}=-\left(\lambda_{1}+\lambda_{2}+\lambda_{3}+\lambda_{4}\right)} \\ {a_{2}=\lambda_{1} \lambda_{2}+\lambda_{1} \lambda_{3}+\lambda_{1} \lambda_{4}+\lambda_{2} \lambda_{3}+\lambda_{2} \lambda_{4}+\lambda_{3} \lambda_{4}} \\ {a_{1}=-\left(\lambda_{1} \lambda_{2} \lambda_{3}+\lambda_{1} \lambda_{2} \lambda_{4}+\lambda_{1} \lambda_{3} \lambda_{4}+\lambda_{4} \lambda_{2} \lambda_{3}\right)} \\ {a_{0}=|A|=\lambda_{1} \lambda_{2} \lambda_{3} \lambda_{4}}\end{array}\right.
    齐次,由于迹在相似变换下保持不变,故由A的约当标准型(或Scur分解),有\left\{\begin{array}{l}{\lambda_{1}+\lambda_{2}+\lambda_{3}+\lambda_{4}=1} \\ {\lambda_{1}^{2}+\lambda_{2}^{2}+\lambda_{3}^{2}+\lambda_{4}^{2}=2} \\ {\lambda_{1}^{3}+\lambda_{2}^{3}+\lambda_{3}^{3}+\lambda_{4}^{3}=3} \\ {\lambda_{1}^{4}+\lambda_{2}^{4}+\lambda_{3}^{4}+\lambda_{4}^{4}=4}\end{array}\right.由(1)和(2)得a_{2}=\lambda_{1} \lambda_{2}+\lambda_{1} \lambda_{3}+\lambda_{1} \lambda_{4}+\lambda_{2} \lambda_{3}+\lambda_{2} \lambda_{4}+\lambda_{3} \lambda_{4}=-\frac{1}{2}
    由(1)两边立方得\begin{aligned} 1=\lambda_{1}^{3}+\lambda_{2}^{3}+\lambda_{3}^{3}+\lambda_{4}^{3}+3 \lambda_{1}^{2}\left(\lambda_{2}+\lambda_{3}+\lambda_{4}\right) & \\+3 \lambda_{2}^{2}\left(\lambda_{1}+\lambda_{3}+\lambda_{4}\right)+3 \lambda_{3}^{2}\left(\lambda_{1}+\lambda_{2}+\lambda_{4}\right) \\+3 \lambda_{4}^{2}\left(\lambda_{1}+\lambda_{2}+\lambda_{3}\right)-6 a_{1} \end{aligned}再由(1)(2)(3)可以得到1=3+3\left(\lambda_{1}^{2}+\lambda_{2}^{2}+\lambda_{3}^{2}+\lambda_{4}^{2}\right)-3\left(\lambda_{1}^{3}+\lambda_{2}^{3}+\lambda_{3}^{3}+\lambda_{4}^{3}\right)-6 a_{1}
    a_{1}=-\frac{1}{6}最后,由p(\lambda)=\lambda^{4}-\lambda^{3}-\frac{1}{2} \lambda^{2}-\frac{1}{6} \lambda+a_{0}\left\{\begin{array}{l}{p\left(\lambda_{1}\right)=0} \\ {\vdots} \\ {p\left(\lambda_{4}\right)=0}\end{array}\right.相加得4-3-\frac{1}{2} \times 2-\frac{1}{6} \times 1+4 a_{0}=0 \Rightarrow a_{0}=\frac{1}{24},即|A|=\frac{1}{24}
    三、【证明】:设C=l+A,B=A²,An各特征值为\lambda_{1}, \lambda_{2}, \cdots, \lambda_{n},则Bn各特征值为\lambda_{1}^{2}, \lambda_{2}^{2}, \cdots, \lambda_{n}^{2}Cn各特征值为\mu_{1}=\lambda_{1}+1, \mu_{2}=\lambda_{2}+1, \cdots, \mu_{n}=\lambda_{n}+1C的特征多项式为p_{C}(\lambda)=\left(\lambda-\mu_{1}\right)\left(\lambda-\mu_{2}\right) \cdots\left(\lambda-\mu_{n}\right)。若XX+A X-X A^{2}=0的解,则有CX=XB;进而有C^{2} X=X B^{2}, \cdots C^{k} X=X B^{k}, \cdots,结果有0=p_{C}(C) X=X p_{C}(B)=X\left(B-\mu_{1} I\right) \cdots\left(B-\mu_{n} I\right)
    注意到Bn各特征值皆为偶数,而Cn各特征值皆为奇数,所以B-\mu_{1} I, \cdots, B-\mu_{n} I皆为可逆矩阵,结果由0=X\left(B-\mu_{1} I\right) \cdots\left(B-\mu_{n} I\right)
    立即可得X=0.

    第八届

    二、【证明】:由秩不等式rankA+rankB\leq rank(BA)+n,得rankA+rankB≤n.
    结果\operatorname{rank} A \leq \frac{n}{2}\operatorname{rank} B \leq \frac{n}{2}
    注意到n为奇数,故有\operatorname{rank} A<\frac{n}{2}\operatorname{rank} B<\frac{n}{2}成立。
    \operatorname{rank} A<\frac{n}{2},则
    \operatorname{rank}\left(A+J_{A}\right) \leq \operatorname{rank} A+\operatorname{rank} J_{A}<n故0\in S_{1};或\operatorname{rank} B<\frac{n}{2},则
    \operatorname{rank}\left(B+J_{B}\right) \leq \operatorname{rank} B+\operatorname{rank} J_{B}<n故0\in S_{2}.
    所以最终有0\in S_{1} \cup S_{2}.
    三、【证明】:记
    A_{1}=\left(p_{1}^{(1)}, \cdots, p_{2016}^{(1)}\right), \cdots, A_{2017}=\left(p_{1}^{(2017)}, \cdots, p_{2016}^{(2017)}\right)
    考虑线性方程组
    x_{1} p_{1}^{(1)}+\cdots+x_{2017} p_{1}^{(2017)}=0
    由于未知数个数大于方程个数,故该线性方程组必有非零解x_{1} p_{1}^{(1)}+\cdots+x_{2017} p_{1}^{(2017)}=0从而c_{1} A_{1}+\cdots+c_{2017} A_{2017}的第一列为0,更有\operatorname{det}\left(c_{1} A_{1}+\cdots+c_{2017} A_{2017}\right)=0.

    第九届

    三、证明:必要性:由迹的性质直接知.(2分)
    充分性:首先,对于可逆矩阵W \in \Gamma,有W W_{1}, \ldots, W W_{r},各不相同.故有W \Gamma \equiv\left\{W W_{1}, W W_{2}, \ldots, W W_{r}\right\}=\left\{W_{1}, W_{2}, \ldots, W_{r}\right\},即,W \Gamma=\Gamma, \forall W \in \Gamma(7分)
    S=\Sigma_{i=1}^{r} W_{i},则W S=S, \forall W \in \Gamma.进而S^{2}=r S,即S^{2}-r S=0\lambdaS的特征值,则\lambda^{2}-r \lambda=0.即\lambda=0r.
    结合条件\sum_{i=1}^{r} t r\left(W_{i}\right)=0知,S的特征值只能为0.因此有S-r I可逆(例如取S的约当分解就可直接看出)再次注意到S(S-r I)=S^{2}-r S=0,此时右乘(S-r I)^{-1}即得S=0.证毕.
    四、证明:反证.若XN+Y^TM^T=0,则有N^{T} X^{T}+M Y=0.
    另外,由(X, Y) \in TX Y+(X Y)^{T}=2 a I
    X Y+Y^{T} X^{T}=2 a I
    类似有
    M N+N^{T} M^{T}=2 a I
    因此,
    \left( \begin{array}{cc}{X} & {Y^{T}} \\ {M} & {N^{T}}\end{array}\right) \left( \begin{array}{cc}{Y} & {N} \\ {X^{T}} & {M^{T}}\end{array}\right)=2 a \left( \begin{array}{cc}{I} & {0} \\ {0} & {I}\end{array}\right)进而
    \frac{1}{2 a} \left( \begin{array}{cc}{Y} & {N} \\ {X^{T}} & {M^{T}}\end{array}\right) \left( \begin{array}{cc}{X} & {Y^{T}} \\ {M} & {N^{T}}\end{array}\right)=\left( \begin{array}{ll}{I} & {0} \\ {0} & {I}\end{array}\right)
    Y Y^{T}+N N^{T}=0所以
    Y=0, N=0导致XY=0,与X Y=a I+A \neq 0矛盾,证毕。

    相关文章

      网友评论

        本文标题:高代——数学竞赛初赛答案

        本文链接:https://www.haomeiwen.com/subject/qsmbeqtx.html