?fx(x,y,z)?[a?2x(ax?by?cz)]e?(x?y?z)?0,???(x2?y2?z2)f(x,y,z)?[b?2y(ax?by?cz)]e?0, ?y??(x2?y2?z2)f(x,y,z)?[c?2z(ax?by?cz)]e?0,??z解得稳定点为
222P1(a2(a?b?c)a2(a?b?c)222222,b2(a?b?c)b222,c2(a?b?c),?c222),
P2(?,?2(a?b?c)2222(a?b?c)222).
可以通过求在每一个稳定点的Hesse矩阵,知f(x,y,z)在P1取极大值,在P2取极小值.由于极大值点与极小值点均唯一,故极大值点与极小值就是最大值点与最小值点,最大
a2?b2?c2?2值为f(Pe,最小值为f(P2)?1)?25.求证:
221a2?b2?c2?2e.
212(1)f(x,y)?Ax?2Bxy?Cy?2Dx?2Ey?F在R有最小值,无最大值.其
2中A?0,B?AC;
(2)f(x,y)?xy?11?在0?x,y???有最小值,无最大值. xy?fx(x,y)?2Ax?2By?2D?0,证明(1)令?
f(x,y)?2Bx?2Cy?2E?0,?y由于
??ABBC?AC?B2?0,
故以上方程组有唯一一组解,即有唯一稳定点P0(x0,y0).又因为
a11?fxx(P0)?2A?0,a22?fyy(P0)?2C,a12?fxy(P0)?2B,
且D?a11a22?a12?4B?4AC?0,故f(x,y)在P0取极小值.由于极小值点唯一,因而就是最小值点,所以f(x,y)在R有最小值,无最大值.
(2)令
2221?f(x,y)?y??0,x2?x? ?1?fy(x,y)?x??0,2?y?解得唯一稳定点P0(1,1),又因为
a11?fxx(P0)?2?0,a22?fyy(P0)?2?0,a12?fxy(P0)?1,
且D?a11a22?a12?3?0,所以f(x,y)?xy?211?在P0(1,1)取极小值,又极小值点xy唯一,故就是最小值点,即f(x,y)在0?x,y???有最小值,无最大值.
6.设F(x,y,z)有二阶连续偏导数,并且
F(x0,y0,z0)?0,Fz(x0,y0,z0)?0.
讨论由F(x,y,z)?0确定的隐函数z?f(x,y)在(x0,y0)取得极值的必要和充分条件,再由
x2?y2?z2?2x?2y?4z?10?0
所确定的z?f(x,y)的极值.
解 取极值的必要条件为在(x0,y0),有
??z??x????z??y??(x0,y0)?f(x0,y0)F(x,y,z)??x000?0,?xFz(x0,y0,z0)?f(x0,y0)???0,?yFz(x0,y0,z0)Fy(x0,y0,z0)
?(x0,y0)即Fx(x0,y0,z0)?0且Fy(x0,y0,z0)?0为在(x0,y0)取得极值的必要条件.
在F(x,y,z)?0两边对x,y二次求导,有
?z?F?F?0,xz??x? ??z?Fy?Fz?0,??y???zF?F?Fzx?xxxy?x???z?F?F?Fzy?xyxz?y???z?Fzy?Fyy?Fyz??y?并利用在(x0,y0)有,
?z?z2?2z?Fzz()?Fz2?0,?x?x?x?z?z?z?2z?Fzz?Fz?0, ?x?x?y?x?y?z?z2?2z?Fzz()?Fz2?0,?y?y?y?z?z?0,就有 ?0,?y?x?2z?x2?2z?y2(x0,y0)Fxx(x0,y0,z0)?2z??,
Fz(x0,y0,z0)?x?y??Fyy(x0,y0,z0)Fz(x0,y0,z0),
??(x0,y0)Fxy(x0,y0,z0)Fz(x0,y0,z0),
(x0,y0)因此在(x0,y0),隐函数z?f(x,y)取极值的充分条件为
Fyy(x0,y0,z0)Fxy(x0,y0,z0)2Fxx(x0,y0,z0)(?)(?)?(?) Fz(x0,y0,z0)Fz(x0,y0,z0)Fz(x0,y0,z0) ?2Fxx(x0,y0,z0)Fyy(x0,y0,z0)?Fxy(x0,y0,z0)Fz2(x0,y0,z0)?0,
Fxx(x0,y0,z0)?0(或
Fz(x0,y0,z0)2即Fxx(x0,y0,z0)Fyy(x0,y0,z0)?Fxy(x0,y0,z0)?0.且当??Fyy(x0,y0,z0)Fz(x0,y0,z0)时,取极小值, 当??0)
Fyy(x0,y0,z0)Fxx(x0,y0,z0)?0(或??0)
Fz(x0,y0,z0)Fz(x0,y0,z0)2时,取极大值.
即当Fxx(x0,y0,z0)Fyy(x0,y0,z0)?Fxy(x0,y0,z0)?0时取极值,当上式为负时,不取极值,为0时不定.
设F(x,y,z)?x?y?z?2x?2y?4z?10,则令
222?Fx(x,y,z)?2x?2?0, ?F(x,y,z)?2y?2?0,?y解出稳定点(1,?1),对应空间中的点为P1(1,?1,6)与P2(1,?1,?2),且
Fxx?2,Fxy?0,Fyy?2,
2因为D?FxxFyy?Fxy?4?0(在P1与P2),且Fz(P1)?(2z?4)|P1?8,
Fz(P2)?(2z?4)|P2??8,所以有:
由于?Fxx(P1)F(P)22???0,故在P1(1,?1,6)取极大值6;而?xx2??0,故在
Fz(P1)8Fz(P2)8P2(1,?1,?2)取极小值?2.
7.求下列隐函数的极大值和极小值: (1)(x?y)?(y?z)?(z?x)?3; (2)z?xyz?x?xy?9?0.
解(1)设F(x,y,z)?(x?y)?(y?z)?(z?x)?3.则由
222222222?Fx(x,y,z)?2(x?y)?2(z?x)?0,??Fy(x,y,z)?2(x?y)?2(z?y)?0, ??F(x,y,z)?0,得到两个稳定点P,?,?).因为Fxx?4,Fxy?2,Fyy?4,所以1(,,?),P2(?2D?FxxFyy?Fxy?12?0,且
112232121212Fz(P1)?[2(y?z)?2(z?x)]|P1??4,Fz(P2)?[2(y?z)?2(z?x)]|P2?4,
因此,?Fxx(P1)F(P)11311?1?0,故在(,)取极小值?;?xx2??1?0,故在(?,?)Fz(P1)Fz(P2)22222取极大值?1. 2222(2)设F(x,y,z)?z?xyz?x?xy?9,由方程
?Fx(x,y,z)?yz?2x?y2?0,? ?Fy(x,y,z)?xz?2xy?0,??F(x,y,z)?0,解得稳定点为P1,2(0,0,?3),P3,4(0,?3,?3),P5,6(1,?2,?22),且
Fxx??2,Fxy?z?2y,Fyy??2x,
在点P1)??2,Fxy(P1)?3,Fyy(P1)?0,D??9?0,故函数在1(0,0,3),有Fxx(PP1(0,0,3)不取极值;
在点P2(0,0,?3),有Fxx(P2)??2,Fxy(P2)??3,Fyy(P2)?0,D??9?0,故函数在P2(0,0,?3)不取极值;
在点P3(0,3,3),有Fxx(P3)??2,Fxy(P3)??3,Fyy(P3)?0,D??9?0,故函数在P3(0,3,3)不取极值;
在点P4(0,?3,?3),有Fxx(P4)??2,Fxy(P4)?3,Fyy(P4)?0,D??9?0,故函数在P4(0,?3,?3)也不取极值;
在点P5(1,2,22),有Fxx(P5)??2,Fxy(P5)?0,Fyy(P5)??2,D?4?0,而且
Fz(P5)?(2z?xy)]|P5?52,因而
?Fxx(P5)1?2?0,
Fz(P5)5故函数在P5(1,2,22)取极小值22;
在点P6(1,?2,?22),有Fxx(P6)??2,Fxy(P6)?0,Fyy(P6)??2,D?4?0,而且Fz(P6)?(2z?xy)|P6??52,因此
?Fxx(P6)1??2?0,
Fz(P6)5所以,函数在P6(1,?2,?22)取极大值?22.
8.在已知周长为2p的一切三角形中,求出面积为最大的三角形. 解 设三角形其中两边长为x,y,则另一边长为2p?x?y,面积为
s?p(p?x)(p?y)(p?(2p?x?y))?p(p?x)(p?y)(x?y?p),
由于s与s的最值点相同,而
2s2?p(p?x)(p?y)(x?y?p)?F(x,y),
令
?Fx(x,y)?p(p?y)(2p?2x?y)?0, ?F(x,y,z)?p(p?x)(2p?x?2y)?0,?y