and the eigenvalues of the matrix M are k
¼ 0 and k
¼ a2 þ ð1 — xÞ , respectively。
að1 — xÞ ð1 — xÞ2
1 2
。
Hence, the matrix M is a semi-positive definite one, i。e。 M P 0。 For any vector x – 0; xT Mx ¼ 0 implies x ¼ h —
。 。
1—x 。
a
1
where
h 2 R and h – 0。 Consider the set U ¼ n 2 R2 : nT n ¼ 1 。 Then U is a bounded closed set。 By Claim 1, if /
¼ 0; fðtÞ ¼ 0, which
implies that nT Mn – 0。 Since the function nT Mn is continuous with respect to n and for any n 2 U; nT Mn – 0, we have that min nT Mn, denoted by k , exists and is larger than zero。
Consider (23) which can be rewritten as
where k· k2 presents the 2-norm。 Since f
2 U; /
P k1jjfjj
P k1 y2, which implies that j/m j1þa P k1þa
jy2 j1þa 。 Therefore, we
have
1 — xÞ2 b2j/
kfk2 m
2a
j
2 2 1
ð1 — xÞ2 b2j/ j2a
!2ðy1 ; y2Þ¼
2a
ð m
a
2a P
2a
2a
m
2a
2 4a 4a 4a 2a 4a
1þa 2a 1þ
1þa
1þa
1þa
1þa 2a
1þa
1þa
1þa 1 1þa
1þa
q1 j/mj
þ q2 j/m j
þ q3 jy2 j
q1 j/m j
þ q2 j/mj
þ q3 ðk Þ
j/mj
2 2 2a 2 2
P ð1 — xÞ b j/mj , ð1 — xÞ b D /
q0
n 1þa
2a
j m j
n 1þa
4a
þ j/m j1þa
1þa 1 1þa oo
q0