820 CHAPTER 24. CRITICAL POINTS

where r→ h(r) is a decreasing function which has values in (0,1] and equals 1 for r ∈ [0,1]and equals 1/r for r > 1.

1h(r) = 1

r

h(r) = 1

Here u is given and the η ′ is the time derivative is with respect to t. By Corollary24.1.11, there exists a solution to this for t ∈ [0,1].

Then for this solution, η (0,u) = u because of the above initial condition. If u ∈I−1 (X \ [c− ε,c+ ε]) , then u ∈ A and so g(u) = 0 so η (t,u) = u for all t ∈ [0,1]. Thisgives the first two conditions. Consider the third.

ddt

(I (η (t,u))) =⟨I′ (η) ,η ′

⟩=−

⟨I′ (η) ,g(u)h(∥V (η (t,u))∥)V (η (t,u))

⟩= −g(u)h(∥V (η (t,u))∥)

⟨I′ (η) ,V (η (t,u))

⟩≤ −g(u)h(∥V (η (t,u))∥)

∥∥I′ (η)∥∥2

X ′ ≤ 0

this last inequality from the inequalities of 24.1.5, and so this implies the third conditionsince it says that the function t→ I (η (t,u)) is decreasing.

It remains to consider the last condition. This involves an appropriate choice of smallδ . It was chosen small and now it will be seen how small. It is desired to verify that

η(1, I−1 ((−∞,c+δ ])

)⊆ I−1 ((−∞,c−δ ])

Suppose it is not so. Then there exists u such that I (u) ∈ (c− δ ,c+ δ ] but I (η (1,u)) >c−δ . We can assume that I (u)∈ (c−δ ,c+δ ] because if I (u)≤ c−δ , then so is I (η (1,u))from what was just shown. Hence g(u) = 1. Then using the fact that g(u) = 1,

c−δ < I (η (1,u)) = I (u)+∫ 1

0

ddt

(I (η))dt

= I (u)−∫ 1

0

⟨I′ (η) ,h(∥V (η)∥)V (η)

⟩dt

= I (u)+∫ 1

0−h(∥V (η)∥)

⟨I′ (η) ,V (η)

⟩dt

≤≤c+δ

I (u) +∫ 1

0

(−h(∥V (η)∥)

∥∥I′ (η)∥∥2)

dt

Then

c−δ +∫ 1

0h(∥V (η)∥)

∥∥I′ (η)∥∥2 dt < I (u)≤ c+δ

Thus

c−2δ +∫ 1

0h(∥V (η)∥)

∥∥I′ (η)∥∥2 dt < c

820 CHAPTER 24. CRITICAL POINTSwhere r > /(r) is a decreasing function which has values in (0, 1] and equals 1 for r € [0, 1]and equals 1/r for r > 1.Here u is given and the 77’ is the time derivative is with respect to t. By Corollary24.1.11, there exists a solution to this for t € [0, 1].Then for this solution, 7 (0,u) = u because of the above initial condition. If u €I-'(X\|e—€,c+6]), then u € A and so g(u) =0 so n(t,u) =u for all t € [0,1]. Thisgives the first two conditions. Consider the third.a(n (tu))) = ('(n).n')=-(U'(n),8 (ACV (n (tI) V (n (t,4)))= —g(u)h(\Vn(e#))I) 0), a u)))< ~g(w)h(|V(n(e,u))I)) |) Il,this last inequality from the inequalities of 24.1.5, and so this implies the third conditionsince it says that the function t > I (17 (t,u)) is decreasing.It remains to consider the last condition. This involves an appropriate choice of small6. It was chosen small and now it will be seen how small. It is desired to verify that(LL! ((-2,¢+6])) CI! ((-29,¢— 8])Suppose it is not so. Then there exists uw such that I(uw) € (c— 6,c+ 6] but J(n (1,u)) >c—6. We can assume that J(u) € (c—6,c+6] because if J(u) <c—6, then so is /(7 (1,u))from what was just shown. Hence g(u) = 1. Then using the fact that a (u )=1,ldc—8 <I (n(1u))=Tu)+ J 7 (UE (n))at0 atetre+ [melvin vn) a<7 5el (lv Cnyid (ny) ae1 ; 508+ f n(lv ml) |) [Par < tla) <e+8ThenThus 1c— 26+ | n(l\v (mpl) | (mI) ar <e