• Keine Ergebnisse gefunden

1.2 The Vlasov part

1.2.2 Solutions of the Vlasov part

∫

𝛾𝑇‒+

𝑓+π›Όπœ“π‘‘π›Ύπ›Όβˆ’

∫

π›Ύπ‘‡β€’βˆ’

𝒦𝛼𝑓+𝛼+𝑔𝛼

πœ“π‘‘π›Ύπ›Όβˆ’

∫

Ξ©

∫

R3

π‘“Λšπ›Όπœ“(0)𝑑𝑣𝑑π‘₯

for allπœ“βˆˆΞ¨π‘‡β€’.

We explain in the following remark in what sense we can speak of traces 𝑓𝛼 + of 𝑓𝛼

in a weak solution concept.

Remark 1.2.1. If Definition 1.1.1.(ii) is satisfied,𝑓+𝛼is the trace of 𝑓𝛼in the following sense:

β€’ As we have just seen, 𝑓+𝛼is the restriction of 𝑓𝛼to𝛾𝑇+β€’if 𝑓𝛼 ∈𝐢1 𝐼𝑇

‒×Ω×R3

.

β€’ There is no otherπ‘“Λœπ›Ό + ∈𝐿1

loc

𝛾𝑇+β€’

such that Definition 1.1.1.(ii) is satisfied as well, since for such π‘“Λœ+𝛼we have

∫

𝛾𝑇‒+

𝑓+π›Όβˆ’ Λœπ‘“+𝛼

πœ“π‘‘π›Ύπ›Ό =0

for allπœ“βˆˆπΆβˆž 𝐼𝑇

‒×Ω×R3

with suppπœ“βŠ‚ [0, 𝑇‒[ ×Ω×R3compact that vanish onπ›Ύπ‘‡βˆ’β€’βˆͺ𝛾𝑇0β€’. Consequently,π‘“Λœ+𝛼= 𝑓+𝛼.

1.2.2 Solutions of the Vlasov part

We give a brief introduction to the techniques and statements of Beals and Pro-topopescu [BP87], who used an approach via characteristics to tackle linear transport problems with initial-boundary conditions in a very general setting. Since we do not need the full statements of [BP87], we formulate those results in the way we will need them in our situation.

Throughout this subsection, let𝑇 >0,Ξ© βŠ‚R3 be an open, bounded set with𝐢1,πœ… -boundary for someπœ… >0, andΣ𝑇 B]0, 𝑇[ ×Ω×R3. Furthermore, letπ‘Œbe a first order linear differential operator that is divergence free and whose coefficients are Lipschitz continuous onΣ𝑇. In accordance to our situation, we choose

π‘ŒB πœ•π‘‘+

bπ‘£π›ΌΒ·πœ•π‘₯+πΉΒ·πœ•π‘£.

Thus, the assumptions aboutπ‘Œhere reduce to two conditions on𝐹, namely, that𝐹is Lipschitz continuous onΣ𝑇 and divergence free with respect to 𝑣. We additionally assume that𝐹is bounded onΣ𝑇. By Lipschitz continuity of𝐹, for each(𝑑, π‘₯, 𝑣) βˆˆΞ£π‘‡

there is a well-defined integral curve𝑠↦→ (𝑆, 𝑋 , 𝑉)(𝑠, 𝑑, π‘₯, 𝑣)satisfying 𝑑

𝑑𝑠𝑆=1, 𝑑

𝑑𝑠𝑋=𝑉b𝛼, 𝑑

𝑑𝑠𝑉=𝐹(𝑠, 𝑋 , 𝑉), (𝑆, 𝑋 , 𝑉)(𝑑, 𝑑, π‘₯, 𝑣)=(𝑑, π‘₯, 𝑣).

This curve is defined as long as it remains inΣ𝑇and there is a corresponding maximal time interval𝐼 βŠ‚Rfor which it is defined. We define the length of this curve to be the

length of the maximal time interval for which the curve remains inΣ𝑇, that is to say, the length equals𝑠+βˆ’π‘ βˆ’where

𝑠+Bsup{π‘ βˆˆπΌ | (𝑆, 𝑋 , 𝑉)(𝑠, 𝑑, π‘₯, 𝑣) βˆˆΞ£π‘‡}, π‘ βˆ’Binf{π‘ βˆˆπΌ | (𝑆, 𝑋 , 𝑉)(𝑠, 𝑑, π‘₯, 𝑣) βˆˆΞ£π‘‡}.

The next assumption is that there is a finite upper bound to all lengths of such integral curves. This condition is trivially satisfied in our caseΣ𝑇 =]0, 𝑇[ ×Ω×R3since𝑇is an upper bound. The last assumption is that each integral curve has a left and right limit point, i.e.,

𝑠→𝑠limβˆ’,𝑠>π‘ βˆ’(𝑆, 𝑋 , 𝑉)(𝑠, 𝑑, π‘₯, 𝑣),𝑠→𝑠lim+,𝑠

<𝑠+(𝑆, 𝑋 , 𝑉)(𝑠, 𝑑, π‘₯, 𝑣) βˆˆΞ£π‘‡.

These limits, if they exist, have to be elements ofπœ•Ξ£π‘‡. For their existence it is sufficient that 𝐹 is bounded by some constant 𝐢 > 0 since then both 𝑋€ and 𝑉€ are bounded because of

𝑋€

=

𝑉b𝛼

≀1,

𝑉€

=|𝐹(𝑠, 𝑋 , 𝑉)| ≀𝐢.

Accordingly, we define π·π‘‡βˆ’ (𝐷𝑇+) to be the subset of πœ•Ξ£π‘‡ consisting of all such left (right) limits, often referred to as incoming (outgoing) sets. These sets are Borel sets since π·π‘‡βˆ’ (𝐷𝑇+) is the image of the open set Σ𝑇 under the continuous function that maps a point inΣ𝑇to the left (right) limit point of the integral curve passing through this point. Note that possibly 𝐷𝑇± are not disjoint and/or do not exhaust πœ•Ξ£π‘‡ but both𝐷𝑇+βˆ©π·π‘‡βˆ’ andπœ•Ξ£π‘‡\ 𝐷𝑇+βˆͺπ·π‘‡βˆ’

are negligible in the sense that the union of all associated integral curves inΣ𝑇has Lebesgue measure zero.

We proceed with the definition of the test function space corresponding toπ‘Œ. Definition 1.2.2. LetΞ¦π‘Œπ‘‡be the space of all measurable functionsπœ™:Σ𝑇→Rwith the following three properties:

(i) πœ™is continuously differentiable along each integral curve.

(ii) πœ™andπ‘Œπœ™are bounded functions.

(iii) The support ofπœ™is bounded and there is a positive lower bound to the lengths of the integral curves which meet the support ofπœ™.

Remark 1.2.3. β€’ Here and in the following, the termπ‘Œ β„Ž, where β„Ž ∈ 𝐿1

loc(Σ𝑇), is in general to be understood as a distribution, i.e.,

(π‘Œ β„Ž) πœ‘

=βˆ’

∫

Σ𝑇

πœ•π‘‘πœ‘+

bπ‘£π›ΌΒ·πœ•π‘₯πœ‘+πΉΒ·πœ•π‘£πœ‘β„Ž 𝑑(𝑑, π‘₯, 𝑣), πœ‘βˆˆπΆβˆžπ‘ (Σ𝑇).

In Definition 1.2.2.(ii) or later in Definition 1.2.7.(i), this distribution is assumed to be given by a function onΣ𝑇.

β€’ Because of Definition 1.2.2.(ii) and 1.2.2.(iii) we haveπœ™, π‘Œπœ™βˆˆπΏπ‘ž(Σ𝑇)for anyπœ™βˆˆΞ¦π‘Œπ‘‡ and 1≀ π‘žβ‰€ ∞.

1.2 The Vlasov part 17

β€’ Note that a functionπœ™βˆˆ Ξ¦π‘Œπ‘‡ only has to be continuously differentiable along each integral curve but may be discontinuous in other directions. Because of Defini-tion 1.2.2.(i) and 1.2.2.(ii) every πœ™ ∈ Ξ¦π‘Œπ‘‡ can be extended to be continuous at the endpoints of each integral curve.

SinceΞ¦π‘Œπ‘‡depends on𝐹, it cannot be suitable for the whole nonlinear system (VM), where𝐹is unknown. Thus, an important (technical) statement is that our test function spaceΨ𝑇‒, which is independent of𝐹, belongs toΞ¦π‘Œπ‘‡after a cut-off in the time variable (if𝑇≀𝑇‒). This is verified in the following two lemmas, where we follow the proof of [Guo93, Lemma 2.1.].

Lemma 1.2.4. (i) For any πœ„ > 0 there is a 𝛿 = 𝛿(πœ„) > 0 such that for all (π‘₯, 𝑣) ∈ Λœπ›Ύβˆ’ satisfyingdist (π‘₯, 𝑣),π›ΎΛœ0

> πœ„we haveb𝑣𝛼·𝑛(π‘₯) ≀ βˆ’π›Ώ.

(ii) For anyπœ„ >0there is anπœ‚=πœ‚(πœ„)>0such that for anyπ‘₯ βˆˆπœ•Ξ©,𝑦 ∈R3we haveπ‘¦βˆˆΞ© if

π‘¦βˆ’π‘₯

< πœ‚and π‘¦βˆ’π‘₯

·𝑛(π‘₯) ≀ βˆ’ πœ„

π‘¦βˆ’π‘₯ <0.

Proof. As for part 1.2.4.(i), suppose the contrary. Then we can find a πœ„ > 0 and a sequence(π‘₯π‘˜, π‘£π‘˜) βŠ‚ Λœπ›Ύβˆ’ with dist (π‘₯π‘˜, π‘£π‘˜),π›ΎΛœ0 > πœ„

forπ‘˜ ∈ Nandbπ‘£π‘˜,𝛼 ·𝑛(π‘₯π‘˜) β†’ 0 for π‘˜β†’ ∞. Without loss of generality we can assume that(π‘£π‘˜)is bounded: If|π‘£π‘˜| β‰₯1 let π‘€π‘˜ B |π‘£π‘£π‘˜π‘˜|. Then,

0>𝑀bπ‘˜,𝛼·𝑛(π‘₯π‘˜)=|

𝑀bπ‘˜,𝛼|cos(](

𝑀bπ‘˜,𝛼, 𝑛(π‘₯π‘˜))) β‰₯ |

bπ‘£π‘˜,𝛼|cos(](

bπ‘£π‘˜,𝛼, 𝑛(π‘₯π‘˜)))

=bπ‘£π‘˜,𝛼·𝑛(π‘₯π‘˜) β†’0 forπ‘˜β†’ ∞because of|

𝑀bπ‘˜,𝛼| ≀ | bπ‘£π‘˜,𝛼|.

Therefore,(π‘₯π‘˜, π‘£π‘˜) βŠ‚πœ•Ξ©Γ—R3converges, after extracting a suitable subsequence, to some(π‘₯, 𝑣) βˆˆπœ•Ξ©Γ—R3. On the one hand, we have dist (π‘₯, 𝑣),π›ΎΛœ0β‰₯πœ„, and on the other handb𝑣𝛼·𝑛(π‘₯)=0 which is a contradiction.

The proof of part 1.2.4.(ii) exploits that πœ•Ξ© is of class 𝐢1,πœ…. Suppose that the assertion does not hold, i.e., we can find aπœ„ >0 and sequences(π‘₯π‘˜) βŠ‚ πœ•Ξ©, π‘¦π‘˜ βŠ‚

R3 with

π‘¦π‘˜βˆ’π‘₯π‘˜

< 1π‘˜ and π‘¦π‘˜βˆ’π‘₯π‘˜Β·π‘›(π‘₯

π‘˜) ≀ βˆ’πœ„ π‘¦π‘˜βˆ’π‘₯π‘˜

<0 butπ‘¦π‘˜βˆ‰ Ξ©. We may assume that both sequences converge because of(π‘₯π‘˜) βŠ‚πœ•Ξ©and π‘¦π‘˜βŠ‚

πœ•Ξ©+𝐡1. The limits of both sequences have to be the same; we call the limitπ‘₯ βˆˆπœ•Ξ©. Sinceπ‘₯π‘˜+𝑑 π‘¦π‘˜βˆ’π‘₯π‘˜βˆˆ

Ξ© for𝑑 >0 small enough andπ‘¦π‘˜ βˆ‰ Ξ©, there has to be aπ‘₯Λœπ‘˜ ∈ π‘₯

π‘˜, π‘¦π‘˜

βˆ©πœ•Ξ©. Obviously we have| ˜π‘₯π‘˜βˆ’π‘₯π‘˜| < 1π‘˜ and

( ˜π‘₯π‘˜βˆ’π‘₯π‘˜) ·𝑛(π‘₯π‘˜)= π‘¦π‘˜βˆ’π‘₯π‘˜Β·π‘›(π‘₯

π‘˜)| ˜π‘₯π‘˜βˆ’π‘₯π‘˜| π‘¦π‘˜βˆ’π‘₯π‘˜

≀ βˆ’πœ„| ˜π‘₯π‘˜βˆ’π‘₯π‘˜|<0. (1.2.5) Sinceπœ•Ξ©is compact andπ‘₯Λœπ‘˜ β†’ π‘₯,π‘₯π‘˜ β†’ π‘₯forπ‘˜ β†’ ∞,π‘₯Λœπ‘˜,π‘₯π‘˜, andπ‘₯ lie in the image of the same 𝐢1,πœ…-chart πœ‘:R2 βŠƒ π‘Š β†’ πœ•Ξ© if π‘˜is large enough. Let π‘Λœπ‘˜ B πœ‘βˆ’1( ˜π‘₯π‘˜), π‘π‘˜ B πœ‘βˆ’1(π‘₯π‘˜), and𝑝 B πœ‘βˆ’1(π‘₯). By continuity ofπœ‘βˆ’1, both π‘Λœπ‘˜

and π‘π‘˜

converge to 𝑝. Thus, we may assume thatπ‘Λœπ‘˜, π‘π‘˜βˆˆπΎπ‘ Bπ΅π‘Ÿ 𝑝 βŠ‚π‘Š

for suitableπ‘Ÿ >0 and largeπ‘˜. We expand the left-hand side of (1.2.5) to get the estimate

|( ˜π‘₯π‘˜βˆ’π‘₯π‘˜) ·𝑛(π‘₯π‘˜)|

=

. Together with (1.2.5) and the fact that πœ‘βˆ’1 is Lipschitz continuous on πœ‘ 𝐾𝑝

with some Lipschitz constant πΏπœ‘,𝐾

𝑝 >0β€”see proof belowβ€”, this yields for largeπ‘˜ 0< πœ„

But this contradicts π‘Λœπ‘˜βˆ’π‘π‘˜

β†’0 forπ‘˜β†’ ∞.

So there remains to show the Lipschitz continuity ofπœ‘βˆ’1onπœ‘ 𝐾𝑝

. This relies on the fact that, sinceπœ‘is a chart, the function

𝐺:πΎπ‘Γ—πœ•π΅1 β†’R, 𝐺 𝑝,˜ 𝛿𝑝

is continuous and positive so that it is bounded from below by some positive constant 𝑐>0. Forπ‘₯, π‘₯˜ βˆˆπœ‘ 𝐾𝑝

also converge to the same limit due to the continuity ofπœ‘βˆ’1. But this contradicts (1.2.6). Hence,πœ‘βˆ’1is Lipschitz continuous on πœ‘ 𝐾𝑝

and the proof is finished.

Lemma 1.2.5. For eachπœ“ βˆˆΞ¨π‘‡we haveπœ“

1.2 The Vlasov part 19 Sinceπœ“

Ξ£

𝑇 obviously satisfies Definition 1.2.2.(i) and 1.2.2.(ii), we only have to take care about Definition 1.2.2.(iii). First note that, since the support ofπœ“ is compact in [0, 𝑇[ ×Ω×R3, there is a 0 ≀ 𝑠0 < 𝑇such that πœ“(𝑑, π‘₯, 𝑣)= 0 for𝑠0 ≀ 𝑑 <𝑇,π‘₯ ∈ Ξ©, π‘£βˆˆR3.

We consider an integral curve which meets suppπœ“ = suppπœ“

Σ𝑇. This curve can be written as𝑠↦→ (𝑆, 𝑋 , 𝑉)(𝑠, π‘ βˆ’, π‘₯, 𝑣)and remains inΣ𝑇for a maximal time interval ]π‘ βˆ’, 𝑠+[ βŠ‚ ]0, 𝑇[so that(π‘ βˆ’, π‘₯, 𝑣) βˆˆπ·π‘‡βˆ’. Obviously it holds thatπ‘ βˆ’ ≀ 𝑠0. We have to find a positive lower bound for 𝑠+βˆ’π‘ βˆ’that does not depend onπ‘ βˆ’,𝑠+,π‘₯, and𝑣. In the following, let𝑠 ∈ ]π‘ βˆ’, 𝑠+[.

Case 1. If

dist (π‘ βˆ’, π‘₯, 𝑣),suppπœ“β‰₯ 𝑑0

2

we can find an 𝑠 such that(𝑆, 𝑋 , 𝑉)(𝑠, π‘ βˆ’, π‘₯, 𝑣) ∈ suppπœ“ since the curve meets the support ofπœ“. By

𝑋€

≀1 and 𝑉€

≀supΞ£

𝑇|𝐹|we have 𝑑0

2 ≀dist (π‘ βˆ’, π‘₯, 𝑣),suppπœ“ ≀ |(𝑆, 𝑋 , 𝑉)(𝑠, π‘ βˆ’, π‘₯, 𝑣) βˆ’ (π‘ βˆ’, π‘₯, 𝑣)|

≀ q

2+ k𝐹k2∞(π‘ βˆ’π‘ βˆ’) so that 𝑑0

2

√

2+k𝐹k2∞ is such a desired lower bound in this case.

Case 2. The more complicated case is

dist (π‘ βˆ’, π‘₯, 𝑣),suppπœ“

< 𝑑0 2. Since{𝑇} ×Ω×R3and𝛾𝑇+do not intersectπ·π‘‡βˆ’, we have

π·π‘‡βˆ’βŠ‚π›Ύπ‘‡βˆ’βˆͺ𝛾𝑇0 βˆͺ

{0} ×Ω×R3 .

Clearly, it holds that(π‘ βˆ’, π‘₯, 𝑣)βˆ‰π›Ύπ‘‡0 because of dist (π‘ βˆ’, π‘₯, 𝑣),suppπœ“

< 𝑑0

2 <𝑑0 ≀dist suppπœ“,𝛾𝑇0

. If(π‘ βˆ’, π‘₯, 𝑣) ∈ {0} ×Ω×R3we have

dist(π‘₯,πœ•Ξ©)=dist (π‘ βˆ’, π‘₯, 𝑣),{0} Γ—πœ•Ξ©Γ—R3

β‰₯dist suppπœ“,{0} Γ—πœ•Ξ©Γ—R3βˆ’

dist (π‘ βˆ’, π‘₯, 𝑣),suppπœ“

>𝑑0βˆ’π‘‘0 2 = 𝑑0

2 . Thus,𝑋(𝑠, π‘ βˆ’, π‘₯, 𝑣) ∈Ωfor 0≀𝑠<minn𝑑

0 2 , 𝑇o

again because of 𝑋€

≀1. Therefore, a positive lower bound to the length of the integral curve in this case is minn𝑑

0 2 , 𝑇o

.

Finally, suppose(π‘ βˆ’, π‘₯, 𝑣) βˆˆπ›Ύπ‘‡βˆ’. First note that dist (π‘₯, 𝑣),π›ΎΛœ0

=dist (π‘ βˆ’, π‘₯, 𝑣),𝛾𝑇0

β‰₯

dist suppπœ“,𝛾𝑇0

βˆ’

dist (π‘ βˆ’, π‘₯, 𝑣),suppπœ“

>𝑑0βˆ’ 𝑑0 2 = 𝑑0

2 . (1.2.7)

Let𝛿=𝛿𝑑

0 2

andπœ‚=πœ‚ 𝛿2

according to Lemma 1.2.4. We claim that

π‘šBmin (

π‘‡βˆ’π‘ 0,πœ‚

2, 𝛿

9

2k𝐹k𝐿∞(Σ𝑇;R3)+1 )

is such a positive lower bound (to the length of the integral curve) we search for.

Indeed, we firstly have[π‘ βˆ’, π‘ βˆ’+π‘š] βŠ‚ [0, 𝑇]due toπ‘ βˆ’ ≀𝑠0. Secondly, let 𝑠Bsup{π‘ βˆˆ ]π‘ βˆ’, π‘ βˆ’+π‘š] |𝑋(Λœπ‘ , π‘ βˆ’, π‘₯, 𝑣) ∈Ωfor allπ‘ Λœβˆˆ ]π‘ βˆ’, 𝑠[}.

Because of

|𝑋(𝑠, π‘ βˆ’, π‘₯, 𝑣) βˆ’π‘₯| β‰€π‘ βˆ’π‘ βˆ’ < πœ‚ and

(𝑋(𝑠, π‘ βˆ’, π‘₯, 𝑣) βˆ’π‘₯) ·𝑛(π‘₯)=

∫ 𝑠

π‘ βˆ’

𝑉b𝛼(𝜏, π‘ βˆ’, π‘₯, 𝑣)π‘‘πœ

!

·𝑛(π‘₯)

=(π‘ βˆ’π‘ βˆ’)

b𝑣𝛼·𝑛(π‘₯) +

∫ 𝑠

π‘ βˆ’

∫ 𝜏 π‘ βˆ’

𝑑 𝑑𝑠𝑉b𝛼

(𝑙, π‘ βˆ’, π‘₯, 𝑣)π‘‘π‘™π‘‘πœΒ·π‘›(π‘₯)

≀ βˆ’π›Ώ(π‘ βˆ’π‘ βˆ’) +9

2k𝐹k𝐿∞(Σ𝑇;R3)Β·1

2(π‘ βˆ’π‘ βˆ’)2 ≀ βˆ’π›Ώ

2(π‘ βˆ’π‘ βˆ’) ≀ βˆ’π›Ώ

2|𝑋(𝑠, π‘ βˆ’, π‘₯, 𝑣) βˆ’π‘₯| (which also implies𝑋(𝑠, π‘ βˆ’, π‘₯, 𝑣)β‰  π‘₯ sinceβˆ’π›Ώ

2(π‘ βˆ’π‘ βˆ’) <0) by (1.2.7) and

𝑑b𝑣𝛼,𝑖 𝑑𝑣𝑗

≀

3 2, 𝑖, 𝑗 = 1,2,3, we have 𝑋(𝑠, π‘ βˆ’, π‘₯, 𝑣) ∈ Ξ© and thus 𝑠 = π‘ βˆ’+π‘š. This completes the proof.

We should remark that the three conditions onπœ“βˆˆΞ¨π‘‡in (1.1.1) are really necessary:

Letπœ„ >0 be small and, for simplicity, take𝐹 =0. Firstly, if we allow a test functionπœ“ that does not vanish before time𝑇and has support onπ›Ύπ‘‡βˆ’, we can find an integral curve enteringΣ𝑇onπ›Ύπ‘‡βˆ’βˆ©suppπœ“at timeπ‘ βˆ’ =π‘‡βˆ’πœ„. Secondly, if we allow a test function πœ“ with support on𝛾𝑇0, then for some(𝑑, π‘₯, 𝑣) ∈ 𝛾𝑇0β€”such that in a neighborhood of π‘₯ there are no common points of Ξ© and the tangent space of πœ•Ξ© at π‘₯β€”the curves (𝑆, 𝑋 , 𝑉)(𝑠) = (𝑠, π‘₯βˆ’πœ„π‘›(π‘₯) + (π‘ βˆ’π‘‘)

b𝑣𝛼, 𝑣), defined for all 𝑠 ∈ [0, 𝑇], will meet the support ofπœ“. Thirdly, if we allow a test functionπœ“ with support on{0} Γ—πœ•Ξ©Γ—R3 we can find an integral curve meeting the support ofπœ“, (its𝑋-coordinate) starting at time 0 nearπœ•Ξ©and leavingΞ©at timeπœ„. In all three cases, there will be no positive lower bound to the length of these curves.

1.2 The Vlasov part 21 Conversely, these restrictions cause no problems for later considerations. Firstly, we do not want to test a solution of (1.2.1) at time𝑇since we are interested in an initial, and not final, value problem. Secondly, we only want to impose a boundary condition on π›Ύβˆ’ and not on 𝛾0. Thirdly, proper initial data of the distribution function have to satisfy the boundary condition at time 0 a priori so that this property need not be tested, and{0} Γ—πœ•Ξ©Γ—R3is even a null set with respect to𝑑𝛾𝛼.

We now proceed with some important results of [BP87]. There, the main idea is to use the β€œidentifications”

Σ𝑇 β‰ˆ(𝑠, 𝑧) |π‘§βˆˆπ·βˆ’

𝑇,0<𝑠 <𝑙(𝑧) , π‘Œβ‰ˆ 𝑑 𝑑𝑠,

where𝑙(𝑧)is the length of the integral curve corresponding to𝑧. The first important result is the following property which is closely related to Green’s identity; see [BP87, Proposition 7].

Proposition 1.2.6. There are two unique Borel measures𝜈±on𝐷±

𝑇such that

∫

Σ𝑇

π‘Œπœ™π‘‘(𝑑, π‘₯, 𝑣)=

∫

𝐷+π‘‡πœ™π‘‘πœˆ+βˆ’

∫

π·π‘‡βˆ’πœ™π‘‘πœˆβˆ’ for allπœ™βˆˆΞ¦π‘Œπ‘‡.

We have to define the space of functions in which we search for solutions of some initial-boundary problem.

Definition 1.2.7. For 1 ≀ 𝑝 < ∞let𝐸𝑝(Σ𝑇;π‘Œ)be the space of functions 𝑓 ∈ 𝐿𝑝(Σ𝑇) with the following two properties:

(i) π‘Œ 𝑓 βˆˆπΏπ‘(Σ𝑇).

(ii) There is a trace of 𝑓 on𝐷±

𝑇, i.e., a pair of functions 𝑓± βˆˆπΏπ‘ 𝐷±

𝑇, π‘‘πœˆΒ±

satisfying the extended Green’s identity

∫

Σ𝑇

πœ™π‘Œ 𝑓 +𝑓 π‘Œπœ™π‘‘(𝑑, π‘₯, 𝑣)

=

∫

𝐷+ 𝑇

𝑓+πœ™π‘‘πœˆ+βˆ’

∫

π·βˆ’ 𝑇

π‘“βˆ’πœ™π‘‘πœˆβˆ’

for allπœ™βˆˆΞ¦π‘Œπ‘‡.

Note that a trace in the sense as stated above is unique and that all terms are well-defined according to Remark 1.2.3.

Lemma 1.2.8. Let 1 ≀ 𝑝 < ∞, 𝑓 ∈ 𝐸𝑝(Σ𝑇;π‘Œ) and 𝑀 ∈ 𝐢∞(Σ𝑇) ∩𝐢1𝑏 Σ𝑇

. Then, 𝑀 𝑓 βˆˆπΈπ‘(Σ𝑇;π‘Œ)and 𝑀 𝑓±

=𝑀 𝑓±. Proof. Because of

π‘Œ 𝑀 𝑓 πœ‘

=βˆ’

∫

Σ𝑇

πœ•π‘‘πœ‘+

bπ‘£π›ΌΒ·πœ•π‘₯πœ‘+πΉΒ·πœ•π‘£πœ‘π‘€ 𝑓 𝑑(𝑑, π‘₯, 𝑣)

=βˆ’

∫

Σ𝑇

πœ•π‘‘ π‘€πœ‘+

bπ‘£π›ΌΒ·πœ•π‘₯ π‘€πœ‘+𝐹·

πœ•π‘£ π‘€πœ‘ 𝑓 𝑑(𝑑, π‘₯, 𝑣)

+

∫

Σ𝑇

(πœ•π‘‘π‘€+

bπ‘£π›ΌΒ·πœ•π‘₯𝑀+πΉΒ·πœ•π‘£π‘€)π‘“πœ‘π‘‘(𝑑, π‘₯, 𝑣)

=

∫

Σ𝑇

π‘Œ 𝑓𝑀

πœ‘π‘‘(𝑑, π‘₯, 𝑣) +

∫

Σ𝑇

(π‘Œπ‘€)π‘“πœ‘π‘‘(𝑑, π‘₯, 𝑣)

for any πœ‘ ∈ πΆβˆžπ‘ (Σ𝑇), it holds thatπ‘Œ 𝑀 𝑓

= π‘€π‘Œ 𝑓 + 𝑓 π‘Œπ‘€ ∈ 𝐿𝑝(Σ𝑇). Now letπœ™ ∈ Ξ¦π‘Œπ‘‡. We haveπ‘€πœ™ ∈ Ξ¦π‘Œπ‘‡ since Definition 1.2.2.(i) and 1.2.2.(ii) are satisfied because of the regularity of𝑀and Definition 1.2.2.(iii) is satisfied because of supp π‘€πœ™ βŠ‚

suppπœ™.

Thus, it holds that

∫

Σ𝑇

πœ™π‘Œ 𝑀 𝑓+𝑀 𝑓 π‘Œ

πœ™π‘‘(𝑑, π‘₯, 𝑣)

=

∫

Σ𝑇

π‘€πœ™π‘Œ 𝑓 +𝑓 π‘Œ π‘€πœ™ 𝑑(𝑑, π‘₯, 𝑣)

=

∫

𝐷+ 𝑇

𝑓+π‘€πœ™π‘‘πœˆ+βˆ’

∫

π·βˆ’ 𝑇

π‘“βˆ’π‘€πœ™π‘‘πœˆβˆ’,

which proves the assertion.

In the following it is convenient to split𝐷𝑇±as follows:

𝐷±𝑇B (𝑑, π‘₯, 𝑣) βˆˆπ·π‘‡Β±|0<𝑑<𝑇 ,

𝐷0 B (𝑑, π‘₯, 𝑣) βˆˆπ·βˆ’π‘‡ |𝑑=0 , 𝐷𝑇𝑇 B (𝑑, π‘₯, 𝑣) βˆˆπ·π‘‡+ |𝑑=𝑇 , so thatπ·βˆ’

𝑇 =π·βˆ’π‘‡βˆͺ𝐷0and𝐷+

𝑇 =𝐷+𝑇βˆͺ𝐷𝑇

𝑇. Note that𝐷0does not depend on𝑇(in the sense that any 0<π‘‡Λœ <𝑇yields the same set𝐷0). According to this decomposition we writeπ‘‘πœˆβˆ’=π‘‘πœˆβˆ’

𝐷𝑇

βˆ’,π‘‘πœˆ0=π‘‘πœˆβˆ’ 𝐷

0,π‘‘πœˆ+=π‘‘πœˆ+ 𝐷𝑇

+,π‘‘πœˆπ‘‡=π‘‘πœˆ+ 𝐷𝑇

𝑇, π‘“βˆ’= π‘“βˆ’ 𝐷𝑇

βˆ’, 𝑓0= π‘“βˆ’ 𝐷

0, 𝑓+= 𝑓+

𝐷𝑇

+, and 𝑓𝑇= 𝑓+ 𝐷𝑇

𝑇. We have

{0} ×Ω×R3 βŠ‚π·0 βŠ‚ {0} ×Ω×R3, {𝑇} ×Ω×R3βŠ‚π·π‘‡π‘‡βŠ‚ {𝑇} ×Ω×R3, π›Ύπ‘‡βˆ’βŠ‚π·π‘‡βˆ’ βŠ‚π›Ύπ‘‡βˆ’βˆͺ𝛾𝑇0, 𝛾𝑇+ βŠ‚π·+π‘‡βŠ‚π›Ύπ‘‡+βˆͺ𝛾𝑇0.

Therefore, we can identify 𝐿𝑝-functions on 𝐷0 (or𝐷𝑇𝑇) with𝐿𝑝-functions onΩ×R3 since

Ω×R3

\ Ω×R3

has(π‘₯, 𝑣)-Lebesgue measure zero. Additionally, we may write 𝑓(0)and 𝑓(𝑇)instead of 𝑓0and 𝑓𝑇pointing out that we may evaluate 𝑓 at time 0 and𝑇in some sense.

For eachπœ“βˆˆΞ¨π‘‡we have

∫

Σ𝑇

π‘Œπœ“π‘‘(𝑑, π‘₯, 𝑣)=βˆ’

∫

Ξ©

∫

R3

πœ“(0)𝑑𝑣𝑑π‘₯+

∫ 𝑇

0

∫

πœ•Ξ©

∫

R3

πœ“b𝑣𝛼·𝑛 𝑑𝑣𝑑𝑆π‘₯𝑑𝑑

=βˆ’

∫

Ξ©

∫

R3

πœ“(0)𝑑𝑣𝑑π‘₯+

∫

𝛾𝑇+

πœ“π‘‘π›Ύπ›Όβˆ’

∫

π›Ύβˆ’π‘‡

πœ“π‘‘π›Ύπ›Ό.

1.2 The Vlasov part 23 This shows thatπ‘‘πœˆ0 =𝑑(π‘₯, 𝑣)on𝐷0andπ‘‘πœˆΒ± =𝑑𝛾𝛼on𝛾𝑇±. With an analog reasoning (consider test functions πœ“Λœ(𝑑, π‘₯, 𝑣) = πœ“(π‘‡βˆ’π‘‘, π‘₯, 𝑣),πœ“ ∈ Ψ𝑇) we conclude that π‘‘πœˆπ‘‡ = 𝑑(π‘₯, 𝑣)on𝐷𝑇𝑇as well.

We proceed with a definition of some properties of operators.

Definition 1.2.9. Let 𝑂 be an operator between two function spaces on subsets of someR𝑛, whose first component we call time. 𝑂is called

(i) local in time if𝑂(𝑒𝑣)=𝑒𝑂(𝑣)for all continuous functions𝑒that only depend on time and all possible𝑣;

(ii) nonnegative if𝑂(𝑣) β‰₯0 for all𝑣 β‰₯0.

Now we are ready to state the following result regarding the unique solvability of linear transport problems with initial-boundary conditions; see [BP87, Proposition 1, Theorems 1 and 2].

Proposition 1.2.10. Let1 ≀ 𝑝 < ∞, β„Ž ∈ 𝐿∞(Σ𝑇),𝐹:Σ𝑇 β†’ R be Lipschitz continuous, differentiable with respect to𝑣, and divergence free with respect to𝑣, andπ‘Œ=πœ•π‘‘+

bπ‘£π›ΌΒ·πœ•π‘₯+πΉΒ·πœ•π‘£. (i) For all 𝑓 βˆˆπΈπ‘(Σ𝑇;π‘Œ)we have

∫

𝐷𝑇 𝑇

𝑓𝑇

π‘π‘‘πœˆπ‘‡+

∫

𝐷𝑇 +

𝑓+

π‘π‘‘πœˆ++𝑝

∫

Σ𝑇

β„Ž 𝑓

𝑝𝑑(𝑑, π‘₯, 𝑣)

=

∫

𝐷0

𝑓0

π‘π‘‘πœˆ0+

∫

π·π‘‡βˆ’

π‘“βˆ’

π‘π‘‘πœˆβˆ’+𝑝

∫

Σ𝑇

sign 𝑓 𝑓

π‘βˆ’1(π‘Œ+β„Ž)𝑓 𝑑(𝑑, π‘₯, 𝑣). (1.2.8)

(ii) Let moreoverπ”Ž: 𝐿𝑝 𝐷+𝑇, π‘‘πœˆ+ β†’ 𝐿𝑝 π·π‘‡βˆ’, π‘‘πœˆβˆ’

be a bounded linear operator, that is local in time and has operator norm less than1, and𝑔0 βˆˆπΏπ‘(𝐷0),π‘”βˆ’βˆˆ 𝐿𝑝 π·π‘‡βˆ’, π‘‘πœˆβˆ’

. Then the problem

π‘Œ 𝑓 =0 onΣ𝑇, (1.2.9a)

𝑓0 =𝑔0 on𝐷0, (1.2.9b)

π‘“βˆ’ = π”Žπ‘“++π‘”βˆ’ onπ·π‘‡βˆ’ (1.2.9c)

has a unique solution𝑓 βˆˆπΈπ‘(Σ𝑇;π‘Œ). Here,(1.2.9a)holds pointwise almost everywhere (cf. Definition 1.2.7.(i) and Remark 1.2.3), and(1.2.9b)and (1.2.9c)hold pointwise almost everywhere (with respect to the corresponding measures) in the sense of trace (cf. Definition 1.2.7.(ii)). Moreover, the solution is nonnegative ifπ”Ž, 𝑔0, and π‘”βˆ’ are nonnegative.

Here and in the following, for functions the property β€œnonnegative” usually means

β€œnonnegative almost everywhere”. We want to express, in some way, the theorem above in words that fit to our problem (1.2.1), that is to say, we should somehow replace 𝐷0,π·π‘‡βˆ’ (and so on) byΩ×R3,π›Ύπ‘‡βˆ’(and so on). Moreover, we search for solutions of (1.2.1) on𝐼𝑇

β€’ instead of solutions on some time interval [0, 𝑇]. To this end, we first have to define what we call a strong solution of (1.2.1). From now on, the force term 𝐹shall satisfy the following condition.

Condition 1.2.11. 𝐹:𝐼𝑇

β€’ ×Ω×R3 β†’R3 is Lipschitz continuous and bounded onΣ𝑇

for any𝑇 ∈ 𝐼𝑇‒, and moreover differentiable and divergence free (both) with respect to𝑣.

Definition 1.2.12. Assume that 1 ≀ 𝑝 < ∞,𝒦: 𝐿𝑝

lt

𝛾𝑇+β€’, 𝑑𝛾𝛼

β†’ 𝐿𝑝

lt

π›Ύπ‘‡βˆ’β€’, 𝑑𝛾𝛼

, π‘“Λš ∈ 𝐿𝑝 Ω×R3

, and 𝑔 ∈ 𝐿𝑝

lt

π›Ύπ‘‡βˆ’β€’

, 𝑑𝛾𝛼

. We call a function 𝑓:𝐼𝑇

β€’ ×Ω×R3 β†’Ra strong solution of (1.2.1) if:

(i) πœ’π‘‡π‘“ βˆˆπΈπ‘(Σ𝑇;π‘Œ)for all 0<π‘‡βˆˆπΌπ‘‡

β€’. (ii) For allπœ“ βˆˆΞ¨π‘‡β€’it holds that

∫ 𝑇‒ 0

∫

Ξ©

∫

R3

πœ•π‘‘πœ“+

bπ‘£π›ΌΒ·πœ•π‘₯πœ“+πΉΒ·πœ•π‘£πœ“π‘“ 𝑑𝑣𝑑π‘₯𝑑𝑑

=

∫

𝛾+𝑇‒

𝑓+πœ“π‘‘π›Ύπ›Όβˆ’

∫

π›Ύβˆ’π‘‡β€’

𝒦𝑓++𝑔

πœ“π‘‘π›Ύπ›Όβˆ’

∫

Ξ©

∫

R3

π‘“Λšπœ“(0)𝑑𝑣𝑑π‘₯.

Note that, for each 0<𝑇 βˆˆπΌπ‘‡

β€’, at first only a trace ofπœ’π‘‡π‘“ is defined. By uniqueness, for another 𝐼𝑇

β€’ 3 𝑇0 β‰₯ 𝑇, the traces of πœ’π‘‡0𝑓 and πœ’π‘‡π‘“ coincide on the common time interval[0, 𝑇]. Thus, we may write 𝑓±, which is defined on all of𝐼𝑇

β€’, and may drop the dependence on some𝑇.

Proposition 1.2.13. Let1 ≀ 𝑝 <∞,𝐹 satisfy Condition 1.2.11, and𝒦:𝐿𝑝

lt

𝛾𝑇+β€’, 𝑑𝛾𝛼

β†’ 𝐿𝑝

lt

π›Ύπ‘‡βˆ’β€’, 𝑑𝛾𝛼

be a linear operator, that is local in time and such that there is a0 ≀ π‘˜0 < 1 satisfying

k𝒦𝑐k𝐿𝑝(π›Ύπ‘‡βˆ’,𝑑𝛾𝛼) β‰€π‘˜0k𝑐k𝐿𝑝(𝛾𝑇+,𝑑𝛾𝛼) for all 𝑐 ∈ 𝐿𝑝

lt

𝛾𝑇+β€’, 𝑑𝛾𝛼

, 0 < 𝑇 ∈ 𝐼𝑇

β€’. Furthermore, let π‘“Λš ∈ 𝐿𝑝 Ω×R3

and 𝑔 ∈ 𝐿𝑝

lt

π›Ύπ‘‡βˆ’β€’, 𝑑𝛾𝛼

. Then:

(i) There is exactly one strong solution of (1.2.1)in the sense of Definition 1.2.12.

(ii) This solution is nonnegative if𝒦, π‘“Λš, and𝑔are nonnegative.

Proof. Let 0<π‘‡βˆˆπΌπ‘‡β€’and define

π‘”π‘‡βˆ’:π·π‘‡βˆ’ β†’R, π‘”π‘‡βˆ’(𝑑, π‘₯, 𝑣)=

(𝑔(𝑑, π‘₯, 𝑣), (𝑑, π‘₯, 𝑣) βˆˆπ›Ύπ‘‡βˆ’, 0, otherwise;

𝑔0:𝐷0β†’R, 𝑔0(0, π‘₯, 𝑣)= π‘“Λš(π‘₯, 𝑣).

Note that the latter definition makes sense since, as mentioned above,𝐷0 coincides with{0} ×Ω×R3up to a negligible set. We have

𝑔

𝑇

βˆ’

𝐿𝑝(π·π‘‡βˆ’,π‘‘πœˆβˆ’)= πœ’π‘‡π‘”

𝐿𝑝(π›Ύβˆ’,𝑑𝛾𝛼), 𝑔0

𝐿𝑝(𝐷0)=

π‘“Λš 𝐿𝑝(Ω×R3)

1.2 The Vlasov part 25 so that π‘”βˆ’π‘‡βˆˆπΏπ‘ π·π‘‡βˆ’, π‘‘πœˆΒ±

and𝑔0βˆˆπΏπ‘(𝐷0). Furthermore, forβ„Ž βˆˆπΏπ‘ 𝐷+𝑇, π‘‘πœˆ+ let β„Ž:𝛾𝑇+β€’ β†’R, β„Ž(𝑑, π‘₯, 𝑣)=

(β„Ž(𝑑, π‘₯, 𝑣), (𝑑, π‘₯, 𝑣) βˆˆπ›Ύπ‘‡+, 0, otherwise and

π”Žπ‘‡β„Ž:π·βˆ’π‘‡β†’R, (π”Žπ‘‡β„Ž)(𝑑, π‘₯, 𝑣)=

( π’¦β„Ž

(𝑑, π‘₯, 𝑣), (𝑑, π‘₯, 𝑣) βˆˆπ›Ύπ‘‡βˆ’, 0, otherwise. Because of

kπ”Žπ‘‡β„Žk𝐿𝑝(π·π‘‡βˆ’,π‘‘πœˆβˆ’)= πœ’π‘‡π’¦β„Ž

𝐿𝑝(π›Ύπ‘‡βˆ’,𝑑𝛾𝛼) β‰€π‘˜0 πœ’π‘‡β„Ž

𝐿𝑝(𝛾𝑇+,𝑑𝛾𝛼)=π‘˜0kβ„Žk𝐿𝑝(𝐷𝑇+,π‘‘πœˆ+) we conclude that π”Žπ‘‡ maps 𝐿𝑝 𝐷+𝑇, π‘‘πœˆ+

to 𝐿𝑝 π·βˆ’π‘‡, π‘‘πœˆβˆ’

and has operator norm less than 1. Moreover,π”Žπ‘‡is local in time. Thus, by Proposition 1.2.10.(ii) there is a solution of (1.2.9) (withπ”Žπ‘‡, 𝑔0, π‘”π‘‡βˆ’ given). By uniqueness andπ·βˆ’π‘‡ βŠ‚ π·βˆ’π‘‡0 for𝑇 ≀ 𝑇0, such a solution (and its trace) does not depend on𝑇, whence there is a function 𝑓 such that πœ’π‘‡π‘“ ∈ 𝐸𝑝(Σ𝑇;π‘Œ)is the unique solution of (1.2.9) for any given𝑇. Now takeπœ“ ∈ Ψ𝑇‒

and 0<𝑇 <𝑇‒such thatπœ“(𝑑, π‘₯, 𝑣)=0 if𝑑 β‰₯𝑇. By Lemma 1.2.5 we haveπœ“

Σ𝑇 βˆˆΞ¦π‘Œπ‘‡. Applying the definition of trace and using the properties ofπœ“, this leads to

∫

Σ𝑇‒

𝑓 π‘Œπœ“π‘‘(𝑑, π‘₯, 𝑣)=

∫

Σ𝑇

𝑓 π‘Œπœ“π‘‘(𝑑, π‘₯, 𝑣)

=βˆ’

∫

𝐷0

π‘“πœ“π‘‘πœˆ0+

∫

𝐷+𝑇

𝑓+πœ“π‘‘πœˆ+βˆ’

∫

π·βˆ’π‘‡

π”Žπ‘‡π‘“++π‘”π‘‡βˆ’πœ“π‘‘πœˆ

βˆ’

=βˆ’

∫

Ξ©

∫

R3

π‘“Λšπœ“(0)𝑑𝑣𝑑π‘₯+

∫

𝛾+𝑇

𝑓+πœ“π‘‘π›Ύπ›Όβˆ’

∫

π›Ύπ‘‡βˆ’

π”Žπ‘‡π‘“++π‘”βˆ’π‘‡ πœ“π‘‘π›Ύπ›Ό

=βˆ’

∫

Ξ©

∫

R3

π‘“Λšπœ“(0)𝑑𝑣𝑑π‘₯+

∫

𝛾+𝑇‒

𝑓+πœ“π‘‘π›Ύπ›Όβˆ’

∫

π›Ύπ‘‡β€’βˆ’

𝒦𝑓++π‘”πœ“π‘‘π›Ύ

𝛼

and the proof of part 1.2.13.(i) is complete.

Part 1.2.13.(ii) follows from the fact thatπ”Žπ‘‡,𝑔0, andπ‘”π‘‡βˆ’are nonnegative if𝒦,π‘“Λš, and 𝑔are nonnegative, and Proposition 1.2.10.(ii).

We now turn to the special situation that𝒦=π‘ŽπΎ, where𝐾is the reflection operator.

According to (0.2), 𝐾 𝑓 is defined for any function 𝑓 (that is defined on a subset of RΓ—πœ•Ξ©Γ—R3) and is self inverse, i.e., 𝐾2 = id. Its restriction to 𝐿𝑝

lt

𝛾𝑇+β€’

, 𝑑𝛾𝛼

yields an operator 𝐾: 𝐿𝑝

lt

𝛾𝑇+β€’, 𝑑𝛾𝛼

β†’ 𝐿𝑝

lt

π›Ύπ‘‡βˆ’β€’, 𝑑𝛾𝛼

for any 1 ≀ 𝑝 ≀ ∞. Its inverse is 𝐾:𝐿𝑝

lt

π›Ύπ‘‡βˆ’β€’, 𝑑𝛾𝛼

→𝐿𝑝

lt

𝛾𝑇+β€’, 𝑑𝛾𝛼

.

Lemma 1.2.14. For any𝑇 >0,1≀𝑝 <∞, and anyπ‘Ž, 𝑏 ∈ 𝐿∞ π›Ύπ‘‡βˆ’

,π‘βˆˆ 𝐿𝑝 𝛾𝑇+, 𝑑𝛾𝛼 , and β„Ž βˆˆπΏπ‘ π›Ύβˆ’π‘‡, 𝑑𝛾𝛼

we have

∫

π›Ύπ‘‡βˆ’

𝑏|π‘ŽπΎπ‘+β„Ž|𝑝𝑑𝛾𝛼=

∫

𝛾𝑇+

𝐾𝑏|π‘πΎπ‘Ž+𝐾 β„Ž|𝑝𝑑𝛾𝛼.

If additionallyπ‘Ž0B kπ‘Žk𝐿∞(π›Ύπ‘‡βˆ’)<1andπ‘Ž, 𝑏, 𝑐, β„Žare nonnegative, the estimate

∫

π›Ύπ‘‡βˆ’

𝑏|π‘ŽπΎπ‘+β„Ž|𝑝𝑑𝛾𝛼 ≀ π‘Ž0

∫

𝛾+𝑇

(𝐾𝑏)𝑐𝑝𝑑𝛾𝛼+ (1βˆ’π‘Ž0)1βˆ’π‘

∫

π›Ύπ‘‡βˆ’

𝑏 β„Žπ‘π‘‘π›Ύπ›Ό

holds.

Proof. We compute

∫

π›Ύπ‘‡βˆ’

𝑏|π‘ŽπΎπ‘+β„Ž|𝑝𝑑𝛾𝛼

=

∭

π›Ύπ‘‡βˆ’

𝑏(𝑑, π‘₯, 𝑣)|π‘Ž(𝑑, π‘₯, 𝑣)𝑐(𝑑, π‘₯, π‘£βˆ’2(𝑣·𝑛(π‘₯))𝑛(π‘₯)) +β„Ž(𝑑, π‘₯, 𝑣)|𝑝|

b𝑣𝛼·𝑛(π‘₯)|𝑑𝑣𝑑𝑆π‘₯𝑑𝑑

=

∭

𝛾𝑇+

𝑏(𝑑, π‘₯, π‘£βˆ’2(𝑣·𝑛(π‘₯))𝑛(π‘₯))|π‘Ž(𝑑, π‘₯, π‘£βˆ’2(𝑣·𝑛(π‘₯))𝑛(π‘₯))𝑐(𝑑, π‘₯, 𝑣) +β„Ž(𝑑, π‘₯, π‘£βˆ’2(𝑣·𝑛(π‘₯))𝑛(π‘₯))|𝑝|βˆ’

b𝑣𝛼·𝑛(π‘₯)|𝑑𝑣𝑑𝑆π‘₯𝑑𝑑

=

∫

𝛾𝑇+

𝐾𝑏|π‘πΎπ‘Ž+𝐾 β„Ž|𝑝𝑑𝛾𝛼.

using the change of variables𝑣 β†¦β†’π‘£βˆ’2(𝑣·𝑛(π‘₯))𝑛(π‘₯). Note that the determinant of the corresponding Jacobian equalsβˆ’1 since the map is a reflection. As for the second statement, we estimate

∫

π›Ύπ‘‡βˆ’

𝑏|π‘ŽπΎπ‘+β„Ž|𝑝𝑑𝛾𝛼=

∫

𝛾+𝑇

𝐾𝑏|π‘πΎπ‘Ž+𝐾 β„Ž|𝑝𝑑𝛾𝛼

≀

∫

𝛾+𝑇

𝐾𝑏

π‘Ž0𝑐+ (1βˆ’π‘Ž0)(1βˆ’π‘Ž0)

βˆ’1𝐾 β„Ž

𝑝

𝑑𝛾𝛼

β‰€π‘Ž0

∫

𝛾𝑇+

(𝐾𝑏)𝑐𝑝𝑑𝛾𝛼+ (1βˆ’π‘Ž0)

∫

𝛾𝑇+

𝐾𝑏

(1βˆ’π‘Ž0)βˆ’1𝐾 β„Ž

𝑝

𝑑𝛾𝛼

=π‘Ž0

∫

𝛾𝑇+

(𝐾𝑏)𝑐𝑝𝑑𝛾𝛼+ (1βˆ’π‘Ž0)1βˆ’π‘

∫

π›Ύπ‘‡βˆ’

𝑏 β„Žπ‘π‘‘π›Ύπ›Ό

using the convexity of the𝑝-th power and the first statement.

1.2 The Vlasov part 27 Then there is a unique, nonnegative strong solution 𝑓 ∈ 𝐿∞

lt 𝐼𝑇‒; 𝐿1∩𝐿∞

in the sense that the conditions of Definition 1.2.12 are satisfied for all1≀𝑝 <∞. Moreover, we have the estimates

(1βˆ’π‘Ž0)1𝑝 nonnegative𝐢1-functionπœƒ=πœƒ(𝑣)onR3that only depends on|𝑣|, is monotonically increasing in |𝑣|(i.e.,πœƒ(𝑣) =πœƒΛœ(|𝑣|)for some monotonically increasingπœƒΛœ ∈ 𝐢1(Rβ‰₯0)withπœƒΛœ0(0)=0), linear and local in time. We have

k𝒦𝑐k𝐿𝑝(π›Ύπ‘‡βˆ’,𝑑𝛾𝛼)=kπ‘πΎπ‘Žk𝐿𝑝(𝛾𝑇+,𝑑𝛾𝛼) β‰€π‘Ž0k𝑐k𝐿𝑝(𝛾𝑇+,𝑑𝛾𝛼)

for allπ‘βˆˆ 𝐿𝑝

lt

𝛾𝑇+β€’, 𝑑𝛾𝛼

and 0<𝑇 βˆˆπΌπ‘‡

β€’by Lemma 1.2.14. Thus, by Proposition 1.2.13 there is a unique solution for this 𝑝 in the sense of Definition 1.2.12. Since 𝑝 was arbitrary, it follows that, for all 1≀𝑝 <∞, 𝑓 βˆˆπΏπ‘

lt(Σ𝑇‒)and 𝑓± βˆˆπΏπ‘

lt

𝛾𝑇±‒, 𝑑𝛾𝛼

, and all conditions of Definition 1.2.12 are satisfied.

Let 0 < 𝑇 ∈ 𝐼𝑇

β€’ and recall that in the proof of Proposition 1.2.13 the solution on [0, 𝑇]was given by a solution of (1.2.9) withπ”Žπ‘‡,𝑔0,π‘”βˆ’π‘‡. Thus, π‘“βˆ’ = π”Žπ‘‡π‘“++π‘”π‘‡βˆ’ =0 on π·βˆ’π‘‡\π›Ύπ‘‡βˆ’. Applying Proposition 1.2.10.(i) (withβ„Ž =0), dropping negligible terms in (1.2.8), and using Lemma 1.2.14 we arrive at

∫

altogether and therefore π‘“βˆ’ ∈ 𝐿1 that it is nonnegative and only depends on|𝑣|since𝛽is nonnegative and only depends on|𝑣|. Proceeding similarly as before, we define π‘“Λœ(𝑑, π‘₯, 𝑣)Bπ›½πœ„(𝑣)𝑓(𝑑, π‘₯, 𝑣) β‰₯0,

1.2 The Vlasov part 29 Taking the limitπœ„β†’0 does not cause any problem because we have

π›½πœ„βˆ’π›½ finite. Hence, (1.2.14) holds withπœ„removed. Next we insert the definition of𝛽and drop the terms where|𝑣| β‰₯𝑅on the left-hand side to get

sinceπœƒis monotonically increasing in|𝑣|. Note that it is important thatβˆ‡π›½vanishes for |𝑣| > 𝑅. This proves (1.2.11) for 0 < 𝑅 < ∞. Because of βˆ‡πœƒ ∈ πΏπ‘ž R3;R3

is a bounded function.

As for (1.2.13), let 0<𝑅≀ ∞and first derive the following key estimate:

∫

where we setπ‘ŸB ∫

𝐡𝑅𝑣0𝛼𝑓(𝑇, π‘₯, 𝑣)𝑑𝑣14

∈ [0,∞[and used (1.2.10); ifπ‘Ÿ=0, the second step makes no sense but clearly both the left-hand side and the right-hand side of (1.2.15) are zero in this case. Note that the integral on the left-hand side exists for almost all(𝑇, π‘₯) βˆˆπΌπ‘‡

β€’ Γ—Ξ©by Fubini’s theorem and that the first estimate above holds trivially ifπ‘Ÿ >𝑅by 𝑓 β‰₯0 and is an equality ifπ‘Ÿβ‰€ 𝑅. Taking both sides of (1.2.15) to the power 43 and then integrating overΞ©yields (1.2.13).

Remark 1.2.16. The𝐿∞-spaces on𝛾𝑇± with respect to𝑑𝛾𝛼 and the standard surface measure are the same and the respective norms coincide since null sets with respect to 𝑑𝛾𝛼 are null sets with respect to the standard surface measure and vice versa by b𝑣𝛼·𝑛(π‘₯)>0 (<0) on𝛾𝑇+(π›Ύπ‘‡βˆ’). Consequently, from now on we will (mostly) not point out the measure in the denotation of such𝐿∞-spaces and simply write𝐿∞ 𝛾𝑇±

.