SLIDE 1
Computing with Compact Sets: The Gray Code Case
Dieter Spreen University of Siegen
joint work with
Hideki Tsuiki Kyoto University
Continuity, Computability, Constructivity: From Logic to Algorithms Kochel am See, Bavaria, 14-18 September 2015
SLIDE 2
- 1. Signed digit repesentation
Let x = 0.a1a2a3 . . . be a real number in I = [−1, +1]. Then writing a digit a ∈ SD := {−1, 0, 1} in front of a1a2a3 . . . corresponds to applying the contraction da defined by da(x) = x + a 2 to x. In this way we can think of 0.a1a2a3 . . . as an iterated application of operations da ∈ D := { da | a ∈ SD } da1 ◦ da2 ◦ da3 ◦ · · ·
SLIDE 3 Lemma
I is a complete bounded metric space and D a finite set of contractions that covers I: I =
In particular I is compact. Consequently, for α = α0α1 . . . ∈ SDω
dα0 ◦ · · · ◦ dαn−1[I] contains exactly one element. Let [ [α] ] be the unique x ∈
dα0 ◦ · · · ◦ dαn−1[I]
SLIDE 4 Lemma
SDω is a bounded complete and hence compact metric space with metric δ(α, β) :=
2− min{ n|αn=βn }
Proposition
◮ [
[·] ]: SDω → I is onto and uniformly continuous.
◮ The Euclidean topology on I is equivalent to the quotient
topology induced by [ [·] ]. In particular we have that each α ∈ SDω denotes a real number in I, and vice versa.
SLIDE 5
- 2. Pre-Gray code representation
Let pG := {U, D, L, R, FinL, FinR} G := {U, L, R} H := {D, FinL, FinR} Now, not every sequence in pGω is meaningful! The set coG of meaningful sequences is defined together with a set
coH by mutual coninduction: (coG, coH) is the largest subset of
pGω × pGω such that α ∈ coG → (∃β ∈ coG) α ∈ {L, R}β ∨ (∃β ∈ coH) α = Uβ α ∈ coH → (∃β ∈ coG) α ∈ {FinL, FinR}β ∨ (∃β ∈ coH) α = Dβ
SLIDE 6 Proposition
- 1. (coG, δ) is complete.
- 2. (coH, δ) is complete.
Recall that δ is bounded.
Corollary
Both, (coG, δ) and (coH, δ) are compact.
SLIDE 7
Let [ [L] ] := LR−1 [ [R] ] := LR1 [ [U] ] := U [ [FinL] ] := Fin−1 [ [FinR] ] := Fin1 [ [D] ] := D where for a ∈ {−1, 1} LRa(x) := −ax − 1 2 U(x) := x 2 Fina(x) := ax + 1 2 D(x) := x 2
SLIDE 8
Lemma
◮ d1(x) = Fin1(x) = LR1(−x) ◮ d−1(x) = Fin−1(−x) = LR−1(x) ◮ d0(x) = U(x) = D(x)
Corollary
◮ range(d1) = range(Fin1) = −
− → [0, 1], range(LR1) = ← − − [0, 1]
◮ range(d−1) = range(LR−1) = −
− − − → [−1, 0], range(Fin−1) = ← − − − − [−1, 0]
◮ range(d0) = range(U) = range(D) = [−1/2, +1/2]
SLIDE 9 Consequently, both { U, LRa | a ∈ {−1, 1} } and { D, Fina | a ∈ {−1, 1} } are finite covering sets of contractions on I. Hence, for α = α0α1 . . . ∈ coG and β = β0β1 . . . ∈ coH
[ [α0] ] ◦ · · · ◦ [ [αn] ][I] and
[ [β0] ] ◦ · · · ◦ [ [βn] ][I] contain exactly one element. Let {[ [α] ]G} :=
[ [α0] ]◦· · ·◦[ [αn] ][I] and {[ [β] ]H} :=
[ [β0] ]◦· · ·◦[ [βn] ][I]
Proposition
◮ [
[·] ]G : coG → I, [ [·] ]H : coH → I are onto and uniformly continuous.
◮ The Euclidean topology on I is equivalent to the quotient
topology induced by [ [·] ]G ([ [·] ]H).
SLIDE 10
Let K(I) be the collection of all nonempty compact subsets of I.
Proposition
K(I) endowed with the Hausdorff metric µH is a bounded complete, hence compact metric space. Berger/Spreen: There is no finite set of contractions on K(I) that covers K(I). So, it is impossible to represent nonempty compact subsets of I by infinite streams. As we will see, however, they can be represented by infinite trees.
SLIDE 11 Definition
A digital tree is a nonempty set T ⊆ pG∗ of finite sequences of elements of pG that is closed under initial seqments and has no maximal elements, that is, the empty sequence [] is in T and whenever R0 . . . Rn ∈ T, then R0 . . . Rn−1 ∈ T and R0 . . . RnR ∈ T, for some R ∈ pG. Note that each such tree is finitely branching. Moreover, each element R0 . . . Rn ∈ T can be continued to an infinite path α in T. In the following we write α ∈ T to mean that α is an infinite path
- f T and identify T with the set of its infinite paths.
Let T be the set of all digital trees and TG := { T ∩ coG | T ∈ T } TH := { T ∩ coH | T ∈ T }
SLIDE 12 Lemma
(TG, TH) is the largest subset of T × T such that T ∈ TG → (∃T1, . . . , Tn ∈ T)(∃R1, . . . , Rn ∈ G)(∀1 ≤ i ≤ n) [Ri ∈ {L, R} → Ti ∈ TG] ∧ [Ri = U → Ti ∈ TH] ∧ T =
n
RiTi T ∈ TH → (∃T1, . . . , Tn ∈ T)(∃R1, . . . , Rn ∈ H)(∀1 ≤ i ≤ n) [Ri ∈ {FinL, FinR} → Ti ∈ TG] ∧ [Ri = D → Ti ∈ TH] ∧ T =
n
RiTi.
SLIDE 13 Proposition
◮ TG endowed with the Hausdorff metric δH induced by δ is a
bounded complete, hence compact metric space.
◮ TH endowed with the Hausdorff metric δH induced by δ is a
bounded complete, hence compact metric space.
Lemma
Every tree in TG is compact and every nonempty compact subset
- f coG is the set of infinite paths of a digital tree in TG. Similarly,
for TH.
Definition
For T ∈ TG set [ [T] ]TG := { [ [α] ]G | α ∈ T }.
SLIDE 14
Since [ [·] ]G is continuous, [ [T] ]TG is compact. Conversely, for any K ∈ K(I), [ [·] ]−1
G [K] is closed and hence compact, as TD is
compact.
Lemma
The nonempty compact subsets of I are exactly the values of trees in TG.
Proposition
◮ [
[·] ]TG : TG → I is onto and uniformly continuous.
◮ The topology on K(I) induced by the Hausdorff metric is
equivalent to the quotient topology induced by [ [·] ]TG .
SLIDE 15 Let D := { da | a ∈ {−1, 0, +1} }, G := {U, LR−1, LR1}, H := {D, Fin−1, Fin1}. In Berger/Spreen a representation-free coinductive approach to computations with compact subsets of I was presented. Let CK be the largest subset of K(I) such that CK(A) → (∃B ⊆ D)(∃(Ab)b∈B ∈ K(I)B)(∀b ∈ B)CK(Ab) ∧ A =
b[Ab].
Lemma (Berger/Spreen)
K(I) = CK.
SLIDE 16 Let (GK, HK) be the largest subset of K(I) × K(I) such that GK(A) → (∃E ⊆ G)(∃(Ae)e∈E ∈ K(I)E) A =
e[Ae] ∧ (∀e ∈ E) [e ∈ {LR−1, LR1} → GK(Ae)] ∧ [e = U → HK(Ae)], HK(A) → (∃F ⊆ H)(∃(Af )f ∈F ∈ K(I)F) A =
f [Af ] ∧ (∀f ∈ F) [f ∈ {Fin−1, Fin1} → GK(Af )] ∧ [f = D → HK(Af )].
Theorem
GK = CK.
SLIDE 17
◮ From the proof a computable translator between realizers for
both sides can be derived, that is, between trees in TG and trees over SD.
◮ In Berger/Spreen the equivalence between the latter
representation and the Cauchy representation for nonempty compact sets was shown.
◮ Consequently, pre-Gray computability of nonempty compact
subsets of I is equivalent to TTE computability.