Intro and main Background, Zinbiel, and Hopf …gilliam/ttu/cib_mem_webpage/Chris...Intro and main...
Transcript of Intro and main Background, Zinbiel, and Hopf …gilliam/ttu/cib_mem_webpage/Chris...Intro and main...
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
1
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Algebra and the geometryof nonlinear control systems
Matthias Kawski †
School of Mathematical and Statistical SciencesArizona State UniversityTempe, AZ 85287, USA
http://math.asu.edu/~kawski
Workshop in Celebration of the Life, Mathematics,and Memories of Christopher I. Byrnes
Lubbock, TX, September 2010
1This work was partially supported by the National Science Foundationthrough the grants DMS 09-08204 and DMS 09-60589,.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Trimester: Combinatorics & Control 2010, Madrid
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Outline
• Background: From control to algebra and combinatorics
• Recall: Hall sets and coord’s of 2nd kind
• Hopf algebras and convolutions
• Coord’s of 1st kind w/ examples
• Conclusion and outlook
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
In a nut-shell
• Recall: Chen Fliess series is an exponential Lie series
∑w∈Z∗
w ⊗ w = exp
(∑b∈B
ζb ⊗ b
)=←∏
b∈Bexp(ξb ⊗ b)
where B is ordered basis of free Lie algebra L(Z ) ⊂ A(Z )
• Known (coord’s of 2nd kind): ξHK = ξH ∗ ξK
• New/recent (coord’s of the 1st kind): ζ = π′1 ◦ ξ
• Use in control / geometric integration: explicit formula foriterated integral functionals Υζb and Υξb .
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
In a nut-shell
• Recall: Chen Fliess series is an exponential Lie series
∑w∈Z∗
w ⊗ w = exp
(∑b∈B
ζb ⊗ b
)=←∏
b∈Bexp(ξb ⊗ b)
where B is ordered basis of free Lie algebra L(Z ) ⊂ A(Z )
• Known (coord’s of 2nd kind): ξHK = ξH ∗ ξK
• New/recent (coord’s of the 1st kind): ζ = π′1 ◦ ξ
• Use in control / geometric integration: explicit formula foriterated integral functionals Υζb and Υξb .
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
In a nut-shell
• Recall: Chen Fliess series is an exponential Lie series
∑w∈Z∗
w ⊗ w = exp
(∑b∈B
ζb ⊗ b
)=←∏
b∈Bexp(ξb ⊗ b)
where B is ordered basis of free Lie algebra L(Z ) ⊂ A(Z )
• Known (coord’s of 2nd kind): ξHK = ξH ∗ ξK
• New/recent (coord’s of the 1st kind): ζ = π′1 ◦ ξ
• Use in control / geometric integration: explicit formula foriterated integral functionals Υζb and Υξb .
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
In a nut-shell
• Recall: Chen Fliess series is an exponential Lie series
∑w∈Z∗
w ⊗ w = exp
(∑b∈B
ζb ⊗ b
)=←∏
b∈Bexp(ξb ⊗ b)
where B is ordered basis of free Lie algebra L(Z ) ⊂ A(Z )
• Known (coord’s of 2nd kind): ξHK = ξH ∗ ξK
• New/recent (coord’s of the 1st kind): ζ = π′1 ◦ ξ
• Use in control / geometric integration: explicit formula foriterated integral functionals Υζb and Υξb .
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Basic set-up: splitting
Splitting into geometric state-dependent and time-varying parts
x = u1(t)f1(x) + . . .+ um(t)fm(x) = F (t , x)y = ϕ(x)
fi : M 7→ TM smooth vector fields on manifold Mn,
u : [0,T ] 7→ U ⊂⊂ Rm measurable controls/perturbations
,
and
φ ∈ Cω(M) measured output
.
not only control, but also geometric numerical integration,stochastic DE (rough paths [Lyons]) ldots.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Basic set-up: splitting
Splitting into geometric state-dependent and time-varying parts
x = u1(t)f1(x) + . . .+ um(t)fm(x) = F (t , x)y = ϕ(x)
fi : M 7→ TM smooth vector fields on manifold Mn,
u : [0,T ] 7→ U ⊂⊂ Rm measurable controls/perturbations
,
and
φ ∈ Cω(M) measured output
.
not only control, but also geometric numerical integration,stochastic DE (rough paths [Lyons]) ldots.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Basic set-up: splitting
Splitting into geometric state-dependent and time-varying parts
x = u1(t)f1(x) + . . .+ um(t)fm(x) = F (t , x)y = ϕ(x)
fi : M 7→ TM smooth vector fields on manifold Mn,
u : [0,T ] 7→ U ⊂⊂ Rm measurable controls/perturbations,
and
φ ∈ Cω(M) measured output
.
not only control, but also geometric numerical integration,stochastic DE (rough paths [Lyons]) ldots.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Basic set-up: splitting
Splitting into geometric state-dependent and time-varying parts
x = u1(t)f1(x) + . . .+ um(t)fm(x) = F (t , x)y = ϕ(x)
fi : M 7→ TM smooth vector fields on manifold Mn,
u : [0,T ] 7→ U ⊂⊂ Rm measurable controls/perturbations, and
φ ∈ Cω(M) measured output.
not only control, but also geometric numerical integration,stochastic DE (rough paths [Lyons]) ldots.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Long history
Combinatorics and algebra behind integration of ODEs
Just a few names
Campbell-Baker-Hausdorff, Magnus, Chen,
Schützenberger, Reutenauer, Fer, Ferfera,
Fliess, Lamnabhi-Lagarrigue, Crouch, Gray,
Grossmann, Larson, Rota and Baxter, Ebrahimi-Fard, . . .
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
A closer look at manipulating series expansions
φ(x(t ,u)) = 1 · φ(x0)
+∫ t
0ua(s)ds (faφ)(x0)
+∫ t
0ub(s)ds (fbφ)(x0)
+ 12
∫ t0
∫ s10
∫ s10 ua(s1)ua(s2)ds2ds1 (fafaφ)(x0)
+ 12
∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafbφ)(x0)
+ 12
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 (fbfaφ)(x0)
+ 12
∫ t0
∫ s10 ub(s1)ub(s2)ds2ds1 (fbfbφ)(x0)
+ 16
∫ t0
∫ s10
∫ s20 ua(s1)ua(s2)ua(s3)ds3ds2ds1 (fafafaφ)(x0)
+ . . .
Objective: Collect first order differential operators, and minimizenumber of higher order differential operators involved
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
A closer look at manipulating series expansions
φ(x(t ,u)) = 1 · φ(x0)
+∫ t
0ua(s)ds (faφ)(x0)
+∫ t
0ub(s)ds (fbφ)(x0)
+ 12
∫ t0
∫ s10
∫ s10 ua(s1)ua(s2)ds2ds1 (fafaφ)(x0)
+ 12
∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafbφ)(x0)
+ 12
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 (fbfaφ)(x0)
+ 12
∫ t0
∫ s10 ub(s1)ub(s2)ds2ds1 (fbfbφ)(x0)
+ 16
∫ t0
∫ s10
∫ s20 ua(s1)ua(s2)ua(s3)ds3ds2ds1 (fafafaφ)(x0)
+ . . .Objective: Collect first order differential operators, and minimizenumber of higher order differential operators involved
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Integrate by parts: The wrong way to do it∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 fafb+
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 fbfa =
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 fafb −
∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 fbfa
+∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 fbfa +
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 fbfa
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafb − fbfa)
+∫ t
0
(ua(s1)
∫ s10 ub(s2) + ub(s1)
∫ s10 ua(s2)ds2
)ds1 fbfa
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafb − fbfa)
+
(∫ t0ua(s) ds
)·(∫ t
0ua(s) ds)
fbfa
Lie brackets together w/ iterated integrals in right orderhigher order deriv’s (wrong order) w/ pointwise prod’s of int’s
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Integrate by parts: The wrong way to do it∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 fafb+
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 fbfa =
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 fafb −
∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 fbfa
+∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 fbfa +
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 fbfa
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafb − fbfa)
+∫ t
0
(ua(s1)
∫ s10 ub(s2) + ub(s1)
∫ s10 ua(s2)ds2
)ds1 fbfa
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafb − fbfa)
+
(∫ t0ua(s) ds
)·(∫ t
0ua(s) ds)
fbfa
Lie brackets together w/ iterated integrals in right orderhigher order deriv’s (wrong order) w/ pointwise prod’s of int’s
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Integrate by parts: The wrong way to do it∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 fafb+
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 fbfa =
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 fafb −
∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 fbfa
+∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 fbfa +
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 fbfa
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafb − fbfa)
+∫ t
0
(ua(s1)
∫ s10 ub(s2) + ub(s1)
∫ s10 ua(s2)ds2
)ds1 fbfa
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafb − fbfa)
+
(∫ t0ua(s) ds
)·(∫ t
0ua(s) ds)
fbfa
Lie brackets together w/ iterated integrals in right orderhigher order deriv’s (wrong order) w/ pointwise prod’s of int’s
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Integrate by parts: The wrong way to do it∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 fafb+
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 fbfa =
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 fafb −
∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 fbfa
+∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 fbfa +
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 fbfa
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafb − fbfa)
+∫ t
0
(ua(s1)
∫ s10 ub(s2) + ub(s1)
∫ s10 ua(s2)ds2
)ds1 fbfa
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafb − fbfa)
+
(∫ t0ua(s) ds
)·(∫ t
0ua(s) ds)
fbfa
Lie brackets together w/ iterated integrals in right orderhigher order deriv’s (wrong order) w/ pointwise prod’s of int’s
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Integrate by parts: The wrong way to do it∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 fafb+
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 fbfa =
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 fafb −
∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 fbfa
+∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 fbfa +
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 fbfa
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafb − fbfa)
+∫ t
0
(ua(s1)
∫ s10 ub(s2) + ub(s1)
∫ s10 ua(s2)ds2
)ds1 fbfa
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafb − fbfa)
+
(∫ t0ua(s) ds
)·(∫ t
0ua(s) ds)
fbfa
Lie brackets together w/ iterated integrals in right orderhigher order deriv’s (wrong order) w/ pointwise prod’s of int’s
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Integrate by parts: The wrong way to do it∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 fafb+
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 fbfa =
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 fafb −
∫ t0
∫ s10 ua(s1)ub(s2)ds2ds1 fbfa
+∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 fbfa +
∫ t0
∫ s10 ub(s1)ua(s2)ds2ds1 fbfa
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafb − fbfa)
+∫ t
0
(ua(s1)
∫ s10 ub(s2) + ub(s1)
∫ s10 ua(s2)ds2
)ds1 fbfa
=∫ t
0
∫ s10 ua(s1)ub(s2)ds2ds1 (fafb − fbfa)
+
(∫ t0ua(s) ds
)·(∫ t
0ua(s) ds)
fbfa
Lie brackets together w/ iterated integrals in right orderhigher order deriv’s (wrong order) w/ pointwise prod’s of int’s
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Integrate by parts, smart way. Hopf algebra preview
Do not manipulate iterated integrals and iterated Lie brackets ofvector fields by hand – work on level of “words” (their indices)∑
I∈{a,b}∗I ⊗ I = 1⊗ 1
+ a ⊗ a
+ b ⊗ b
+ 12 aa ⊗ aa
+ 12 ab ⊗ ab
+ 12 ba ⊗ ba
+ 12 bb ⊗ bb
+ 16 aaa ⊗ aaa
+ . . .
= 1⊗ 1
+ a ⊗ a
+ b ⊗ b
+ 12 aa ⊗ aa
+ 12 ab ⊗ (ab − ba)
+ 12 (ab + ba) ⊗ ba
+ 12 bb ⊗ bb
+ 16 aaa ⊗ aaa
+ . . .
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Drop everything except the indices
• The iterated integral
Υi1i2...in =
∫ t
0
∫ t1
0· · ·∫ tn−1
0ui1(t1)ui2(t2) · · · uin (tn) dtn dtn−1 · · · dt1
is uniquely identified by the multi-index (“word”) i1i2 . . . in
• The n-th order partial differential operator fin ◦ fin−1 ◦ . . . fi1is uniquely identified by the multi-index (“word”) i1i2 . . . in
• The Chen series is identified with the identity map on freeassociative algebra A(Z ) over set of indices Z = {1, . . .m}
CF ∼ IdA(Z ) =∑n≥0
∑w∈Z n
w ⊗ w ∈ A(Z )⊗ A(Z )
with shuffle product on left and concatenation on right
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Drop everything except the indices
• The iterated integral
Υi1i2...in =
∫ t
0
∫ t1
0· · ·∫ tn−1
0ui1(t1)ui2(t2) · · · uin (tn) dtn dtn−1 · · · dt1
is uniquely identified by the multi-index (“word”) i1i2 . . . in
• The n-th order partial differential operator fin ◦ fin−1 ◦ . . . fi1is uniquely identified by the multi-index (“word”) i1i2 . . . in
• The Chen series is identified with the identity map on freeassociative algebra A(Z ) over set of indices Z = {1, . . .m}
CF ∼ IdA(Z ) =∑n≥0
∑w∈Z n
w ⊗ w ∈ A(Z )⊗ A(Z )
with shuffle product on left and concatenation on right
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Drop everything except the indices
• The iterated integral
Υi1i2...in =
∫ t
0
∫ t1
0· · ·∫ tn−1
0ui1(t1)ui2(t2) · · · uin (tn) dtn dtn−1 · · · dt1
is uniquely identified by the multi-index (“word”) i1i2 . . . in
• The n-th order partial differential operator fin ◦ fin−1 ◦ . . . fi1is uniquely identified by the multi-index (“word”) i1i2 . . . in
• The Chen series is identified with the identity map on freeassociative algebra A(Z ) over set of indices Z = {1, . . .m}
CF ∼ IdA(Z ) =∑n≥0
∑w∈Z n
w ⊗ w ∈ A(Z )⊗ A(Z )
with shuffle product on left and concatenation on right
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Homomorphisms I
• For fixed smooth vector fields fi
F : A(Z ) 7→ partial diff operators on C∞(M)
F : (i1i2 . . . in) 7→ fi1 ◦ fi2 ◦ . . . finassociative algebras: concatenation 7→ composition
• For fixed control u ∈ UZ
Υ(u) : A(Z ) 7→ AC([0,T ],R)
Υ(u) : (i1i2 . . . in) 7→∫ T
0
∫ t1
0· · ·∫ tp−1
0uip (tp) . . . ui1(t1) dt1 . . . dtp
associative algebras (Ree’s theorem):shuffle of words 7→ pointwise multiplication of functions
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Homomorphisms I
• For fixed smooth vector fields fi
F : A(Z ) 7→ partial diff operators on C∞(M)
F : (i1i2 . . . in) 7→ fi1 ◦ fi2 ◦ . . . finassociative algebras: concatenation 7→ composition
• For fixed control u ∈ UZ
Υ(u) : A(Z ) 7→ AC([0,T ],R)
Υ(u) : (i1i2 . . . in) 7→∫ T
0
∫ t1
0· · ·∫ tp−1
0uip (tp) . . . ui1(t1) dt1 . . . dtp
associative algebras (Ree’s theorem):shuffle of words 7→ pointwise multiplication of functions
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Homomorphisms II
• Restriction is Lie algebra homomorphism
F : L(Z ) ⊆ A(Z ) 7→ Γ∞(M) (vector fields)
• Do not fix controls: iterated integral functionals
Υ: ∈ A(Z ) 7→ IIF(UZ )
Υ: (i1i2 . . . in) 7→
(u 7→
∫ T
0
∫ t1
0· · ·∫ tp−1
0uip (tp) . . . ui1(t1) dt1 . . . dtp
)associative algebras: shuffle of words 7→ pointwisemultiplication of iterated integral functionals
• Much better:
Theorem: If U = L1([0,T ], [−1,1]) thenΥ: (A(Z ), ∗) 7→ IIF(UZ ) is a Zinbiel algebra isomorphism.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Homomorphisms II
• Restriction is Lie algebra homomorphism
F : L(Z ) ⊆ A(Z ) 7→ Γ∞(M) (vector fields)
• Do not fix controls: iterated integral functionals
Υ: ∈ A(Z ) 7→ IIF(UZ )
Υ: (i1i2 . . . in) 7→
(u 7→
∫ T
0
∫ t1
0· · ·∫ tp−1
0uip (tp) . . . ui1(t1) dt1 . . . dtp
)associative algebras: shuffle of words 7→ pointwisemultiplication of iterated integral functionals
• Much better:
Theorem: If U = L1([0,T ], [−1,1]) thenΥ: (A(Z ), ∗) 7→ IIF(UZ ) is a Zinbiel algebra isomorphism.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Homomorphisms II
• Restriction is Lie algebra homomorphism
F : L(Z ) ⊆ A(Z ) 7→ Γ∞(M) (vector fields)
• Do not fix controls: iterated integral functionals
Υ: ∈ A(Z ) 7→ IIF(UZ )
Υ: (i1i2 . . . in) 7→
(u 7→
∫ T
0
∫ t1
0· · ·∫ tp−1
0uip (tp) . . . ui1(t1) dt1 . . . dtp
)associative algebras: shuffle of words 7→ pointwisemultiplication of iterated integral functionals
• Much better:
Theorem: If U = L1([0,T ], [−1,1]) thenΥ: (A(Z ), ∗) 7→ IIF(UZ ) is a Zinbiel algebra isomorphism.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Homomorphisms II
• Restriction is Lie algebra homomorphism
F : L(Z ) ⊆ A(Z ) 7→ Γ∞(M) (vector fields)
• Do not fix controls: iterated integral functionals
Υ: ∈ A(Z ) 7→ IIF(UZ )
Υ: (i1i2 . . . in) 7→
(u 7→
∫ T
0
∫ t1
0· · ·∫ tp−1
0uip (tp) . . . ui1(t1) dt1 . . . dtp
)associative algebras: shuffle of words 7→ pointwisemultiplication of iterated integral functionals
• Much better: Theorem: If U = L1([0,T ], [−1,1]) thenΥ: (A(Z ), ∗) 7→ IIF(UZ ) is a Zinbiel algebra isomorphism.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Zinbiel product, Hall sets, and formula for ξ
The product of control theory (U ∗ V )(t) =∫ t
0 U(s) V ′(s) dssatisfies the right Zinbiel identity
U ∗ (V ∗W ) = (U ∗ V ) ∗W + (V ∗ U) ∗W
Theorem (Schützenberger, Sussmann, Grossman,Reutenauer-Melançon)For a Hall set H over Z and letters a ∈ Z and H,K ,HK ∈ H
ξa = aξHK = µHK ξH ∗ ξK
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Zinbiel product, Hall sets, and formula for ξ
The product of control theory (U ∗ V )(t) =∫ t
0 U(s) V ′(s) dssatisfies the right Zinbiel identity
U ∗ (V ∗W ) = (U ∗ V ) ∗W + (V ∗ U) ∗W
Theorem (Schützenberger, Sussmann, Grossman,Reutenauer-Melançon)For a Hall set H over Z and letters a ∈ Z and H,K ,HK ∈ H
ξa = aξHK = µHK ξH ∗ ξK
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Illustration: Normal form for nilpotent systemNormal from for a free nilpotent system (of rank r = 5)using a typical Hall set on the alphabet Z = {a,b}
ξa = ua
ξb = ubξab = ξa · ξb = ξa ubξaab = ξa · ξab = ξ2
a ub using ψ−b(aab) = (a(ab))
ξbab = ξb · ξab = ξbξa ub using ψ−b(bab) = (b(ab))
ξaaab = ξa · ξaab = ξ3a ub using ψ−b(aaab) = (a(a(ab)))
ξbaab = ξb · ξaab = ξbξ2a ub using ψ−b(baab) = (b(a(ab)))
ξbbab = ξb · ξbab = ξ2bξa ub using ψ−b(bbab) = (b(b(ab)))
ξaaaab = ξa · ξaaab = ξ4a ub using ψ−b(aaaab) = (a(a(a(ab))))
ξbaaab = ξb · ξaaab = ξbξ3a ub using ψ−b(baaab) = (b(a(a(ab))))
ξbbaab = ξb · ξbaab = ξ2bξ
2a ub using ψ−b(bbaab) = (b(b(a(ab))))
ξabaab = ξab · ξaab = ξabξ3a ub using ψ−b(abaab) = ((ab)(a(ab)))
ξabbab = ξab · ξbab = ξabξ2bξa ub using ψ−b(abbab) = ((ab)(b(ab)))
Cost: accept label coord’s not by integers, but by Hall wordsBenefit: One-line formula for all iterated integrals! ξHK = ξH ∗ ξK
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Illustration: Normal form for nilpotent systemNormal from for a free nilpotent system (of rank r = 5)using a typical Hall set on the alphabet Z = {a,b}
ξa = ua
ξb = ubξab = ξa · ξb = ξa ubξaab = ξa · ξab = ξ2
a ub using ψ−b(aab) = (a(ab))
ξbab = ξb · ξab = ξbξa ub using ψ−b(bab) = (b(ab))
ξaaab = ξa · ξaab = ξ3a ub using ψ−b(aaab) = (a(a(ab)))
ξbaab = ξb · ξaab = ξbξ2a ub using ψ−b(baab) = (b(a(ab)))
ξbbab = ξb · ξbab = ξ2bξa ub using ψ−b(bbab) = (b(b(ab)))
ξaaaab = ξa · ξaaab = ξ4a ub using ψ−b(aaaab) = (a(a(a(ab))))
ξbaaab = ξb · ξaaab = ξbξ3a ub using ψ−b(baaab) = (b(a(a(ab))))
ξbbaab = ξb · ξbaab = ξ2bξ
2a ub using ψ−b(bbaab) = (b(b(a(ab))))
ξabaab = ξab · ξaab = ξabξ3a ub using ψ−b(abaab) = ((ab)(a(ab)))
ξabbab = ξab · ξbab = ξabξ2bξa ub using ψ−b(abbab) = ((ab)(b(ab)))
Cost: accept label coord’s not by integers, but by Hall words
Benefit: One-line formula for all iterated integrals! ξHK = ξH ∗ ξK
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Illustration: Normal form for nilpotent systemNormal from for a free nilpotent system (of rank r = 5)using a typical Hall set on the alphabet Z = {a,b}
ξa = ua
ξb = ubξab = ξa · ξb = ξa ubξaab = ξa · ξab = ξ2
a ub using ψ−b(aab) = (a(ab))
ξbab = ξb · ξab = ξbξa ub using ψ−b(bab) = (b(ab))
ξaaab = ξa · ξaab = ξ3a ub using ψ−b(aaab) = (a(a(ab)))
ξbaab = ξb · ξaab = ξbξ2a ub using ψ−b(baab) = (b(a(ab)))
ξbbab = ξb · ξbab = ξ2bξa ub using ψ−b(bbab) = (b(b(ab)))
ξaaaab = ξa · ξaaab = ξ4a ub using ψ−b(aaaab) = (a(a(a(ab))))
ξbaaab = ξb · ξaaab = ξbξ3a ub using ψ−b(baaab) = (b(a(a(ab))))
ξbbaab = ξb · ξbaab = ξ2bξ
2a ub using ψ−b(bbaab) = (b(b(a(ab))))
ξabaab = ξab · ξaab = ξabξ3a ub using ψ−b(abaab) = ((ab)(a(ab)))
ξabbab = ξab · ξbab = ξabξ2bξa ub using ψ−b(abbab) = ((ab)(b(ab)))
Cost: accept label coord’s not by integers, but by Hall wordsBenefit: One-line formula for all iterated integrals! ξHK = ξH ∗ ξK
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Hopf algebra and dynamical systems
Long history of using combinatorial Hopf algebras for analysisof dynamical systems.
E.g. Grossmann and Larson, Hopf algebra structure of labeledtrees, and of differential operators
Main reference for this work: Reutenauer, Free Lie Algebras
Recent related work, numerical “geometric” integration:Murua, Ebrahimi-Fard, Munthe-Kaas, . . .
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Easy intro to Hopf algebra: shuffle
Combinatorially: for words w , z ∈ Z ∗ and letters a,b ∈ Z
( w a ) X ( z b ) = (( w a ) X z ) b + ( w X ( z b )) a
Example: (ab) X (cd) = a b c d + a c b d + c a b d +a c d b + c a d b + c d a b
Algebraically: transpose of the coproduct ∆
< v X w , z >=< v ⊗ w , ∆(z) >
where
∆: A(Z ) 7→ A(Z )⊗ A(Z ) by ∆(a) = 1⊗a+a⊗1 for a ∈ Z
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Easy intro to Hopf algebra: shuffle
Combinatorially: for words w , z ∈ Z ∗ and letters a,b ∈ Z
( w a ) X ( z b ) = (( w a ) X z ) b + ( w X ( z b )) a
Example: (ab) X (cd) = a b c d + a c b d + c a b d +a c d b + c a d b + c d a b
Algebraically: transpose of the coproduct ∆
< v X w , z >=< v ⊗ w , ∆(z) >
where
∆: A(Z ) 7→ A(Z )⊗ A(Z ) by ∆(a) = 1⊗a+a⊗1 for a ∈ Z
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Two Hopf algebra structures on A(Z )
• Products: Concatenation and shuffle
• Coproducts:∆: A(Z ) 7→ A(Z )⊗ A(Z ) concatenation homo∆′ : A(Z ) 7→ A(Z )⊗ A(Z ) shuffle homomorphismFor a ∈ Z , ∆(a) = ∆′(a) = 1⊗ a + a⊗ 1, and for w ∈ Z ∗
∆(w) =∑
u,v∈Z∗〈w ,u x v〉 u ⊗ v .∆′(w) =
∑u,v∈Z∗〈w ,uv〉 u ⊗ v .
• Antipode (same for both): linear map α : A(Z ) 7→ A(Z )if w = a1a2 . . . ak ∈ Z k then α(w) = (−)k ak . . . a2a1.
• Example: If a,b ∈ Z then
∆′(ab) = 1⊗ ab + a⊗ b + ab⊗ 1and
∆(ab) = ∆(a)∆(b) = 1⊗ ab + a⊗ b + b ⊗ a + ab ⊗ 1.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Two Hopf algebra structures on A(Z )
• Products: Concatenation and shuffle
• Coproducts:∆: A(Z ) 7→ A(Z )⊗ A(Z ) concatenation homo∆′ : A(Z ) 7→ A(Z )⊗ A(Z ) shuffle homomorphismFor a ∈ Z , ∆(a) = ∆′(a) = 1⊗ a + a⊗ 1, and for w ∈ Z ∗
∆(w) =∑
u,v∈Z∗〈w ,u x v〉 u ⊗ v .∆′(w) =
∑u,v∈Z∗〈w ,uv〉 u ⊗ v .
• Antipode (same for both): linear map α : A(Z ) 7→ A(Z )if w = a1a2 . . . ak ∈ Z k then α(w) = (−)k ak . . . a2a1.
• Example: If a,b ∈ Z then
∆′(ab) = 1⊗ ab + a⊗ b + ab⊗ 1and
∆(ab) = ∆(a)∆(b) = 1⊗ ab + a⊗ b + b ⊗ a + ab ⊗ 1.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Two Hopf algebra structures on A(Z )
• Products: Concatenation and shuffle
• Coproducts:∆: A(Z ) 7→ A(Z )⊗ A(Z ) concatenation homo∆′ : A(Z ) 7→ A(Z )⊗ A(Z ) shuffle homomorphismFor a ∈ Z , ∆(a) = ∆′(a) = 1⊗ a + a⊗ 1, and for w ∈ Z ∗
∆(w) =∑
u,v∈Z∗〈w ,u x v〉 u ⊗ v .∆′(w) =
∑u,v∈Z∗〈w ,uv〉 u ⊗ v .
• Antipode (same for both): linear map α : A(Z ) 7→ A(Z )if w = a1a2 . . . ak ∈ Z k then α(w) = (−)k ak . . . a2a1.
• Example: If a,b ∈ Z then
∆′(ab) = 1⊗ ab + a⊗ b + ab⊗ 1and
∆(ab) = ∆(a)∆(b) = 1⊗ ab + a⊗ b + b ⊗ a + ab ⊗ 1.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Two Hopf algebra structures on A(Z )
• Products: Concatenation and shuffle
• Coproducts:∆: A(Z ) 7→ A(Z )⊗ A(Z ) concatenation homo∆′ : A(Z ) 7→ A(Z )⊗ A(Z ) shuffle homomorphismFor a ∈ Z , ∆(a) = ∆′(a) = 1⊗ a + a⊗ 1, and for w ∈ Z ∗
∆(w) =∑
u,v∈Z∗〈w ,u x v〉 u ⊗ v .∆′(w) =
∑u,v∈Z∗〈w ,uv〉 u ⊗ v .
• Antipode (same for both): linear map α : A(Z ) 7→ A(Z )if w = a1a2 . . . ak ∈ Z k then α(w) = (−)k ak . . . a2a1.
• Example: If a,b ∈ Z then ∆′(ab) = 1⊗ ab + a⊗ b + ab⊗ 1and ∆(ab) = ∆(a)∆(b) = 1⊗ ab + a⊗ b + b ⊗ a + ab ⊗ 1.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Two Hopf algebra structures on A(Z )
• Products: Concatenation and shuffle
• Coproducts:∆: A(Z ) 7→ A(Z )⊗ A(Z ) concatenation homo∆′ : A(Z ) 7→ A(Z )⊗ A(Z ) shuffle homomorphismFor a ∈ Z , ∆(a) = ∆′(a) = 1⊗ a + a⊗ 1, and for w ∈ Z ∗
∆(w) =∑
u,v∈Z∗〈w ,u x v〉 u ⊗ v .∆′(w) =
∑u,v∈Z∗〈w ,uv〉 u ⊗ v .
• Antipode (same for both): linear map α : A(Z ) 7→ A(Z )if w = a1a2 . . . ak ∈ Z k then α(w) = (−)k ak . . . a2a1.
• Example: If a,b ∈ Z then ∆′(ab) = 1⊗ ab + a⊗ b + ab⊗ 1and ∆(ab) = ∆(a)∆(b) = 1⊗ ab + a⊗ b + b ⊗ a + ab ⊗ 1.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Two convolution productsEach Hopf algebra structure yields an associative convolutionproduct of linear endomorphisms f ,g : A(Z ) 7→ A(Z )
f ? g = conc ◦ (f ⊗ g) ◦∆, and
f ?′ g = shu ◦ (f ⊗ g) ◦∆′,
For w ∈ A(Z )
(f ? g)(w) =∑
u,v∈Z+
〈w ,u x v〉 f (u)g(v) and
(f ?′ g)(w) =∑
u,v∈Z+
〈w ,uv〉 f (u) x g(v)
Note: Apparent confusion of ? and ?′ led to incorrect formulasfor ζh in the literature.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Fundamental subspaces and projections• Direct sum decomposition A(Z ) =
⊕∞k=1 Uk where for
k ∈ Z+, Uk ⊆ A(Z ) is linear subspace spanned by `k
(concatenation) for ` ∈ L(Z ). Alternatively, Uk is spannedby symmetric products of Lie polynomials `1, . . . `k ∈ L(Z )
Sym(`1, . . . `k ) = 1k!
∑σ∈Sk
`σ(1)`σ(2) . . . `σ(k)
• associated projections (NOT orthogonal) πk : A(Z ) 7→ Uk
π1 =∑k≥1
(−)k−1 1k I?k
=∑k≥1
(−)k−1 1k conck ◦ I⊗k ◦∆k , and
πk = 1k!π
?k1 .
where I : A(Z ) 7→ A(Z ), I(1) = 0, I(w) = w for w ∈ Z ∗.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Fundamental subspaces and projections• Direct sum decomposition A(Z ) =
⊕∞k=1 Uk where for
k ∈ Z+, Uk ⊆ A(Z ) is linear subspace spanned by `k
(concatenation) for ` ∈ L(Z ). Alternatively, Uk is spannedby symmetric products of Lie polynomials `1, . . . `k ∈ L(Z )
Sym(`1, . . . `k ) = 1k!
∑σ∈Sk
`σ(1)`σ(2) . . . `σ(k)
• associated projections (NOT orthogonal) πk : A(Z ) 7→ Uk
π1 =∑k≥1
(−)k−1 1k I?k
=∑k≥1
(−)k−1 1k conck ◦ I⊗k ◦∆k , and
πk = 1k!π
?k1 .
where I : A(Z ) 7→ A(Z ), I(1) = 0, I(w) = w for w ∈ Z ∗.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Fundamental subspaces and projections• Direct sum decomposition A(Z ) =
⊕∞k=1 Uk where for
k ∈ Z+, Uk ⊆ A(Z ) is linear subspace spanned by `k
(concatenation) for ` ∈ L(Z ). Alternatively, Uk is spannedby symmetric products of Lie polynomials `1, . . . `k ∈ L(Z )
Sym(`1, . . . `k ) = 1k!
∑σ∈Sk
`σ(1)`σ(2) . . . `σ(k)
• associated projections (NOT orthogonal) πk : A(Z ) 7→ Uk
π1 =∑k≥1
(−)k−1 1k I?k
=∑k≥1
(−)k−1 1k conck ◦ I⊗k ◦∆k , and
πk = 1k!π
?k1 .
where I : A(Z ) 7→ A(Z ), I(1) = 0, I(w) = w for w ∈ Z ∗.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Fundamental subspaces and projections• Direct sum decomposition A(Z ) =
⊕∞k=1 Uk where for
k ∈ Z+, Uk ⊆ A(Z ) is linear subspace spanned by `k
(concatenation) for ` ∈ L(Z ). Alternatively, Uk is spannedby symmetric products of Lie polynomials `1, . . . `k ∈ L(Z )
Sym(`1, . . . `k ) = 1k!
∑σ∈Sk
`σ(1)`σ(2) . . . `σ(k)
• associated projections (NOT orthogonal) πk : A(Z ) 7→ Uk
π1 =∑k≥1
(−)k−1 1k I?k
=∑k≥1
(−)k−1 1k conck ◦ I⊗k ◦∆k , and
πk = 1k!π
?k1 .
where I : A(Z ) 7→ A(Z ), I(1) = 0, I(w) = w for w ∈ Z ∗.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Fundamental subspaces and projections• Direct sum decomposition A(Z ) =
⊕∞k=1 Uk where for
k ∈ Z+, Uk ⊆ A(Z ) is linear subspace spanned by `k
(concatenation) for ` ∈ L(Z ). Alternatively, Uk is spannedby symmetric products of Lie polynomials `1, . . . `k ∈ L(Z )
Sym(`1, . . . `k ) = 1k!
∑σ∈Sk
`σ(1)`σ(2) . . . `σ(k)
• associated projections (NOT orthogonal) πk : A(Z ) 7→ Uk
π1 =∑k≥1
(−)k−1 1k I?k
=∑k≥1
(−)k−1 1k conck ◦ I⊗k ◦∆k
, and
πk = 1k!π
?k1 .
where I : A(Z ) 7→ A(Z ), I(1) = 0, I(w) = w for w ∈ Z ∗.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Fundamental subspaces and projections• Direct sum decomposition A(Z ) =
⊕∞k=1 Uk where for
k ∈ Z+, Uk ⊆ A(Z ) is linear subspace spanned by `k
(concatenation) for ` ∈ L(Z ). Alternatively, Uk is spannedby symmetric products of Lie polynomials `1, . . . `k ∈ L(Z )
Sym(`1, . . . `k ) = 1k!
∑σ∈Sk
`σ(1)`σ(2) . . . `σ(k)
• associated projections (NOT orthogonal) πk : A(Z ) 7→ Uk
π1 =∑k≥1
(−)k−1 1k I?k
=∑k≥1
(−)k−1 1k conck ◦ I⊗k ◦∆k , and
πk = 1k!π
?k1 .
where I : A(Z ) 7→ A(Z ), I(1) = 0, I(w) = w for w ∈ Z ∗.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
The adjoint π′1
Define the adjoint map π′1 : A(Z ) 7→ A(Z ) by
〈w , π1(z)〉 = 〈π′1(w), z〉 for all w , z ∈ Z.
Since I is selfadjoint, conc and ∆′ are adjoints of each other, asare x and ∆. Hence, using the second convolution product ?′
π′1 =∑k≥1
(−)k−1 1k I?
′k . =∑k≥1
(−)k−1 1k shuk ◦ I⊗k ◦∆′k
On words w ∈ Z , calculate (sum over uj 6= 1)
π′1(w) =∑k≥1
1k (−)k−1
∑u1...uk=w
u1 x . . . x uk
Example: π′1(ξab) = π′1(ab) = ab − 12 a x b = 1
2 (ab − ba).
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Formulas for the ζh
Theorem[E. Gehrig].On every Hall set ζ = π′1 ◦ ξ.
Moreover: (recall PBW basis and Viennot) every w ∈ Z ∗ factorsuniquely w = h1h2 · · · hk with Hall words h1 ≥ h2 ≥ . . . ≥ hk
Naturally extending ξ and ζ to A(Z ),
ξw = 1r1!···rk !ξ
x r1h1
x ξx r2h2
x . . . x ξx rkhk
(known)
ζw = π′|r |(ξw ) = 1r1!···rk !ζ
x r1h1
x . . . x ζx rkhk
(new)
In general the maps π′k do not commute with shuffles.But surprisingly they do on products of this special form.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Formulas for the ζh
Theorem[E. Gehrig].On every Hall set ζ = π′1 ◦ ξ.
Moreover: (recall PBW basis and Viennot) every w ∈ Z ∗ factorsuniquely w = h1h2 · · · hk with Hall words h1 ≥ h2 ≥ . . . ≥ hk
Naturally extending ξ and ζ to A(Z ),
ξw = 1r1!···rk !ξ
x r1h1
x ξx r2h2
x . . . x ξx rkhk
(known)
ζw = π′|r |(ξw ) = 1r1!···rk !ζ
x r1h1
x . . . x ζx rkhk
(new)
In general the maps π′k do not commute with shuffles.But surprisingly they do on products of this special form.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Sample calculation π′1(aaabb) and ζabaab
π′1(aaabb) = aaabb − 12 (aaab x b + aaa x bb + aa x aab
+a x aabb) + 13(aaa x b x b + aa x ab x b
+2aa x a x bb + a x aab x b + a x a x abb)
−14(aa x a x b x b + a x aa x b x b + a x a x ab x b
+a x a x a x bb) + 15a x a x a x b x b.
Similarly, starting fromξabaab = 1
2 (ab ∗ (a ∗ (a ∗ b))) = 3aaabb + 2aabab + abaab
ζabaab = π′1(ξabaab) = π′1(3aaabb + 2aabab + abaab)
= 110aaabb + 1
15aabab + 115aabba + 1
15abaab − 110ababa
+ 115abbaa− 1
10baaab + 115baaba + 1
15babaa− 110bbaaa.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
The first 14 terms as nilpotent cascade system I
Iterated integrals zH(t ,U ′) = µH Υ(ζH)(U)(t)
za = ua
zb = ub
zab = −16zbua + 1
6zaub
zaab =(−1
2zab − 112zazb
)ua + 1
12z2aub
zbab = − 112z2
bua +(−1
2zab + 112zazb
)ub
zaaab =(−1
2zaab − 112zabza
)ua
zbaab =(−1
2zbab − 112zabzb
)ua −
(12zaab + 1
12zabza)
ub
zbbab =(−1
2zbab − 112zabzb
)ub
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
The first 14 terms as nilpotent cascade system II
zaaaab =(−1
2zaaab − 112zazaab − 1
720z3azb)
ua − 1720z4
aub
zbbbab = 1720z4
bua −(1
2zbbab + 112zbzbab + 1
720zaz3b
)ub
zbaaab =( 1
240z2ay2
b −12zbaab − 1
12zbzaab − 112zazbab
)ua
−(1
2zaaab + 112zazaab + 1
240z3azb)
ub
zabaab =(−1
2zbaab + 112zbzaab − 1
12zazbab − 112y2
ab
+ 1360z2
ay2b)
ua +(−1
6zazaab − 1360z3
azb)
ub
zabbab =(−1
2zbbab1
12zbzbab + 1720zaz3
b
)ua
−(1
6zazbab + 112y2
ab + 1720z2
ay2b)
ub
zbbaab =(−1
2zbbab − 112zbzbab + 1
240zaz3b
)ua
−(1
2zbaab + 112zbzaab + 1
12zazbab + 1240z2
ay2b)
ub
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Summary and outlook
• Formulas to obtain coord’s of the first kind ζhfrom known formulas for ξh in terms of Zinbiel product
• The main contribution is not the existence of the map,but expressing it via elementary objects in Hopf algebra.
• Formulas for ζh are explicit and easily computable,and amenable for implementation for path planning,or for geometric integration procedures.
• Combinatorial Hopf algebra: Connections betweennonlinear control, geometric numerical integration,stochastic DEs (rough paths), more to come.Next: geometric, no letters, but use full tensor algebra
For more information, visit http://math.asu.edu/~kawski
Thank you.
This presentation was prepared with LaTeX using the beamer package.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Summary and outlook
• Formulas to obtain coord’s of the first kind ζhfrom known formulas for ξh in terms of Zinbiel product
• The main contribution is not the existence of the map,but expressing it via elementary objects in Hopf algebra.
• Formulas for ζh are explicit and easily computable,and amenable for implementation for path planning,or for geometric integration procedures.
• Combinatorial Hopf algebra: Connections betweennonlinear control, geometric numerical integration,stochastic DEs (rough paths), more to come.Next: geometric, no letters, but use full tensor algebra
For more information, visit http://math.asu.edu/~kawski
Thank you.
This presentation was prepared with LaTeX using the beamer package.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Summary and outlook
• Formulas to obtain coord’s of the first kind ζhfrom known formulas for ξh in terms of Zinbiel product
• The main contribution is not the existence of the map,but expressing it via elementary objects in Hopf algebra.
• Formulas for ζh are explicit and easily computable,and amenable for implementation for path planning,or for geometric integration procedures.
• Combinatorial Hopf algebra: Connections betweennonlinear control, geometric numerical integration,stochastic DEs (rough paths), more to come.Next: geometric, no letters, but use full tensor algebra
For more information, visit http://math.asu.edu/~kawski
Thank you.
This presentation was prepared with LaTeX using the beamer package.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Summary and outlook
• Formulas to obtain coord’s of the first kind ζhfrom known formulas for ξh in terms of Zinbiel product
• The main contribution is not the existence of the map,but expressing it via elementary objects in Hopf algebra.
• Formulas for ζh are explicit and easily computable,and amenable for implementation for path planning,or for geometric integration procedures.
• Combinatorial Hopf algebra: Connections betweennonlinear control, geometric numerical integration,stochastic DEs (rough paths), more to come.Next: geometric, no letters, but use full tensor algebra
For more information, visit http://math.asu.edu/~kawski
Thank you.
This presentation was prepared with LaTeX using the beamer package.
Intro and main Background, Zinbiel, and ξ Hopf algebra, convolution ζh and examples Conclusion
Summary and outlook
• Formulas to obtain coord’s of the first kind ζhfrom known formulas for ξh in terms of Zinbiel product
• The main contribution is not the existence of the map,but expressing it via elementary objects in Hopf algebra.
• Formulas for ζh are explicit and easily computable,and amenable for implementation for path planning,or for geometric integration procedures.
• Combinatorial Hopf algebra: Connections betweennonlinear control, geometric numerical integration,stochastic DEs (rough paths), more to come.Next: geometric, no letters, but use full tensor algebra
For more information, visit http://math.asu.edu/~kawski
Thank you. This presentation was prepared with LaTeX using the beamer package.