Definition 4.2 .

Define a relation \approx on V Λ := { ( x ; m ) : x Λ , m k } assign subscript 𝑉 Λ conditional-set 𝑥 𝑚 formulae-sequence 𝑥 Λ 𝑚 superscript 𝑘 V_{\Lambda}:=\{(x;m):x\in\partial\Lambda,m\in\mathbb{N}^{k}\} italic_V start_POSTSUBSCRIPT roman_Λ end_POSTSUBSCRIPT := { ( italic_x ; italic_m ) : italic_x ∈ ∂ roman_Λ , italic_m ∈ blackboard_N start_POSTSUPERSCRIPT italic_k end_POSTSUPERSCRIPT } by: ( x ; m ) ( y ; p ) 𝑥 𝑚 𝑦 𝑝 (x;m)\approx(y;p) ( italic_x ; italic_m ) ≈ ( italic_y ; italic_p ) if and only if

  1. (V1)

    x ( m d ( x ) ) = y ( p d ( y ) ) 𝑥 𝑚 𝑑 𝑥 𝑦 𝑝 𝑑 𝑦 x(m\wedge d(x))=y(p\wedge d(y)) italic_x ( italic_m ∧ italic_d ( italic_x ) ) = italic_y ( italic_p ∧ italic_d ( italic_y ) ) ; and

  2. (V2)

    m - m d ( x ) = p - p d ( y ) 𝑚 𝑚 𝑑 𝑥 𝑝 𝑝 𝑑 𝑦 m-m\wedge d(x)=p-p\wedge d(y) italic_m - italic_m ∧ italic_d ( italic_x ) = italic_p - italic_p ∧ italic_d ( italic_y ) .

Definition 4.3 .

Define a relation similar-to \sim on P Λ := { ( x ; ( m , n ) ) : x Λ , m n k } assign subscript 𝑃 Λ conditional-set 𝑥 𝑚 𝑛 formulae-sequence 𝑥 Λ 𝑚 𝑛 superscript 𝑘 P_{\Lambda}:=\{(x;(m,n)):x\in\partial\Lambda,m\leq n\in\mathbb{N}^{k}\} italic_P start_POSTSUBSCRIPT roman_Λ end_POSTSUBSCRIPT := { ( italic_x ; ( italic_m , italic_n ) ) : italic_x ∈ ∂ roman_Λ , italic_m ≤ italic_n ∈ blackboard_N start_POSTSUPERSCRIPT italic_k end_POSTSUPERSCRIPT } by: ( x ; ( m , n ) ) ( y ; ( p , q ) ) similar-to 𝑥 𝑚 𝑛 𝑦 𝑝 𝑞 (x;(m,n))\sim(y;(p,q)) ( italic_x ; ( italic_m , italic_n ) ) ∼ ( italic_y ; ( italic_p , italic_q ) ) if and only if

  1. (P1)

    x ( m d ( x ) , n d ( x ) ) = y ( p d ( y ) , q d ( y ) ) 𝑥 𝑚 𝑑 𝑥 𝑛 𝑑 𝑥 𝑦 𝑝 𝑑 𝑦 𝑞 𝑑 𝑦 x(m\wedge d(x),n\wedge d(x))=y(p\wedge d(y),q\wedge d(y)) italic_x ( italic_m ∧ italic_d ( italic_x ) , italic_n ∧ italic_d ( italic_x ) ) = italic_y ( italic_p ∧ italic_d ( italic_y ) , italic_q ∧ italic_d ( italic_y ) ) ;

  2. (P2)

    m - m d ( x ) = p - p d ( y ) 𝑚 𝑚 𝑑 𝑥 𝑝 𝑝 𝑑 𝑦 m-m\wedge d(x)=p-p\wedge d(y) italic_m - italic_m ∧ italic_d ( italic_x ) = italic_p - italic_p ∧ italic_d ( italic_y ) ; and

  3. (P3)

    n - m = q - p 𝑛 𝑚 𝑞 𝑝 n-m=q-p italic_n - italic_m = italic_q - italic_p .

Definition 4.5 .

Define r ~ , s ~ : P Λ ~ V Λ ~ : ~ 𝑟 ~ 𝑠 ~ subscript 𝑃 Λ ~ subscript 𝑉 Λ \widetilde{r},\widetilde{s}:\widetilde{P_{\Lambda}}\to\widetilde{V_{\Lambda}} ~ start_ARG italic_r end_ARG , ~ start_ARG italic_s end_ARG : ~ start_ARG italic_P start_POSTSUBSCRIPT roman_Λ end_POSTSUBSCRIPT end_ARG → ~ start_ARG italic_V start_POSTSUBSCRIPT roman_Λ end_POSTSUBSCRIPT end_ARG by:

r ~ ( [ x ; ( m , n ) ] ) = [ x ; m ] and s ~ ( [ x ; ( m , n ) ] ) = [ x , n ] . formulae-sequence ~ 𝑟 𝑥 𝑚 𝑛 𝑥 𝑚 and ~ 𝑠 𝑥 𝑚 𝑛 𝑥 𝑛 \widetilde{r}([x;(m,n)])=[x;m]\qquad\text{and}\qquad\widetilde{s}([x;(m,n)])=[% x,n]. ~ start_ARG italic_r end_ARG ( [ italic_x ; ( italic_m , italic_n ) ] ) = [ italic_x ; italic_m ] and ~ start_ARG italic_s end_ARG ( [ italic_x ; ( italic_m , italic_n ) ] ) = [ italic_x , italic_n ] .

Definition 3.7 .

A Bayesian graphical calculus is called classical if it satisfies the following equivalent conditions:

  1. (a)

    modifiers can move through the Frobenius structure:

    . \raisebox{-8.535827pt}{\psfig{width=276.0pt}}\ . . (47)
  2. (b)

    modifiers are of the form:

    , \raisebox{-5.690551pt}{\psfig{width=126.0pt}}\,, , (48)
Definition 3.14 .

A Bayesian graphical calculus is a Q 1 / 2 1 2 {}_{1/2} start_FLOATSUBSCRIPT 1 / 2 end_FLOATSUBSCRIPT -calculus when modifiers are of the form:

. \psfig{width=68.0pt}\,. . (108)

Definizione 1

- A constraint is said:

  1. 1.

    holonomic, any restriction on the possible configurations [ References ] of the system:

    f ( 𝐪 , t ) = 0 𝑓 𝐪 𝑡 0 f(\mathbf{q},t)=0 italic_f ( bold_q , italic_t ) = 0 (1)

    and it is an integrable relation;

  2. 2.

    nonholonomic, any restriction on the movements possible [ References ] of the system:

    g ( 𝐪 ; 𝐪 ˙ , t ) = 0 𝑔 𝐪 ˙ 𝐪 𝑡 0 g(\mathbf{q};\dot{\mathbf{q}},t)=0 italic_g ( bold_q ; ˙ start_ARG bold_q end_ARG , italic_t ) = 0 (2)

    and it is not an integrable relation.

If the nonholonomic constraints represents a holonomic constraint, then it is integrable.


Definition 1

A kei or involutory quandle is a set X 𝑋 X italic_X with a binary operation \triangleright satisfying for all x , y , z X 𝑥 𝑦 𝑧 𝑋 x,y,z\in X italic_x , italic_y , italic_z ∈ italic_X


Definition 4.33 .

A quasigroup ( Q , ) 𝑄 (Q,\circ) ( italic_Q , ∘ ) is an ( α , β , γ ) 𝛼 𝛽 𝛾 (\alpha,\beta,\gamma) ( italic_α , italic_β , italic_γ ) -inverse quasigroup if there exist permutations α , β , γ 𝛼 𝛽 𝛾 \alpha,\beta,\gamma italic_α , italic_β , italic_γ of the set Q 𝑄 Q italic_Q such that

α ( x y ) β x = γ y 𝛼 𝑥 𝑦 𝛽 𝑥 𝛾 𝑦 \alpha(x\circ y)\circ\beta x=\gamma y italic_α ( italic_x ∘ italic_y ) ∘ italic_β italic_x = italic_γ italic_y (50)

for all x , y Q 𝑥 𝑦 𝑄 x,y\in Q italic_x , italic_y ∈ italic_Q [ 51 , 7 , 50 , 96 ] .


Definition 3 (Channel)

The channel outputs the transmitted vector, corrupted by independent and identically distributed (i.i.d.) Gaussian noise:

𝐲 = 𝐱 + 𝐳 𝐲 𝐱 𝐳 \displaystyle\mathbf{y}=\mathbf{x}+\mathbf{z} bold_y = bold_x + bold_z (6)

where 𝐳 𝒩 ( 𝟎 , N 𝐈 n × n ) similar-to 𝐳 𝒩 0 𝑁 superscript 𝐈 𝑛 𝑛 \mathbf{z}\sim\mathcal{N}(\mathbf{0},N\mathbf{I}^{n\times n}) bold_z ∼ caligraphic_N ( bold_0 , italic_N bold_I start_POSTSUPERSCRIPT italic_n × italic_n end_POSTSUPERSCRIPT ) for some noise variance N > 0 𝑁 0 N>0 italic_N > 0 .


Definition 3.9 .

A ternary function g 𝑔 g italic_g is said to be cyclically commutative if it is invariant under the cyclic permutation of variables, i.e.

g ( x , y , z ) g ( y , z , x ) g ( z , x , y ) . 𝑔 𝑥 𝑦 𝑧 𝑔 𝑦 𝑧 𝑥 𝑔 𝑧 𝑥 𝑦 g(x,y,z)\approx g(y,z,x)\approx g(z,x,y). italic_g ( italic_x , italic_y , italic_z ) ≈ italic_g ( italic_y , italic_z , italic_x ) ≈ italic_g ( italic_z , italic_x , italic_y ) .

Definition 5.1 .

A deformation retract datum of linear factorisations of W 𝑊 W italic_W consists of

\xymatrix @ C + 2 p c ( L , b ) \ar @ < - 0.8 e x > [ r ] i & ( M , b ) , \ar @ < - 0.8 e x > [ l ] p h \xymatrix @ 𝐶 2 𝑝 𝑐 𝐿 𝑏 \ar @ expectation 0.8 𝑒 𝑥 subscript delimited-[] 𝑟 𝑖 & 𝑀 𝑏 \ar @ expectation 0.8 𝑒 𝑥 subscript delimited-[] 𝑙 𝑝 \xymatrix@C+2pc{(L,b)\ar@<-0.8ex>[r]_{i}&(M,b),\ar@<-0.8ex>[l]_{p}}\quad h @ italic_C + 2 italic_p italic_c ( italic_L , italic_b ) @ < - 0.8 italic_e italic_x > [ italic_r ] start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT & ( italic_M , italic_b ) , @ < - 0.8 italic_e italic_x > [ italic_l ] start_POSTSUBSCRIPT italic_p end_POSTSUBSCRIPT italic_h

where ( L , b ) 𝐿 𝑏 (L,b) ( italic_L , italic_b ) and ( M , b ) 𝑀 𝑏 (M,b) ( italic_M , italic_b ) are linear factorisations of W 𝑊 W italic_W , p 𝑝 p italic_p and i 𝑖 i italic_i are morphisms of factorisations, and h h italic_h is a degree one map M M 𝑀 𝑀 M\longrightarrow M italic_M ⟶ italic_M , which together satisfy the following two conditions:

Notice that in particular p 𝑝 p italic_p is a homotopy equivalence with inverse i 𝑖 i italic_i .


Definition 1

Let f : V { 1 , 2 } normal-: 𝑓 normal-→ 𝑉 1 2 f:V\rightarrow\{1,2\} italic_f : italic_V → { 1 , 2 }

f ( v ) = { 1 , if v belongs to a component of type (a) or (b) , 2 , if v belongs to a component of type (c) , 𝑓 𝑣 cases 1 if 𝑣 belongs to a component of type (a) or (b) 2 if 𝑣 belongs to a component of type (c) f(v)=\left\{\begin{array}[]{cl}1,&\mbox{ if }v\mbox{ belongs to a component of% type (a) or (b)},\\ 2,&\mbox{ if }v\mbox{ belongs to a component of type (c)},\\ \end{array}\right. italic_f ( italic_v ) = { start_ARRAY start_ROW start_CELL 1 , end_CELL start_CELL if italic_v belongs to a component of type (a) or (b) , end_CELL end_ROW start_ROW start_CELL 2 , end_CELL start_CELL if italic_v belongs to a component of type (c) , end_CELL end_ROW end_ARRAY

Definition 1.3.6

We say a sesquiad A 𝐴 A italic_A has no zero divisors , if

a b = 0 ( a = 0 or b = 0 ) . fragments a b 0 italic- italic- fragments ( a 0 italic- or italic- b 0 ) . ab=0\quad\Rightarrow\quad\left(a=0\ \ \ \text{or}\ \ \ b=0\right). italic_a italic_b = 0 ⇒ ( italic_a = 0 or italic_b = 0 ) .

We call a sesquiad integral , if 1 0 1 0 1\neq 0 1 ≠ 0 and

a f = b f ( a = b or f = 0 ) . fragments a f b f italic- italic- fragments ( a b italic- or italic- f 0 ) . af=bf\quad\Rightarrow\quad\left(a=b\ \ \ \text{or}\ \ \ f=0\right). italic_a italic_f = italic_b italic_f ⇒ ( italic_a = italic_b or italic_f = 0 ) .

An integral sesquiad has no zero divisors, but the converse does not hold in general. A subsesquiad of an integral domain is an integral sesquiad.


Definition 1.2

We set:

q = ln ρ . 𝑞 𝜌 q=\ln\rho. italic_q = roman_ln italic_ρ .

Definition 2 .

A Reed-Solomon-Code 𝒢 𝒮 ( q ; n , k , 𝐯 ) 𝒢 𝒮 𝑞 𝑛 𝑘 𝐯 \mathcal{GRS}(q;n,k,\boldsymbol{v}) caligraphic_G caligraphic_R caligraphic_S ( italic_q ; italic_n , italic_k , bold_italic_v ) with length n = q - 1 𝑛 𝑞 1 n=q-1 italic_n = italic_q - 1 and

𝒗 = ( α 0 , α 1 , , α q - 2 ) 𝒗 superscript 𝛼 0 superscript 𝛼 1 superscript 𝛼 𝑞 2 \boldsymbol{v}=(\alpha^{0},\alpha^{1},\ldots,\alpha^{q-2}) bold_italic_v = ( italic_α start_POSTSUPERSCRIPT 0 end_POSTSUPERSCRIPT , italic_α start_POSTSUPERSCRIPT 1 end_POSTSUPERSCRIPT , … , italic_α start_POSTSUPERSCRIPT italic_q - 2 end_POSTSUPERSCRIPT )

with α 𝛼 \alpha italic_α being a primitive element of 𝔽 𝔽 \mathbb{F} blackboard_F will be referred to as 𝒮 ( q - 1 , k ) 𝒮 𝑞 1 𝑘 \mathcal{RS}(q-1,k) caligraphic_R caligraphic_S ( italic_q - 1 , italic_k ) .
The extended code of length n = q 𝑛 𝑞 n=q italic_n = italic_q obtained by adding the zero element of 𝔽 𝔽 \mathbb{F} blackboard_F to the vector 𝐯 𝐯 \boldsymbol{v} bold_italic_v of 𝒮 ( q - 1 , k ) 𝒮 𝑞 1 𝑘 \mathcal{RS}(q-1,k) caligraphic_R caligraphic_S ( italic_q - 1 , italic_k ) , i.e.

𝒗 = ( 0 , α 0 , α 1 , , α q - 2 ) , 𝒗 0 superscript 𝛼 0 superscript 𝛼 1 superscript 𝛼 𝑞 2 \boldsymbol{v}=(0,\alpha^{0},\alpha^{1},\ldots,\alpha^{q-2}), bold_italic_v = ( 0 , italic_α start_POSTSUPERSCRIPT 0 end_POSTSUPERSCRIPT , italic_α start_POSTSUPERSCRIPT 1 end_POSTSUPERSCRIPT , … , italic_α start_POSTSUPERSCRIPT italic_q - 2 end_POSTSUPERSCRIPT ) ,

will be called 𝒮 ( q , k ) superscript 𝒮 normal-∗ 𝑞 𝑘 \mathcal{RS}^{\ast}(q,k) caligraphic_R caligraphic_S start_POSTSUPERSCRIPT ∗ end_POSTSUPERSCRIPT ( italic_q , italic_k ) .


Definition 3 (Distance forest)

Let \mathcal{F} caligraphic_F be a forest whose leaves are enumerated by Q 𝑄 Q italic_Q . Then \mathcal{F} caligraphic_F is a distance forest for M 𝑀 M italic_M if for every q , p Q 𝑞 𝑝 𝑄 q,p\in Q italic_q , italic_p ∈ italic_Q we have

d ( q , p ) = { level ( lca ( q , p ) ) if tree ( q ) = tree ( p ) , otherwise . 𝑑 𝑞 𝑝 cases level lca 𝑞 𝑝 if tree 𝑞 tree 𝑝 otherwise d(q,p)=\begin{cases}\operatorname{level}(\operatorname{lca}(q,p))&\text{ if }% \mathrm{tree}(q)=\mathrm{tree}(p),\\ \infty&\text{ otherwise}.\end{cases} italic_d ( italic_q , italic_p ) = { start_ROW start_CELL roman_level ( roman_lca ( italic_q , italic_p ) ) end_CELL start_CELL if roman_tree ( italic_q ) = roman_tree ( italic_p ) , end_CELL end_ROW start_ROW start_CELL ∞ end_CELL start_CELL otherwise . end_CELL end_ROW

Definition 4 (Quartet or Four-Point Condition)

The pairwise distances { l ( i , j ) } i , j { a , b , u , v } subscript 𝑙 𝑖 𝑗 𝑖 𝑗 𝑎 𝑏 𝑢 𝑣 \{l(i,j)\}_{i,j\in\{a,b,u,v\}} { italic_l ( italic_i , italic_j ) } start_POSTSUBSCRIPT italic_i , italic_j ∈ { italic_a , italic_b , italic_u , italic_v } end_POSTSUBSCRIPT for the configuration in Fig. 4 satisfy

l ( a , u ) + l ( b , v ) = l ( b , u ) + l ( a , v ) , 𝑙 𝑎 𝑢 𝑙 𝑏 𝑣 𝑙 𝑏 𝑢 𝑙 𝑎 𝑣 l(a,u)+l(b,v)=l(b,u)+l(a,v), italic_l ( italic_a , italic_u ) + italic_l ( italic_b , italic_v ) = italic_l ( italic_b , italic_u ) + italic_l ( italic_a , italic_v ) , (6)

and the configuration is denoted by Q ( a b | u v ) 𝑄 conditional 𝑎 𝑏 𝑢 𝑣 Q(ab|uv) italic_Q ( italic_a italic_b | italic_u italic_v ) .