PhyslibSearch

Physlib.Relativity.Tensors.Contraction.Products

11 declarations

theorem

dropPairEmbi,j(m)=m\text{dropPairEmb}_{i,j}(m) = m for m<im < i and m<jm < j

#dropPairEmb_apply_lt_lt

Given a natural number nn, let ii and jj be two distinct indices in Fin(n+2)\text{Fin}(n+2). For any index mFin(n)m \in \text{Fin}(n), if m<im < i and m<jm < j, then the embedding map dropPairEmb(i,j)\text{dropPairEmb}(i, j) (which maps indices from Fin(n)\text{Fin}(n) to Fin(n+2)\text{Fin}(n+2) by skipping positions ii and jj) satisfies dropPairEmb(i,j)(m)=m\text{dropPairEmb}(i, j)(m) = m.

theorem

dropPairEmbn1+i,n1+j(m)=m\text{dropPairEmb}_{n_1+i, n_1+j}(m) = m for m<n1m < n_1

#dropPairEmb_natAdd_apply_castAdd

Given natural numbers nn and n1n_1, let ii and jj be distinct indices in Fin(n+2)\text{Fin}(n+2). Let dropPairEmb\text{dropPairEmb} be the order-preserving embedding from Fin(n1+n)\text{Fin}(n_1 + n) to Fin(n1+n+2)\text{Fin}(n_1 + n + 2) that skips the indices n1+in_1 + i and n1+jn_1 + j. For any index mFin(n1)m \in \text{Fin}(n_1), when mm is viewed as an element of Fin(n1+n)\text{Fin}(n_1 + n), the embedding satisfies dropPairEmb(n1+i,n1+j)(m)=m\text{dropPairEmb}(n_1 + i, n_1 + j)(m) = m, where the result is viewed as an element of Fin(n1+n+2)\text{Fin}(n_1 + n + 2).

theorem

dropPairEmbn1+i,n1+j(range(castAdd))={kk<n1}\text{dropPairEmb}_{n_1+i, n_1+j}(\text{range}(\text{castAdd})) = \{k \mid k < n_1\}

#dropPairEmb_natAdd_image_range_castAdd

Let nn and n1n_1 be natural numbers, and let ii and jj be distinct indices in Fin(n+2)\text{Fin}(n+2). Let dropPairEmbn1+i,n1+j\text{dropPairEmb}_{n_1+i, n_1+j} be the order-preserving embedding from Fin(n1+n)\text{Fin}(n_1 + n) to Fin(n1+n+2)\text{Fin}(n_1 + n + 2) that skips the indices n1+in_1 + i and n1+jn_1 + j. The image of the range of the inclusion map Fin(n1)Fin(n1+n)\text{Fin}(n_1) \hookrightarrow \text{Fin}(n_1 + n) under this embedding is the set of indices in Fin(n1+n+2)\text{Fin}(n_1 + n + 2) whose value is strictly less than n1n_1.

theorem

dropPairEmb\text{dropPairEmb} commutes with index shifting natAdd\text{natAdd}

#dropPairEmb_comm_natAdd

Let nn and n1n_1 be natural numbers, and let ii and jj be distinct indices in Fin(n+2)\text{Fin}(n+2). Let mm be an index in Fin(n)\text{Fin}(n). Let dropPairEmbi,j:Fin(n)Fin(n+2)\text{dropPairEmb}_{i, j} : \text{Fin}(n) \hookrightarrow \text{Fin}(n+2) be the order-preserving embedding that skips the indices ii and jj. Let natAddn1\text{natAdd}_{n_1} denote the map that shifts an index by n1n_1 (i.e., natAddn1(x)=n1+x\text{natAdd}_{n_1}(x) = n_1 + x). Then, shifting the index and the skipped pair by n1n_1 commutes with the embedding: dropPairEmbn1+i,n1+j(n1+m)=n1+dropPairEmbi,j(m)\text{dropPairEmb}_{n_1+i, n_1+j}(n_1+m) = n_1 + \text{dropPairEmb}_{i, j}(m)

theorem

Identity Permutation Condition for Concatenated and Skipped Indices of Tensor Products

#dropPairEmb_permCond_prod

Let n,n1n, n_1 be natural numbers. Let c:Fin(n+2)Cc : \text{Fin}(n+2) \to C and c1:Fin(n1)Cc_1 : \text{Fin}(n_1) \to C be index maps. Let i,ji, j be distinct indices in Fin(n+2)\text{Fin}(n+2) satisfying the duality condition S.τ(c(i))=c(j)S.\tau(c(i)) = c(j). Let i=n1+ii' = n_1 + i and j=n1+jj' = n_1 + j be the shifted indices in the concatenated index space Fin(n1+n+2)\text{Fin}(n_1 + n + 2). The identity map id\text{id} satisfies the permutation condition between: 1. The concatenation of c1c_1 and cc, composed with the embedding dropPairEmbi,j\text{dropPairEmb}_{i', j'} that skips the shifted indices. 2. The concatenation of c1c_1 and the map cc restricted to the indices remaining after skipping ii and jj via dropPairEmbi,j\text{dropPairEmb}_{i, j}. Mathematically, this identity holds: (c1 append c)dropPairEmbn1+i,n1+j=c1 append (cdropPairEmbi,j)(c_1 \text{ append } c) \circ \text{dropPairEmb}_{n_1+i, n_1+j} = c_1 \text{ append } (c \circ \text{dropPairEmb}_{i, j})

theorem

contrPCoeff(n1+i,n1+j,p1p)=contrPCoeff(i,j,p)\text{contrPCoeff}(n_1 + i, n_1 + j, p_1 \otimes p) = \text{contrPCoeff}(i, j, p)

#contrPCoeff_natAdd

Let SS be a tensor species over a field kk and set of colors CC with a duality map τ\tau. Let p1p_1 be a pure tensor of rank n1n_1 and pp be a pure tensor of rank n+2n+2. For any two distinct indices i,j{0,,n+1}i, j \in \{0, \dots, n+1\} such that the color of the jj-th component of pp is dual to the color of the ii-th component (i.e., τ(ci)=cj\tau(c_i) = c_j), let p1pp_1 \otimes p denote the tensor product (concatenation) of p1p_1 and pp. Then the contraction coefficient of p1pp_1 \otimes p at the indices shifted by n1n_1 is equal to the contraction coefficient of pp at the original indices: contrPCoeff(n1+i,n1+j,p1p)=contrPCoeff(i,j,p)\text{contrPCoeff}(n_1 + i, n_1 + j, p_1 \otimes p) = \text{contrPCoeff}(i, j, p)

theorem

contrPCoeff(pp1,cast(i),cast(j))=contrPCoeff(p,i,j)\text{contrPCoeff}(p \otimes p_1, \text{cast}(i), \text{cast}(j)) = \text{contrPCoeff}(p, i, j)

#contrPCoeff_castAdd

Let SS be a tensor species over a field kk and a set of colors CC with duality map τ\tau. Let pp be a pure tensor with index map c:Fin(n+2)Cc: \text{Fin}(n+2) \to C, and let p1p_1 be a pure tensor with index map c1:Fin(n1)Cc_1: \text{Fin}(n_1) \to C. Suppose ii and jj are two distinct indices in Fin(n+2)\text{Fin}(n+2) such that the color of the jj-th index is dual to the color of the ii-th index, i.e., τ(c(i))=c(j)\tau(c(i)) = c(j). Let pp1p \otimes p_1 denote the tensor product of pp and p1p_1. Then the contraction coefficient of the product pp1p \otimes p_1 at the indices corresponding to ii and jj (embedded into the first n+2n+2 slots of the combined index set) is equal to the contraction coefficient of pp at indices ii and jj.

theorem

p1dropPairi,j(p)=dropPairn1+i,n1+j(p1p)p_1 \otimes \text{dropPair}_{i, j}(p) = \text{dropPair}_{n_1+i, n_1+j}(p_1 \otimes p)

#prodP_dropPair

Let SS be a tensor species over a field kk and a set of colors CC with a duality map τ\tau. Let p1p_1 be a pure tensor of rank n1n_1 and pp be a pure tensor of rank n+2n+2. Let ii and jj be distinct indices in the index set {0,,n+1}\{0, \dots, n+1\} of pp such that the color of the jj-th component is dual to the color of the ii-th component (i.e., τ(c(i))=c(j)\tau(c(i)) = c(j)). The theorem states that taking the tensor product of p1p_1 with the tensor formed by dropping the components at indices ii and jj from pp is equal to first taking the tensor product p1pp_1 \otimes p and then dropping the components at the shifted indices n1+in_1 + i and n1+jn_1 + j: p1dropPairi,j(p)=permP(id,H,dropPairn1+i,n1+j(p1p))p_1 \otimes \text{dropPair}_{i, j}(p) = \text{permP}(\text{id}, H, \text{dropPair}_{n_1+i, n_1+j}(p_1 \otimes p)) where \otimes denotes the concatenation of pure tensors, dropPair\text{dropPair} denotes the removal of a specific pair of components, and permP\text{permP} denotes the canonical re-indexing (permutation) required to identify the resulting index spaces.

theorem

prodT(p1.toTensor,contrP(i,j,p))=permT(contrP(n1+i,n1+j,p1p))\text{prodT}(p_1.\text{toTensor}, \text{contrP}(i, j, p)) = \text{permT}(\text{contrP}(n_1+i, n_1+j, p_1 \otimes p))

#prodP_contrP_snd

Let SS be a tensor species over a field kk and a set of colors CC with a duality map τ\tau. Let p1p_1 be a pure tensor of rank n1n_1 and pp be a pure tensor of rank n+2n+2. Suppose ii and jj are two distinct indices in the index set {0,,n+1}\{0, \dots, n+1\} of pp such that the color of the ii-th component is dual to the color of the jj-th component (i.e., τ(ci)=cj\tau(c_i) = c_j). The theorem states that the tensor product of the tensor representation of p1p_1 and the contraction of pp at indices ii and jj is equal to the contraction of the concatenated pure tensor product p1pp_1 \otimes p at the shifted indices n1+in_1 + i and n1+jn_1 + j, followed by a canonical re-indexing: prodT(p1.toTensor,contrP(i,j,p))=permT(id,H,contrP(n1+i,n1+j,p1p))\text{prodT}(p_1.\text{toTensor}, \text{contrP}(i, j, p)) = \text{permT}(\text{id}, H, \text{contrP}(n_1+i, n_1+j, p_1 \otimes p)) where n1+in_1+i and n1+jn_1+j represent the indices ii and jj shifted into the second part of the concatenated index space Fin(n1+n+2)\text{Fin}(n_1 + n + 2), and HH is the permutation condition identifying the resulting index maps.

theorem

prodT(t1,contrT(i,j,t))=permT(contrT(n1+i,n1+j,t1t))\text{prodT}(t_1, \text{contrT}(i, j, t)) = \text{permT}(\text{contrT}(n_1+i, n_1+j, t_1 \otimes t))

#prodT_contrT_snd

Let SS be a tensor species over a field kk. Let t1t_1 be a tensor of rank n1n_1 with index color sequence c1c_1, and tt be a tensor of rank n+2n+2 with index color sequence cc. Suppose ii and jj are distinct indices in {0,,n+1}\{0, \dots, n+1\} such that the color of the ii-th component is dual to the color of the jj-th component, satisfying S.τ(c(i))=c(j)S.\tau(c(i)) = c(j). The theorem states that the tensor product of t1t_1 with the contraction of tt at indices ii and jj is equal to the contraction of the concatenated tensor product t1tt_1 \otimes t at the shifted indices n1+in_1 + i and n1+jn_1 + j, followed by a canonical re-indexing: prodT(t1,contrT(i,j,t))=permT(id,H,contrT(n1+i,n1+j,prodT(t1,t)))\text{prodT}(t_1, \text{contrT}(i, j, t)) = \text{permT}(\text{id}, H, \text{contrT}(n_1+i, n_1+j, \text{prodT}(t_1, t))) where n1+in_1+i and n1+jn_1+j represent the original indices ii and jj shifted into the second part of the concatenated index space, and HH is the permutation condition identifying the resulting index maps.

theorem

contrT(n1+i,n1+j,t1t)=permT(id,t1contrT(i,j,t))\text{contrT}(n_1+i, n_1+j, t_1 \otimes t) = \text{permT}(\text{id}, t_1 \otimes \text{contrT}(i, j, t))

#contrT_prodT_snd

Let SS be a tensor species over a field kk. Let t1t_1 be a tensor of rank n1n_1 with index color sequence c1c_1, and tt be a tensor of rank n+2n+2 with index color sequence cc. Suppose ii and jj are distinct indices in {0,,n+1}\{0, \dots, n+1\} such that the color of the ii-th component is dual to the color of the jj-th component, satisfying S.τ(c(i))=c(j)S.\tau(c(i)) = c(j). The theorem states that the contraction of the concatenated tensor product t1tt_1 \otimes t at the shifted indices n1+in_1 + i and n1+jn_1 + j is equal to the tensor product of t1t_1 and the contraction of tt at indices ii and jj, followed by a canonical re-indexing: contrT(n1+i,n1+j,prodT(t1,t))=permT(id,H,prodT(t1,contrT(i,j,t)))\text{contrT}(n_1+i, n_1+j, \text{prodT}(t_1, t)) = \text{permT}(\text{id}, H, \text{prodT}(t_1, \text{contrT}(i, j, t))) where n1+in_1+i and n1+jn_1+j represent the original indices ii and jj of tt shifted into the second part of the concatenated index space, and HH is the permutation condition identifying the resulting index maps.