On the flag geometry of simple group of Lie type and multivariate cryptography

We propose some multivariate cryptosystems based on finite BN-pair G defined over the fields Fq. We convert the adjacency graph for maximal flags of the geometry of group G into a finite Tits automaton by special colouring of arrows and treat the largest Schubert cell Sch isomorphic to vector space...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Algebra and Discrete Mathematics
Datum:2015
1. Verfasser: Ustimenko, V.
Format: Artikel
Sprache:English
Veröffentlicht: Інститут прикладної математики і механіки НАН України 2015
Online Zugang:https://nasplib.isofts.kiev.ua/handle/123456789/152794
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Zitieren:On the flag geometry of simple group of Lie type and multivariate cryptography / V. Ustimenko // Algebra and Discrete Mathematics. — 2015. — Vol. 19, № 1. — С. 130-144. — Бібліогр.: 18 назв. — англ.

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id nasplib_isofts_kiev_ua-123456789-152794
record_format dspace
spelling Ustimenko, V.
2019-06-12T21:06:52Z
2019-06-12T21:06:52Z
2015
On the flag geometry of simple group of Lie type and multivariate cryptography / V. Ustimenko // Algebra and Discrete Mathematics. — 2015. — Vol. 19, № 1. — С. 130-144. — Бібліогр.: 18 назв. — англ.
1726-3255
https://nasplib.isofts.kiev.ua/handle/123456789/152794
We propose some multivariate cryptosystems based on finite BN-pair G defined over the fields Fq. We convert the adjacency graph for maximal flags of the geometry of group G into a finite Tits automaton by special colouring of arrows and treat the largest Schubert cell Sch isomorphic to vector space over Fq on this variety as a totality of possible initial states and a totality of accepting states at a time. The computation (encryption map) corresponds to some walk in the graph with the starting and ending points in Sch. To make algorithms fast we will use the embedding of geometry for G into Borel subalgebra of corresponding Lie algebra. We also consider the notion of symbolic Tits automata. The symbolic initial state is a string of variables tα∈Fq, where roots α are listed according Bruhat's order, choice of label will be governed by special multivariate expressions in variables tα, where α is a simple root. Deformations of such nonlinear map by two special elements of affine group acting on the plainspace can produce a computable in polynomial time nonlinear transformation. The information on adjacency graph, list of multivariate governing functions will define invertible decomposition of encryption multivariate function. It forms a private key which allows the owner of a public key to decrypt a ciphertext formed by a public user. We also estimate a polynomial time needed for the generation of a public rule.
The paper is dedicated to the memory of my teacher Lev Kaluzhnin.His wide mathematical vision brought us, participants of his Seminar, thelight of Modern Algebra.
en
Інститут прикладної математики і механіки НАН України
Algebra and Discrete Mathematics
On the flag geometry of simple group of Lie type and multivariate cryptography
Article
published earlier
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
title On the flag geometry of simple group of Lie type and multivariate cryptography
spellingShingle On the flag geometry of simple group of Lie type and multivariate cryptography
Ustimenko, V.
title_short On the flag geometry of simple group of Lie type and multivariate cryptography
title_full On the flag geometry of simple group of Lie type and multivariate cryptography
title_fullStr On the flag geometry of simple group of Lie type and multivariate cryptography
title_full_unstemmed On the flag geometry of simple group of Lie type and multivariate cryptography
title_sort on the flag geometry of simple group of lie type and multivariate cryptography
author Ustimenko, V.
author_facet Ustimenko, V.
publishDate 2015
language English
container_title Algebra and Discrete Mathematics
publisher Інститут прикладної математики і механіки НАН України
format Article
description We propose some multivariate cryptosystems based on finite BN-pair G defined over the fields Fq. We convert the adjacency graph for maximal flags of the geometry of group G into a finite Tits automaton by special colouring of arrows and treat the largest Schubert cell Sch isomorphic to vector space over Fq on this variety as a totality of possible initial states and a totality of accepting states at a time. The computation (encryption map) corresponds to some walk in the graph with the starting and ending points in Sch. To make algorithms fast we will use the embedding of geometry for G into Borel subalgebra of corresponding Lie algebra. We also consider the notion of symbolic Tits automata. The symbolic initial state is a string of variables tα∈Fq, where roots α are listed according Bruhat's order, choice of label will be governed by special multivariate expressions in variables tα, where α is a simple root. Deformations of such nonlinear map by two special elements of affine group acting on the plainspace can produce a computable in polynomial time nonlinear transformation. The information on adjacency graph, list of multivariate governing functions will define invertible decomposition of encryption multivariate function. It forms a private key which allows the owner of a public key to decrypt a ciphertext formed by a public user. We also estimate a polynomial time needed for the generation of a public rule.
issn 1726-3255
url https://nasplib.isofts.kiev.ua/handle/123456789/152794
citation_txt On the flag geometry of simple group of Lie type and multivariate cryptography / V. Ustimenko // Algebra and Discrete Mathematics. — 2015. — Vol. 19, № 1. — С. 130-144. — Бібліогр.: 18 назв. — англ.
work_keys_str_mv AT ustimenkov ontheflaggeometryofsimplegroupoflietypeandmultivariatecryptography
first_indexed 2025-11-24T21:46:39Z
last_indexed 2025-11-24T21:46:39Z
_version_ 1850498513455022080
fulltext Algebra and Discrete Mathematics RESEARCH ARTICLE Volume 19 (2015). Number 1, pp. 130–144 © Journal “Algebra and Discrete Mathematics” On the flag geometry of simple group of Lie type and multivariate cryptography Vasyl Ustimenko Communicated by V. I. Sushchansky Abstract. We propose some multivariate cryptosystems based on finite BN -pair G defined over the fields Fq. We convert the adjacency graph for maximal flags of the geometry of group G into a finite Tits automaton by special colouring of arrows and treat the largest Schubert cell Sch isomorphic to vector space over Fq on this variety as a totality of possible initial states and a totality of accepting states at a time. The computation (encryption map) corresponds to some walk in the graph with the starting and ending points in Sch. To make algorithms fast we will use the embedding of geometry for G into Borel subalgebra of corresponding Lie algebra. We also consider the notion of symbolic Tits automata. The symbolic initial state is a string of variables tα ∈ Fq, where roots α are listed according Bruhat’s order, choice of label will be governed by special multivariate expressions in variables tα, where α is a simple root. Deformations of such nonlinear map by two special elements of affine group acting on the plainspace can produce a computable in polynomial time nonlinear transformation. The information on adjacency graph, list of multivariate governing functions will define invertible decomposition of encryption multivariate function. It forms a private key which allows the owner of a public key to decrypt a ciphertext formed by a public user. We also estimate a polynomial time needed for the generation of a public rule. Key words and phrases: Multivariate Cryptography, flag variety, Geometry of Simple Group of Lie type, Schubert cell, symbolic walks. V. Ustimenko 131 1. Introduction Linear codes over finite field Fq are subspaces of Fq n. The subspaces form a finite projective geometry of dimension n−1, which is a very impor- tant object of Pure Mathematics and Classical Coding Theory. Since late 80th some other applications of Finite Projective Geometry to Information Security have appeared (see [3], [4], devoted to Network Coding). In par- ticular, it was used in Cryptography ( see [5], where projective geometry were used for audentification protocols, or [6], [7], where it was used for the symmetric encryption and key exchange protocols respectively). Finite geometries now are widely used as tools for secret sharing. In current article we will introduce a new public key algorithms based on finite Lie geometries of normal and twisted type over the Coxeter Dynkin diagrams An, Bn, Cn and Dn. Public key algorithm of Multivariate Cryptography [1] based on finite projective geometry already was introduced recently at the conference Algebraic Combinatoric and Applications 2015 (ALCOMA 2015, Kloster Banz, Germany , see the program and book of abstracts on its web). In this paper we are going to present more general cryptosystem corresponding to the geometry of a finite simple group of rank n > 2. In section 2 we present the concepts of a family of multivariate maps with polynomially invertible decomposition and its applications to cryp- tography. In section 3 symbolic Tits automaton for the case of variety of maximal flags of the geometry of finite simple group will be defined. Its computa- tions are polynomial maps of the vector space Fq N , where N is a number of positive roots in corresponding root system. In fact, some special Tits automata over various diagrams have been used for the construction of key exchange protocols [7]. In section 3 we also consider the invertibility conditions for the map produced by Tits automaton. We show that the polynomial map produced by Tits automaton can be treated as the polynomial map with invertible decomposition. In section 4 we consider an interpretation of the geometry of simple Lie group of normal type and its flag system in terms of linear algebra. We will use the approach suggested in [9], [10], [11] for the description of the geometry of Shevalley group and the work of corresponding Tits automaton. Conclusions are in section 5. The complexity estimates for the public key algorithm based on flag variety of geometry of Simple Group of Lie type are presented there. 132 On the flag geometry of simple group of Lie type 2. On the general scheme for a cryptosystem based on a family of multivariate maps with an invertible decomposition Multivariate cryptography (see [1]) is one of the directions of Postquan- tum Cryptography, which concerns with algorithms resistant to hypothetic attacks conducted by Quantum Computer. The encryption tools of Multi- variate Cryptography are nonlinear multivariate transformations of affine space Kn, where K is a finite commutative ring. Recall that affine Cremona group C(Kn) is a totality of invertible maps f of affine space Kn over a Commutative ring K into itself, such that the inverse map f−1 is also a polynomial one. Let us refer to the sequence of maps f(n) from C(Kn), n = 1, 2, . . . as a family of polynomial degree if the degree of each transformation is a parameter s of the size O(nt). We say that a family f(n) ∈ C(Kn) has a polynomially invert- ible decomposition if f(n) can be written as a composition of elements f1(n), f2(n), . . . , fk(n)(n) and the knowledge on this decomposition will allow us to compute the value of y = f(x) and the re-image of given y in polynomial time k(n)O(nd) (see [2]). We can create a new encryption map h(n) as a deformation τLf(n)τR of f(n) with special invertible affine transformations τL and τR of Kn. If the family h(n) on a symbolic level is presentable in a computable form and the value of h(n) in given point of affine space is computable in polynomial time, then h(n) is also a map with polynomially invertible decomposition. Let f(n) ∈ C(Kn)) be a family of transformations with polynomially invertible decomposition. Then the following public key can be imple- mented. Alice chooses parameter n. She knows the decomposition f(n) = f(n, 1) f(n, 2) . . . f(n, k(n)). Notice that some transformations f(n, i) can be not a bijective. Additionally she chooses an invertible monomial linear transformations τL of kind xi → λixπ(i), where π is a permutation on the set {1, 2, . . . , n}. Alice also takes a bijective affine transformation τR of kind x → xA + b, where A is a non-singular matrix. She computes the transformation G = τLf(n)τR in affine Cremona group and writes it in the standard form xi → gi(x1, x2, . . . , xn), i = 1, 2, . . . , n. Assume that public rules gi are governed by the lists of monomial terms, written in some chosen order. V. Ustimenko 133 Notice, that the left application of τL does not change the number of monomial terms. The right application of τR can increase the number of monomial terms in n times. So the density den(τLf(n)τR) is O(nd+1). A public user (Bob) gets symbolic transformation G = G(n) in the form of public rule xi → gi(x1, x2, . . . , xn). He can encrypt by computation of the value of G on his plaintext (x1, x2, . . . , xn) in O(ns+d+1). Alice keeps the decomposition f(n) = f(n, 1)f(n, 2) . . . f(n, k(n)) as a deep secret. It allows her to decrypt Bob’s ciphertext for k(n)O(nt+3) elementary steps, because applications of known τL −1 and τR −1 cost O(n) and O(n2), respectively. If Bob does not have an additional information on the transformation G then he has a difficult task of computation G−1. 3. Symbolic Tits automata and bijective multivariate maps Below we consider an abstract scheme for cryptosystem for special class of geometrical maps corresponding to the geometry of finite simple group G = Xn(q) of rank n > 2. In this article we assume that Xn(q) is a Shevalley group (in other terminology a simple group of normal type). This group corresponds to Coxeter-Dynkin diagram Xn which also defines Weyl group W corresponding to root system Ω. Let α1, α2, . . . , αn be the list of simple roots, Ω+ and Ω− correspond to the list of sets of positive and negative roots. We assume that |Ω+| = |Ω−| = N . For each β ∈ Ω subgroup Uβ =< xβ(t)|t ∈ Fq > is isomorphic to additive group Fq (see [12]). Group G is generated by Uβ , β ∈ Ω, its unipotent subgroup U =< Uβ|β ∈ Ω+ > is a Sylow p subgroup of Xn(q), q = pm for prime integer p. In fact, |U | = qN . Let P1, P2, . . . , Pn be the list of maximal parabolic subgroups of G corresponding to nodes of diagram Xn, then the Borel subgroup P1 ∩ P2 ∩ . . . Pn coincides with the normaliser NG(U) of the unipotent subgroup U . The geometry of G is a totality Γ(G) of all left cosets gPi, where g ∈ G, 1 6 i 6 n with the type function t, such that t(gPi) = i), and symmetric and irreflexive incidence relation I, such that a, b ∈ Γ(G) if and only if t(a) 6= t(b) and the intersection of left cosets is nonempty. Let Γi(G) = {a ∈ Γ(G)|t(a) = i}. We refer to a subset F of Γ(G) as a flag of the geometry if each pair of elements from F is incident. We assume that t(F ) = {t(a)|a ∈ F}. In case of maximal flag t(F ) = {1, 2, . . . , n} = M . Let GF (G) be the totality of maximal flags of Γ(G) with the partition onto large Schubert cells, i. e. orbits of unipotent subgroup U on this set. In fact there is the unique largest Schubert cell of size qN on which the unipotant group acts regularly. 134 On the flag geometry of simple group of Lie type Assume that for a flag F its type is t(F ) = M − {i}. The residue Res(F ) = {x ∈ Γi|xIa, a ∈ F}. We assume that Ext(F ) is the totality of all flags of kind F ∪ {x}, x ∈ ResF . In fact |Ext(F )| = |Res(F )| = q + 1. The elements of Ext(F ) are from two different Schubert cells S0 and S1 such that Ext(F ) ∩ S0 and Ext(F ) ∩ S1 have cardinalities 1 and q. If F ∪ {y}, where y ∈ Res(F ) is located in Schubert cell S0 then we join F and maximal flag F ∪ {y} by a directed arrow labeled by pair (i, ∞), where ∞ is a formal symbol. If y ∈ S1, then there is a unique root subgroup Uβ = {xβ(t)|t ∈ Fq} which acts regularly on Ext(F ) ∩ S1. Element z = F ∪{y} from S1 can be identified with corresponding element xr(t) ∈ U and with parameter µ(z) = t ∈ Fq. We join F and z by an arrow with the label (i, µ(z)) in that case. Let F be a maximal flag. For each i ∈ {1, 2, . . . , n} we consider a subflag F i of type M − {i} and join F and representative y from Ext(F i) by an arrow with the label (i, µ(y)), where µ(y) ∈ {∞} ∪ Fq. So we constructed a labeled directed graph T (G) on the set of maximal flags with n(q + 1) output arrows for each vertex, n of them are loops. We convert T (G) to Tits automaton by an announcement that the initial and accepting states of this automaton have to be elements of the largest Schubert cells. If we ignore labels and loops of T (G), this graph can be identified with the directed graph of adjacency relation : two flags are adjacent if their intersection has cardinality n − 1. The action of unipotant subgroup U defines the partition on the large Schubert cells and labeling of T (G). Let F, F ′ be flags joined by an arrow of T (G) labeled by (i, a), i ∈ M ,a ∈ ∞ ∪ Fq. The transition function T i a sends F to F ′ for each pair of flags joined by arrow with the label (i, a). The regular computation of Tits automaton is the composition E of T i1 a1 , T i2 a2 , . . . , T ir ar , where il 6= il+1, l = 1, 2, . . . , r − 1 which sends flag from the largest Schubert cell to another element of this cell. As it follows from the definition of Tits automaton its computation is not a bijective linear map on the vector space of the largest Schubert cell. The existence of the reverse directed walk from the image to the initial state insures that the initial state is uniquely defined by its projection on tangent space and the computation. We also define a Schubert automaton Sch(G) as one obtained from Tits automaton for T (G) by deleting all flags located outside of the largest Schubert cell together with input and output arrows and their labels. The regular action of unipotent subgroup U on the largest Schubert cell Sch(G) allows us to identify Sch(G) with Fq N . V. Ustimenko 135 Finally we define symbolic Tits automaton with the symbolic initial state (t1, t2, . . . , tN ), where ti are variables corresponding to roots from Ω+ and parameters aj 6= ∞ are polynomials in N variables ti1 , ti2 , . . . , tin related to simple roots. The computation of a symbolic Tits automaton in- duces polynomial transformation En of Fq N . The map En corresponds to a ”symbolic walk” Sn,m of length m in An(q). The map En appears together with their decomposition on “symbolic transition functions”. The infinite families of highly nonlinear bijective computable transformations En with polynomially invertible decompositions have been introduced. So the described above (section 1) scheme can be implemented to create a cryp- tosystem in the case of groups of normal type An(q), Bn(q), Cn(q), Dn(q) and twisted groups 2An(q). Symbolic Tits automata of groups of restricted rank E6(q), E7(q), E8(q), F4(q), G2(q), 2E6(q), 2F4(q), q = 22l+1, 2D4(q), 3D4(q) (see [12]) can be used for the construction of S-boxes and hash functions. Special Tits automata can be a base for stream ciphers (see [8]). Let us consider a regular computation of symbolic Tits automaton with the symbolic initial state (t1, t2, . . . , tN ), where ti, i = 1, 2, . . . , N are vari- ables, defined by putting labels from Fq[ti1 , ti2 , . . . , tin ]∪{∞}. The special- ization of symbolic initial state t1 = a1, t2 = a2, . . . , tN = aN produces the computation of Tits automaton with initial state (a1, a2, . . . , aN ) ∈ Fq N . Let e1, e2, . . . , eN be the fixed basis which we use for writing elements of Fq N as tuples. We refer to the symplectic subspace S of all tuples of kind ei1 ,ei2 ,. . . ,ein as a tangent subspace of the largest Schubert cell. The computation in a symbolic Tits automaton sends a symbolic initial state t = (t1, t2, . . . , tN ) into E(t) , where E is a polynomial multivariate map of Fq N into it- self. Tangent subspace is invariant subspace for a transformation E. Let (j1, f1), (j2, f2), . . . , (jk, fk) be the sequence of the labels such that fi ∈ Fq[ti1 , ti2 , . . . , tin ] ∪ ∞ for the computation of a symbolic Tits au- tomaton. We assume that js 6= is+1, s = 1, 2, . . . , k − 1. The computation E is the composition of symbolic transition functions Gi1 fi1 , Gi2 fi2 , . . . , Gim fim . Recall, that the resulting state of our computation is the flag of the largest Schubert cell. Let E′ be the restriction of E on S. Proposition 3.1. The map E is a bijection if and only if its restriction E′ on the tangent space is one to one correspondence. 136 On the flag geometry of simple group of Lie type Proof. Let us assume that bβ, β ∈ Ω+ are coordinates of the image of the vector t. The bijective map E′ can be written in the form hi(tα1 , tα2 , . . . tαn) = bi, i = 1, 2, . . . , n. Theoretically we can solve this system and find tα1 = a1, tα2 = a2, . . . , tαn = an. After that we can compute cj = fj(a1, a2, . . . , an) for j = 1, 2, . . . , m. The map E acts outside of tangent space as tβ → Gi1 c1Gi2 c2 . . . Gim cm(t) for nonsimple root β. It is a linear map depending on N −n variable. Notice that alternatively we can consider the specialisa- tion tα1 = a1, tα2 = a2, . . . , tαn = an of standard presentation of E which exists and write linear equations. We can take a unique solution of system Gi1 c1Gi2 c2 . . . Gim cm(t) = bβ for variables tβ, β ∈ Ω+ − {α1, α2, . . . , αn}. So we get the well defined unique re-image t. The map E corresponds to a symbolic walk of a length k in a symbolic Tits automaton. Proposition 3.2. Let map E be a bijection defined by Tits automaton and it is computable in a given vector from Fq N in a polynomial time. Assume that a tangent space is known and the reimage of E′ in a given point is also computable in a polynomial time. Then the reimage of E in a given point is computable in a polynomial time. Proof. The restriction of map E on symplectic subspace S is the map tαr → ajr [tα1 , tα2 , . . . , tαn ] = t′ αr , r = 1, 2, . . . , n, where jr is the last appearance of r as a first coordinate of label (r, al), al ∈ F [tα1 , tα2 , . . . , tαn ] if such appearance exists. In the opposite case we have simply tαr → tαr . We refer to expressions ajr [tα1 , tα2 , . . . , tαn ] as tangent coordinates of the walk (or computation). The computation of the inverse for E|S in polynomial time will allow us to solve the system of equations ajr [tα1 , tα2 , . . . , tαn ] = t′ αr for tαi = t∗ i, i = 1, 2, . . . , n. It gives us an opportunity to compute parameters as = fis(t∗ 1, t∗ 2, . . . t∗ n−1), s = 1, 2, . . . , m. After that we can consider the computation of Tits automaton Gi1 a1Gi2 a2 . . . Gim am sending vector with coordinates tβ onto the known image of E. Recall that tαi coincides with t∗ i, i = 1, 2, . . . , n. So, we have system of N − n linear equations in N − n variables, which has a unique solution. It is clear that we can solve this system for polynomial time O(N3) and get the re-image of the map E. Notice, that we can get system of linear equations with N − n variable via specialisation tαi = t∗ i, i = 1, 2, . . . , n of polynomial map E. V. Ustimenko 137 Corollary 3.3. Under the conditions of previous statement the map E is a map with the polynomially invertible decomposition E = Gi1 fi1 Gi2 fi2 . . . Gim fim . We say that the computation of Tits automaton is a tame one if it corresponds to symbolic walk given by (j1, f1), (j2, f2), . . . , (jk, fk), fi ∈ Fq[xi1 , xi2 , . . . , xim ] ∪ ∞, for i = 1, 2, . . . , k, js 6= js+1, s = 1, 2, . . . , k − 1, each r ∈ M appears in the list j1, j2, . . . , jk not more than O(1) times and degree of each fi is bounded by O(n) and density O(1). In the last section of the paper we show that the following statement holds. Lemma 3.4. Let En be the map corresponding to tame computation of Tits automaton. Then it has a polynomial density and a polynomial degree bounded by O(n3) and O(n2) respectively. Example. Polynomially invertible En corresponding to the tame com- putation of Tits automaton can be constructed by the following method. We choose the length of the walk m as a linear expressions from n. We assume additionally that the restriction E′ of E onto S is the bijective map tis → bistµ(is) kis + dis , s ∈ {1, 2, . . . , n}, where µ is a permutation on {i1, i2, . . . , in}, and bis , dis are constants from Fq and kis are constant integers such that the equation of kind zkis = b have unique solution. We assume that the field Fq is fixed. It means that the re-image for E′ in given vector can be computed for O(n) elementary steps. Notice that functions fjs which are not a tangent coordinates can be arbitrary polynomials of degree O(n) and density O(1). We refer to such invertible map E′ as invertible tame map. We can use more general maps of kind E′ = τ1Hτ2 where τi(x) = xAi + ci, i = 1, 2 where rows and columns of invertible n×n matrices Ai have finite number of nonzero entries and H is an invertible tame map. Natural example of tame computation is the following. Let E : Fq N → Fq N be the polynomial map induced by a regular tame computation of symbolic Tits automaton, such that different from ∞ labels are “shifted” monomial terms of kind at s1(n) i1 t s2(n) i2 , . . . , t sn(n) in + b, where a and b are constants, parameters si(n), i = 1, 2, . . . , n are linear expressions in n, such that s1(n) + s2(n) + · · · + sn−1(n) is also a linear expression. We choose the length of the walk k as a linear expressions from n. 138 On the flag geometry of simple group of Lie type Remark. Let us consider the deformation Ẽn = τ1Enτ2 of ∆, where τ1 is invertible monomial transformation ti → λitπ(i), λi ∈ F ∗ q , i = 1, 2, . . . , N , π is a permutation on {1, 2, . . . , N}, τ2 is a general bijective affine map of Fq N into itself. For the deformation Ẽ of E the equality deg(Ẽ) = deg(E) holds. Transformations E and τ1E have the same densities. Density of Eτ2 is at most n2 times larger than the density of E. Let us evaluate the complexity of the multivariate cryptosystem based on tame computation of the Tits automaton. Notice that evaluations of degree and density shows that Bob can compute the value of Ẽ in a given point for O(n7) = O(N3+1/2) elementary steps. We can use invertible tame computations defined in the Example above. The time of its generation and the time of the computation of its inverse will be evaluated in the last section. 4. On the interpretation of Lie geometry and Tits automa- ton in terms of linear algebra 4.1. Graphs and incidence system The missing definitions of graph-theoretical concepts which appear in this paper can be found in [13], [14] or [16]. Simple graphs are undirected graphs without loops and multiple edges. Let V (G) and E(G) denote the set of vertices and the set of edges of G, respectively. Then |V (G)| is called the order of G, and |E(G)| is called the size of G. When it is convenient, we shall identify G with the corresponding anti-reflexive binary relation on V (G), i.e. E(G) is a subset of V (G) × V (G) and write vGu for the adjacent vertices u and v (or neighbours). The sequence of distinct vertices v0, v1, . . . , vt, such that viGvi+1 for i = 1, . . . , t − 1 is the pass in the graph. The length of a pass is a number of its edges. The distance dist(u, v) between two vertices is the length of the shortest pass between them. The degree of vertex v is the number of its neighbours. The incidence structure is the set V with partition sets P (points) and L (lines) and symmetric binary relation I such that the incidence of two elements implies that one of them is a point and another one is a line. We shall identify I with the simple graph of this incidence relation (bipartite graph). If number of neighbours of each element is finite and depends only on its type (point or line), then the incidence structure is a tactical configuration in the sense of Moore (see [15]). V. Ustimenko 139 The graph is k-regular if each of its vertices has degree k, where k is a constant. The incidence system is the triple (Γ, I, t) where I is a symmetric antireflexive relation (simple graph) on the vertex set Γ, t : Γ → ∆ is a type function onto the set of types ∆ such that αIβ and t(α) = t(β) implies α = β. The flag F is a nonempty subset in Γ such that α, β ∈ F implies αIβ. We assume that t(F ) = {t(x)|x ∈ F} 4.2. Groups, Coxeter systems and BN -pairs An important example of the incidence system as above is the so-called group incidence system Γ(G, Gs)s∈S . Here G is the abstract group and Gss∈S is the family of distinct subgroups of G. The objects of Γ(G, Gs)s∈S are the left cosets of Gs in G for all possible s ∈ S Cosets α and β are incident precisely when α ∩ β 6= ∅. The type function is defined by t(α) = s where α = gGs for some s ∈ S. Let (W, S) be a Coxeter system, i.e. W is a group with a set of distinguished generators given by S = {s1, s2, . . . , sl} ana generic relation (si × sj) mi,j = e. Here M = (mi,j) is a symmetrical l × l matrix with mi,i = 1 and off-diagonal entries satisfying mi,j > 2 (allowing mi,j = ∞ as a possibility, in which case the relation (si × sj) mi,j = e is omitted). Letting Wi =< S − {si} >, 1 6 i 6 l we obtain a group incidence system ΓW = Γ(W, Wi)16i6l is called the Coxeter geometry of W . The Wi are referred to as the maximal standard subgroups of W (see [17]). Let G be a group, B and N subgroups of G, and S a collection of cosets of B ∩ N in N . We call (G, B, N, S) Tits system ( or we say that G has a BN -pair) if (i) G =< B, N > and B ∩ N is normal in N , (ii) S is a set of involutions which generate W = N/(B ∩ N), (iii) sBw is a subset in BuB ∪ BswB for any s ∈ S and w ∈ W , (iv) sBs 6= B for all s ∈ S. Properties (i)–(iv) imply that (W, S) is a Coxeter system (see [17] or [12]). Whenever (G, B, N, S) is Tits system, we call the group W Weyl group of the system, or more usually the Weyl group of G. The subgroups Pi of G defined by BWiB are called the standard maximal parabolic subgroups of G. The group incidence system ΓG = Γ(G, Pi)16i6l is commonly referred to as Lie geometry of G (see [16]). Note that Lie 140 On the flag geometry of simple group of Lie type geometry of G and Coxeter geometry of the corresponding Weyl group have the same rank. In fact there is a type preserving morphism from ΓG onto ΓW given by gPi → wWi, where w is determined from the equality BgPi = BwPi. This morphism Ret is called a retraction (see [18]). Throughout this section (G, B, N, S) is Tits system which arises in connection with Chevalley group G, although we point that the results of this section remain valid in a far more general setting (see [18],[12], [17]). We write G = Xl(K) to signify that G is Chevalley group over the field K, with associated Dynkin diagram Xl. We are most interested in the case when K is finite, and we shall write Xl(q) instead of Xl(Fq) in that case. So, fix Chevalley group G = Xl(K) with corresponding Weyl group W . As in the previous section ΓW and ΓG are associated Coxeter and Lie geometries. 4.3. On the embeddings of geometries into Lie algebra Below we turn our attention to a method of embedding ΓW to a lattice. Let A = (ai,j) be Cartan matrix corresponding to the root system Ω of W . We consider the lattice R which is generated by simple roots α1, α2, . . . , αl and the reflection r1, r2, . . . rl of R defined by the equality (αi) rj = αi − ai,jαj . Let S = {r1, r2, . . . , rl} be the set of Coxeter generators of Weyl group W . Let α1 ∗, α2 ∗, . . . , αl ∗ be a dual basis of α1, α2, . . . , αl, i.e. αi ∗ is the linear functional on R which satisfies αi ∗(αj) = δi,j . We define the action of W on the dual lattice R∗ by l(x)s = l(xs), where l(x) ∈ R∗ and s ∈ S. Consider the orbit Hi = {αi ∗w|w ∈ W} of permutation group (W, R∗), which contains α1 ∗. Let H be the disjoint union of Hi. We give the set H the structure of an incidence system as follows. Linear functionals l1(x) and l2(x) are incident if and only if products l1(α)l2(α) > 0 for all α ∈ Ω. The type function t is defined by t(l(x)) = i where l(x) ∈ Hi. It can be shown that (H, I, t) is isomorphic to Coxeter geometry ΓW . In fact there is a unique isomorphism of ΓW with (H, I, t) which sends Wi to αi, 1 6 i 6 l.) This gives the desired embedding since H is a subset in R∗. We now consider an analogous embedding of Lie geometry ΓG into Borel subalgebra U = L0 + L+ of L. Let d = α1 ∗ + α2 ∗ + . . . αl ∗. Then we can take Ω+ = {α ∈ Ω|d(α) > 0} to be our set of positive roots in Ω. For any l(x) ∈ R∗ define η(L) = {α ∈ Ω+|l(α) < 0}. Let Lα be the root space corresponding to positive root α. For each h ∈ H we define the subalgebra Lh as the sum of Lα, α ∈ η−(h). Let V. Ustimenko 141 Ui = {h + v|h ∈ Hi, v ∈ Lh} and U be a disjoint union of Ui. We give U the structure of an incident system as follows. Elements h1 + v1 and h2 + v2 are incident if and only if each of the following holds: (i) h1(α)h2(α) > 0 for all α ∈ Ω, i.e. h1 and h2 are incident in (H, I, t). (ii) [h1 + v1, h2 + v2] = 0. Element h + v has type i if h + v ∈ Ui. In [10] (see [11]) it is shown that this newly defined incident system is isomorphic to Lie geometry ΓG, provided that the characteristic of K is zero or sufficiently large one to ensure the isomorphism at the level of the subgeometries (H, I, t) and ΓW . Then analogous to Weyl case there exists a unique isomorphism Retr of Γ(G) into (U, I, t) which sends Pi to αi, 1 6 i 6 l. Proposition 4.1. Let Γ = Γ(G) be the geometry of group G = Xn(q). The above interpretation of Γ(G) allows (i) generate Γ in O(|Γ|) elementary steps and check whether or not two elements of Γ are incident for time O(N2), where N is the number of positive roots. (ii) evaluate the degree and density of a deformation of tame com- putation of Tits automaton as values of size O(n2) and O(n3) respectively. (iii) generate a tame computation of Tits automaton consisting of k elementary steps for time O(n5). Proof. The part (i) is straightforward. Let us discuss point (ii). Let us consider the tame computation corresponding to the password p = (j1, f1), (j2, f2), . . . , (jk, fs), where fi ∈ Fq[ti1 , ti2 , . . . tin ] ∪ ∞. The maximal flag F of the Γ(G) can be treated as sequence (h1, v1), (h2, v2), . . . (hn, vn) of elements of geometry of type 1, 2, . . . , n such {h1, h2, . . . hn} is a flag in Weyl geometry and [hi + vi, hj + vj ] = 0, i, j ∈ M . We refer to such a sequence as expanded data of the flag. Let us denote η(hi) as Di and treat vectors from Lhi as functions from Di to Fq. It is easy to see that the flag is uniquely defined by the flag h1,h2,. . . ,hn of Weyl geometry, vector v1, v2|D2 − D1, v3|(D3 − (D1 ∪ D2)), . . . , vn|(Dn − (D1 ∪ D2 · · · ∪ Dn−1)). We refer to such a data of the flag as compressed data of the flag. We assume that the projection of vector on empty set is the formal symbol ∞. 142 On the flag geometry of simple group of Lie type We have the following interpretation of Tits automaton. Assume that a flag F ′ of type M − {i} is a subset of F of type M as above. The residue Res(F ′) = {x ∈ Γi|xIa, a ∈ F}. There are two elements h, h′ of type i such that sets S1 = {h + x|x ∈ Lh} and S0 = {h′ + x′|x′ ∈ Lh′ have nonempty intersections with Res(F ′) of type i of cardinality q and 1. Then we join flags F and F̃ of kind (h1 + v1), (h2 + v2), . . . , (hi−1 + vi−1), (h + x), (hi+1 + vi+1), . . . , (hn + vn), h + x ∈ Res(F ) by (i, ai), ai = x|η(h) − ((D1 ∪ D2 · · · ∪ Dn) − Di). Additionally we join F and F ′ ∪ S0 ∩ Res(F ′) by symbol (i, ∞). We have to do such labeling for each subset of F ′ of cardinality n − 1. Repetition of this procedure for each maximal flag bring us Tits automaton for which elements from the largest Schubert cells are accepting states. An initial state also has to be chosen as element of the largest Schubert cell. Let us consider the tame computation corresponding to the password p = (j1, f1), (j2, f2), . . . , (jk, fs), where fi ∈ Fq[x1, x2, . . . , xn] ∪ ∞. We define a type Sig(p) o as sequence (σ1, σ2, . . . , σk), where σi = 1 if ai 6= ∞ and σi = ∞ in remaining cases. Let F be an initial state. Notice that for symbolic transition functions produces the chain F = F0, F1 = Gj1 f1(F ), F2 = Gj2 f2(F1), . . . Fk = Gjk fik (Fk−1). The sequence Ri = Res(Fi), i = 0, 1, 2, . . . , k does not depend on the choice of different from ∞ functions f1, f2, . . . , fk or their specialisations. It depends only on sequences (j1, j2, . . . , jk) and (σ1, σ2, . . . , σk). Conse- qutive elements Ri and Ri+1, i = 0, 1, . . . , k − 1 are adjacent elements of Γ(W ) or Ri = Ri+1. Recall that R0 = Rk, corresponds to element of group W with the maximal length of irreducible representation in generators from S (Coxeter element). So the map E is a composition of multivariate maps Gj1 f1 : LR0 → LR1 , Gj2 f2 : LR1 → LR2 , . . . , Gjk fk : LRk−1 → LRk . The best way to compute is a consequtive computation of F1, F2,. . . ,Fk. We assume that F0 is given in its compressed form. For O(n3) operations we can convert compressed data into the expanded form. Further, we make all steps of Tits computation and present the final data in compressed form. Each step gives us addition of O(n2) new monomial terms, number of steps is O(n). It means that we can evaluate the density as O(n3). During the com- putation a monomial terms of degree O(n2) will appear. So we justify (ii). The computation of each monomial term costs O(n2). The product of two monomial terms can be done in time O(n2). Each step of compu- V. Ustimenko 143 tation costs O(n2) operations and the number of terms is n. It means that the generation of the map in the standard form costs O(n5) and we prove (iii). 5. Conclusion and complexity estimates Let us consider the deformation Ẽn = τ1Enτ2 of ∆, where τ1 is invertible monomial transformation ti → λitπ(i), λi ∈ F ∗ q , i = 1, 2, . . . , N , π is a permutation on {1, 2, . . . , N}, τ2 is a general bijective affine map of Fp N into itself. Recall that for the deformation Ẽ of E the equality deg(Ẽ) = deg(E) holds. Transformations E and τ1E have the same densities. Density of Eτ2 is at most n2 times larger than the density of E. Let us evaluate the complexity of the multivariate cryptosystem based on tame computation of Tits automaton. Notice that evaluations of degree and density shows that Bob can compute the value of Ẽ in a given point for O(n7) elementary steps. Alice can use described above method of proposition 4.1 as algorithm for the decryption of ciphertext c given by Bob. The application of τ−1 2 to c costs O(N2) = O(n4). The solution of the nonlinear equations and the computation of E′ costs O(n). The solution of linear system costs O(N3) = O(n6). So re-image of E can be computed for O(n7). The application of known τ1 −1 costs O(N). It means that for O(n13) Alice can decrypt. Generation of E costs O(n5), its deformation by τ1 and τ2 need O(n2) and O(n4) operations, respectively. So generation of Ẽ costs O(n11). As usually, we have to evaluate complexity as function from the dimension N of plainspace. So Bob applies public rule for O(N3+1/2). Alice can decrypt in time 0(N6+1/2). The generation of public key costs O(N5+1/2). The paper is dedicated to the memory of my teacher Lev Kaluzhnin. His wide mathematical vision brought us, participants of his Seminar, the light of Modern Algebra. References [1] Ding J., Gower J. E., Schmidt D. S., Multivariate Public Key Cryptosystems, 260. Springer, Advances in Information Security, v. 25, (2006). [2] V. Ustimenko, On Multivariate Cryptosystems Based on Computable Maps with Invertible Decompositions, Annales of UMCS, Informatica, volume 14 (2014) , Special issue ”Proceedings of International Conference Cryptography and Security Systems”, pp. 7-18. 144 On the flag geometry of simple group of Lie type [3] Anton Betten, Mihael Braun, Adalbert Kerber, Axel Kohnert, Alfred Wasserman Error Correcting Linear Codes Isometry and Applications, Springer, 2006. [4] Andreas Stephan Essenhans, Axel Kohnert, Alfed Wassermann, Constructions of codes for Network Coding, arXiv:1005, 2839[cs]. [5] A. Beultespacher, Enciphered Geometry, Some Applications of Geometry to Cryp- tography, Annals of Discrete Mathematics, v. 37, 1988, 59-68. [6] V. A. Ustimenko, Graphs with Special Arcs and Cryptography, Acta Applicandae Mathematicae, vol. 71, N2, November 2002, 117-153. [7] V. Ustimenko, Schubert cells in Lie geometries and key exchange via symbolic com- putations, Proceedings of the International Conference ”Applications of Computer Algebra, ACA 2010”, Vlora, Albanian Math. J., 2010, Vol 4, n. 4, 135-145. [8] V. Ustimenko, On walks of variable length in Schubert incidence systems and multivariate flow ciphers, Dopovidi of Nathional Acad. Sci. of Ukraine, 2014, No 3, P. 55 - 150. [9] V. A. Ustimenko, On some properties of Chevalley groups and their generalisations, In: Investigations in Algebraic Theory of Combinatorial objects, Moskow, Institute of System Studies, 1985, 134 - 138 (in Russian), Engl.trans.: Kluwer, Dordrecht, 1992, pp. 112-119 [10] V. A. Ustimenko, Linear interpretation of Chevalley group flag geometries, Ukraine Math. J. 43, Nos. 7,8 (1991), pp. 1055–1060 (in Russian). [11] V. A. Ustimenko, On the Varieties of Parabolic Subgroups, their Generalizations and Combinatorial Applications, Acta Applicandae Mathematicae 52 (1998): pp. 223–238. [12] R. W. Carter, Simple Groups of Lie Type, Wiley, New York (1972). [13] F. Harary, Graph Theory, Addison-Wesley Publishing Co, Reading, MA (1966). [14] R. Wilson, Introduction to Graph Theory, Oliver & Boyd, Edinburg, (1972). [15] E. Moore, Tactical Memoranda 1-3, Amer. J. of Math., vol. 18, n3 (1896), 264-29o. [16] A. Brower, A. Cohen, A. Nuemaier, Distance regular graphs, Springer, Berlin, 1989 [17] N. Bourbaki, Lie Groups and Lie Algebras, Chapters 1 - 9, Springer, 1998-2008. [18] F. Buekenhout (Editor), Handbook on Incidence Geometry, North Holland, Ams- terdam, 1995. Contact information Vasyl Ustimenko University of Maria Curie Skłodowska, Pl. Maria Curie Skłodowska 1, pok. 323, 20-931, Lublin, Poland E-Mail(s): vasyl@hektor.umcs.lublin.pl Received by the editors: 23.01.2015 and in final form 21.02.2015.