Factorization of mappings of topological spaces and homomorphisms of topological groups in accordance with weight and dimension ind

Symbols w(X), nw(X), and hl(X) denote the weight, the network weight, and the hereditary Lindelöf number of a space X, respectively. We prove the following factorization theorems.

1. Let X and Y be Tychonoff spaces, φ: X→Y a continuous mapping, hl(X)≤τ, and w(Y)≤τ. Then there exist a Tychonoff space Z and continuous mappings ψ: X→Z, χ: Z→Y such that φ=χ o ψ, Z=ψ(X), w(Z)≤τ andind Z≤ind X. Moreover, if nw(X)≤τ, then mapping ψ is one-to-one.
2. Let π: G→H be a continuous homomorphism of a Hausdorff topological group G to a Hausdorff topological group H, hl(G)≤τ and w(H)≤τ. Then there are a Hausdorff topological group G^* and continuous homomorphisms g: G→G^*, h: G^*→H so that π=h o g, G^*=g(G), w(G^*)≤τ andind G^*≤ind G. If nw(G)≤τ, then g is one-to-one.
3. For every continuous mapping φ: X→Y of a regular Lindelöf space X to a Tychonoff space Y one can find a Tychonoff space Z and continuous mappings ψ: X→Z, χ: Z→Y such that φ=χ o ψ, Z=ψ(X), w(Z)≤w(Y),dim Z≤dim X, andind ₀ Z≤ind ₀ X, whereind ₀ is the dimension function defined by V.V.Filippov with the help of G_δ-partitions. If we additionally suppose that X has a countable network, then ψ can be chosen to be one-to-one. The analogous result also holds for topological groups.
4. For each continuous homomorphism π: G→H of a Hausdorff Lindelöf Σ-group G (in particular, of a σ-compact group G) to a Hausdorff group H there exist a Hausdorff group G^* and continuous homomorphisms g: G→G^*, h:G^*→H so that π=h o g, G^*=g(G), w(G^*)≤w(H),dimG^*≤dimG, andind G^*≤ind G. Bibliography: 25 titles.

     

On the length of a finite group and of its 2-generator subgroups

The nonsoluble length λ(G) of a finite group G is defined as the minimum number of nonsoluble factors in a normal series of G each of whose quotients either is soluble or is a direct product of nonabelian simple groups. The generalized Fitting height of a finite group G is the least number h = h* (G) such that F* _h (G) = G, where F* ₁ (G) = F* (G) is the generalized Fitting subgroup, and F* _i+1(G) is the inverse image of F* (G/F*_i (G)). In the present paper we prove that if λ(J) ≤ k for every 2-generator subgroup J of G, then λ(G) ≤ k. It is conjectured that if h* (J) ≤ k for every 2-generator subgroup J, then h* (G) ≤ k. We prove that if h* (〈x, x^g 〉) ≤ k for allx, g ∈ G such that 〈x, x^g 〉 is soluble, then h* (G) is k-bounded.     

Elementary divisors of induced transformations on symmetry classes of tensors

Let V_χ(G) denote the symmetry class of tensors over the vector space V associated with the permutation group G and irreducible character χ. Write v₁*v₂*...*v_m for the decomposable symmetrized product of the indicated vectors (m=degG). If T is a linear operator on V, let K(T) denote the associated operator on V_χ(G), i.e., K(T)v₁*v₂*...*v_m=Tv₁*Tv₂*...*Tv_m. Denote by D(T) the derivation operator D(T)v₁*v₂*...*v_m=Tv₁*v₂...*v_m+v₁*Tv₂*v₃* ...*v_m+...+v₁*v₂*...*v_m–1*Tv_m. The article concerns the elementary divisors of K(T) and D(T).     

Balanced extensions of graphs and hypergraphs

For a hypergraphG withv vertices ande _i edges of sizei, the average vertex degree isd(G)= ∑ie ₁/v. Callbalanced ifd(H)≦d(G) for all subhypergraphsH ofG. Let $$m(G) = \mathop {\max }\limits_{H \subseteqq G} d(H).$$ A hypergraphF is said to be abalanced extension ofG ifG?F, F is balanced andd(F)=m(G), i.e.F is balanced and does not increase the maximum average degree. It is shown that for every hypergraphG there exists a balanced extensionF ofG. Moreover everyr-uniform hypergraph has anr-uniform balanced extension. For a graphG let ext (G) denote the minimum number of vertices in any graph that is a balanced extension ofG. IfG hasn vertices, then an upper bound of the form ext(G) ₁ n ² is proved. This is best possible in the sense that ext(G)>c ₂ n ² for an infinite family of graphs. However for sufficiently dense graphs an improved upper bound ext(G) ₃ n can be obtained, confirming a conjecture of P. Erdõs.     

Differences of functions and groups generators for compact Abelian groups

LetG be a Hausdorff compact Abelian group andC be the component of the identity element ofG. We consider a special class, ?(G), of functions inL ² (G) whose Fourier series satisfy certain convergence conditions (stronger than absolute convergence). We show thatG/C is topologically generated by not more thann elements if and only if, for each functionf in ?(G), there area ₁,...,a _n inG and functionf ₁,...f _n in ?(G) such that $$f = \sum\limits_{j = 1}^n {(f_j - \delta _{aj} * f_j ),}$$ where * is convolution defined in the usual sense, and δ _a denotes the Dirac measure ataεG.     

A solution concept forn-person noncooperative games

The paper describes a solution concept forn-person noncooperative games, developed jointly by the author and Reinhard Selten. Its purpose is to select one specific perfect equilibrium points=s (G) as the solution of any given noncooperative gameG. The solution is constructed by an inductive procedure. In defining the solutions (G) of gameG, we use the solutionss (G ^*) of the component gamesG ^* (if any) ofG; and in defining the solutions (G^*) of any such component gameG ^*, we use the solutionss (G ^**) of its own component gamesG ^** (if any), etc. This inductive procedure is well-defined because it always comes to an end after a finite number of steps. At each level, the solution of a game (or of a component game) is defined in two steps. First, aprior subjectiveprobability distribution p _i is assigned to the pure strategies of each playeri, meant to represent the other players' initial expectations about playeri's likely strategy choice. Then, a mathematical procedure, called thetracing procedure, is used to define the solution on the basis of these prior probability distributionsp _i . The tracing procedure is meant to provide a mathematical representation for thesolution process by which rational players manage to coordinate their strategy plans and their expectations, and make them converge to one specific equilibrium point as solution for the game     

Conjugacy class sizes and solvability of finite groups

Let G be a finite group and let G* be the set of elements of primary, biprimary and triprimary orders of G. We show that suppose that the conjugacy class sizes of G* are exactly {1, p ^a , n, p ^a n} with (p, n)?=?1 and a??? 0, then G is solvable.     

Progression-free sets in finite abelian groups

Let G be a finite abelian group. Write and denote by rk(2G) the rank of the group 2G.Extending a result of Meshulam, we prove the following. Suppose that A⊆G is free of “true” arithmetic progressions; that is, a₁+a₃=2a₂ with a₁,a₂,a₃∈A implies that a₁=a₃. Then |A|<2|G|/rk(2G). When G is of odd order this reduces to the original result of Meshulam.As a corollary, we generalize a result of Alon and show that if an integer k?2 and a real ε>0 are fixed, |2G| is large enough, and a subset A⊆G satisfies |A|?(1/k+ε)|G|, then there exists A₀⊆A such that 1?|A₀|?k and the elements of A₀ add up to zero. When G is of odd order or cyclic this reduces to the original result of Alon.     

The structure of 0-bisimple strongly E *-unitary inverse monoids

We introduce a class of strongly E ^*-unitary inverse semigroups S _i (G, P) (i = 1,2) determined by a group G and a submonoid P of G and give an embedding theorem for S _i (G, P). Moreover we characterize 0-bisimple strongly E ^*-unitary inverse monoids and 0-bisimple strongly F ^*-inverse monoids by using S _i (G, P).     

Critical graphs for chromatic difference sequences

Let α_k(G) denote the maximum number of vertices in a k-colorable subgraph of G. Set α_k=α_k(G)-α(_k-1)(G). The sequence a₁(G),a₂(G),… is called the chromatic difference sequence (cds) of G. We call a graph G critical if no proper subgraph of G has the same cds as G. We prove that (with a single exception) if there exists a graph G having cds (G)=〈a₁,a₂,a₃〉 (a₃>) and if a₁?a₂+a₃, then there exists a connected critical graph H with cds(H)= 〈a₁, a₂,a₂〉.     

On a co-induction question of Kechris

This note answers a question of Kechris: if H < G is a normal subgroup of a countable group G, H has property MD and G/H is amenable and residually finite, then G also has property MD. Under the same hypothesis we prove that for any action a of G, if b is a free action of G/H, and b _G is the induced action of G, then CInd _H ^G (a|H) × b _G weakly contains a. Moreover, if H < G is any subgroup of a countable group G, and the action of G on G/H is amenable, then CInd _H ^G (a|H) weakly contains a whenever a is a Gaussian action.     

On Lie algebra decompositions related to spherical homogeneous spaces

LetG be a connected, reductive, linear algebraic group over an algebraically closed fieldk of characteristik zero. LetH ₁ andH ₂ be two spherical subgroups ofG. It is shown that for allg in a Zariski open subset ofG one has a Lie algebra decomposition g = h₁ + Adg ? h₂, where a is the Lie algebra of a torus and dim a ≤ min (rankG/H ₁,rankG/H ₂). As an application one obtains an estimate of the transcendence degree of the fieldk(G/H ₁ xG/H ₂) ^G for the diagonal action ofG. Ifk = ? andG _a is a real form ofG defined by an antiholomorphic involution σ :G→G then for a spherical subgroup H ? G and for allg in a Hausdorff open subset ofG one has a decomposition g = g_a + a Adg ? h, where a is the Lie algebra of σ-invariant torus and dim a ≤ rankG/H.     

The chromatic difference sequence of a graph

Let α_k(G) denote the maximum number of vertices in a k-colorable subgraph of G. Set α_k(G)=α_k(G)?α_(k?1)(G). The sequence a₁(G), a₂(G),… is called the chromatic difference sequence of the graph G. We present necessary and sufficient conditions for a sequence to be the chromatic difference sequence of some 4-colorable graph.     

Cocycle rigidity of abelian partially hyperbolic actions

Suppose G is a higher-rank connected semisimple Lie group with finite center and without compact factors. Let G = G or G = G ? V, where V is a finite-dimensional vector space V. For any unitary representation (π,H) of G, we study the twisted cohomological equation π(a)f ? λf = g for partially hyperbolic element a ∈ G and λ ∈ U(1), as well as the twisted cocycle equation π(a₁)f ? λ1f = π(a₂)g ? λ₂g for commuting partially hyperbolic elements a₁, a₂ ∈ G. We characterize the obstructions to solving these equations, construct smooth solutions and obtain tame Sobolev estimates for the solutions. These results can be extended to partially hyperbolic flows in parallel.As an application, we prove cocycle rigidity for any abelian higher-rank partially hyperbolic algebraic actions. This is the first paper exploring rigidity properties of partially hyperbolic that the hyperbolic directions don’t generate the whole tangent space. The result can be viewed as a first step toward the application of KAM method in obtaining differential rigidity for these actions in future works.     

EQUIVARIANT SELF EQUIVALENCES OF PRINCIPAL FIBRE BUNDLES

Let E be a compact Lie group, G a closed subgroup of E, and H a closed normal sub-group of G. For principal fibre bundle (E,p, E,/G;G) tmd (E/H,p‘,E/G;G/H), the relation between auta(E) (resp. autce (E)) and autG/H(E/H) (resp. autGe/H(E/H)) is investigated by using bundle map theory and transformation group theory. It will enable us to compute the group JG(E) (resp. SG(E)) while the group J G/u(E/H) is known.     

<Emphasis Type="Italic">C</Emphasis><Superscript>*</Superscript>-Subalgebras Generated by a Single Operator in <Emphasis Type="Italic">B</Emphasis>(<Emphasis Type="Italic">H</Emphasis>)

In this paper, we characterize a C ^*-subalgebra C ^*(x) of B(H), generated by a single operator x. We show that if x is polar-decomposed by aq, where a is the partial isometry part and q is the positive operator part of x, then C ^*(x) is *-isomorphic to the groupoid crossed product algebra A_q×_a\mathbbG_a\mathcal{A}_{q}\times_{\alpha }\mathbb{G}_{a} , where A_q=C^*(q)\mathcal{A}_{q}=C^{*}(q) and \mathbbG_a\mathbb{G}_{a} is the graph groupoid induced by a partial isometry part a of x.     

Products of powers in finite simple groups

LetG be a group. For a natural numberd≥1 letG ^d denote the subgroup ofG generated by all powersa ^d ,a∈G. A. Shalev raised the question if there exists a functionN=N(m, d) such that for anm-generated finite groupG an arbitrary element fromG ^d can be represented asa ₁ ^d ...a _N ^d ,a _i ∈G. The positive answer to this question would imply that in a finitely generated profinite groupG all power subgroupsG ^d are closed and that an arbitrary subgroup of finite index inG is closed. In [5,6] the first author proved the existence of such a function for nilpotent groups and for finite solvable groups of bounded Fitting height. Another interpretation of the existence ofN(m, d) is definability of power subgroupsG ^d (see [10]). In this paper we address the question for finite simple groups. All finite simple groups are known to be 2-generated. Thus, we prove the following: THEOREM:There exists a function N=N(d) such that for an arbitrary finite simple group G either G ^d =1 orG={a ₁ ^d ...a _N ^d |a _i ∈G}. The proof is based on the Classification of finite simple groups and sometimes resorts to a case-by-case analysis.     

Fractal Property in B(H) Induced by Partial Isometries

We consider the groupoid C*-dynamical systems (A,\mathbb G,a),{(A,{\mathbb G},\alpha ),} where A is a C*-algebra in B(H),\mathbb G{B(H),{\mathbb G}} is a certain groupoid, and α is an embedding groupoid action of \mathbb G,{{\mathbb G},} acting on A in B(H). In particular, we are interested in the case where the groupoid \mathbb G{{\mathbb G}} is a fractaloid, a groupoid with fractal property. Moreover, we restrict our interests to the case where \mathbb G{{\mathbb G}} is generated by finitely many partial isometries in B(H). We observe the basic properties of such C*-dynamical systems.     

Duality theorems for regular homotopy of finite directed graphs

Dati uno spazio topologico normale e numerabilmente paracompattoS ed un grafo finito ed orientatoG si prova che tra gli insiemiQ(S, G) eQ ^*(S, G) delle classi dio-omotopia e dio ^*-omotopia esiste una biiezione naturale. Nelle stesse condizioni, seS′ è un sottospazio chiuso diS eG′ un sottografo diG, esiste ancora una biiezione naturale tra gli insiemiQ (S, S′; G, G′) eQ ^* (S, S′; G, G′) delle classi di omotopia. Si mostra infine che in condizioni meno restrittive per lo spazioS le precedenti biiezioni possono non sussistere.