مجله موجکها و جبر خطی
سال سوم شماره 1 (Spring and Summer 2016)

In this paper we generalize the max plus algebra system of real matrices to the class of real tensors and derive its fundamental properties.
Also we give some basic properties for the left (right) inverse, under the new system. The existence of order 2 left (right) inverses of tensors is characterized.

A Legendre wavelet method is presented for numerical solutions of stochastic Volterra-Fredholm integral equations. The main characteristic of the proposed method is that it reduces stochastic Volterra-Fredholm integral equations into a linear system of equations. Convergence and error analysis of the Legendre wavelets basis are investigated. The efficiency and accuracy of the proposed method was demonstrated by some non-trivial examples and comparison with the block pulse functions method.

$K$-frames which are generalization of frames on Hilbert spacesý, ýwere introducedý ýto study atomic systems with respect to a bounded linear operatorý. ýIn this paperý, ý$*$-$K$-frames on Hilbert $C^*$-modulesý, ýas a generalization of $K$-framesý, ýare introduced and some of their properties are obtainedý. ýThen some relationsý ýbetween $*$-$K$-frames and $*$-atomic systems with respect to an adjointable operator are considered and some characterizations of $*$-$K$-frames are givený. ýFinally perturbations of $*$-$K$-frames are discussedý.

Inverse Young inequality asserts that if $nu >1$, then $|zw|ge nu|z|^{frac{1}{nu}}(1-nu)|w|^{frac{1}{1-nu}}$, for all complex numbers $z$ and $w$, and equality holds if and only if $|z|^{frac{1}{nu}}=|w|^{frac{1}{1-nu}}$. In this paper the complex representation of quaternion matrices is applied to establish the inverse Young inequality for matrices of quaternions. Moreover, a necessary and sufficient condition for equality is given.

Let $A=U|A|$ be the polar decomposition of an operator $A$ on a Hilbert space $mathscr{H}$ and $lambdain(0,1)$. The $lambda$-Aluthge transform of $A$ is defined by $tilde{A}_lambda:=|A|^lambda U|A|^{1-lambda}$. In this paper we show that emp {i}) when $mathscr{N}(|A|)=0$, $A$ is self-adjoint if and only if so is $tilde{A}_lambda$ for some $lambdaneq{1over2}$. Also $A$ is self adjoint if and only if $A=tilde{A _lambda^ast$, emph{ii}) if $A$ is normaloid and either $sigma(A)$ has only finitely many distinct nonzero value or $U$ is unitary, then from $A=ctilde{A}_lambda$ for some complex number $c$, we can conclude that $A$ is quasinormal, emph{iii}) if $A^2$ is self-adjoint and any one of the $Re(A)$ or $-Re(A)$ is positive definite then $A$ is self adjoint, emph{iv}) and finally we show that $$opnorm{|A|^{2lambda}|A^ast|^{2- lambda}oplus0}leqopnorm{|A|^{2-2lambda}oplus|A|^{2lambda}} opnorm{tilde{A}_lambdaoplus(tilde{A}_lambda)^ast}$$ where $opnorm{cdot}$ stand for some unitarily invariant norm. From that we conclude that $$||A|^{2lambda |A^ast|^{2-2lambda}|leqmax(|A|^{2lambda},|A|^{2-2lambda})|tilde{A}_lambda|.$$

ýWe use of two notions functionally convex (brieflyý, ýF--convex) and functionally closed (brieflyý, ýF--closed) in functional analysis and obtain more resultsý. ýWe show that if $lbrace A_{alpha} rbrace _{alpha in I}$ is a family $F$--convex subsets with non empty intersection of a Banach space $X$ý, ýthen $bigcup_{alphain I}A_{alpha}$ is F--convexý.
ýMoreoverý, ýwe introduce new definition of notion F--convexiyý.