Let us have two different points a,bβRn. Prove that the set of points which in the Euclidean norm are closer to the point a than to b make up a half-space. Is this true for another norm?
Find ΟSβ(y)=Ο if S={xβRnβ£β₯xβxcββ₯β€R}, yβ/S
Find ΟSβ(y)=Ο if S={xβRnβ£cTx=b}, yβ/S
Find ΟSβ(y)=Ο if S={xβRnβ£Ax=b,AβRmΓn,bβRm}, yβ/S
Illustrate the geometric inequality that connects ΟSβ(y),yβ/S,xβS, from which it follows that ΟSβ(y) is a projection of the y point onto a convex set of S.
For which sets does the projection of the point outside this set exist? Unique?
Find ΟSβ(y)=Ο if S={xβRnβ£cTxβ₯b}
Find ΟSβ(y)=Ο if S={xβRnβ£x=x0β+XΞ±,XβRnΓm,Ξ±βRm}, yβS
Let SβRn be a closed set, and xβRn be a point not lying in it. Show that the projection in l2β norm will be unique, while in lββ norm this statement is not valid.
Find the projection of the matrix X on a set of matrices of rank k,XβRmΓn,kβ€nβ€m. In Frobenius norm and spectral norm.
Find a projection of the X matrix on a set of symmetrical positive semi-definite matrices of XβRnΓn. In Frobenius norm and the scalar product associated with it.
Find the projection ΟSβ(y) of point y onto the set S={x1β,x2ββR2β£β£β£x1ββ£+β£x2ββ£=1} in β₯β
β₯1β norm. Consider the different positions of y.
Find ΟSβ(y)=Ο, if S={xβRnβ£Ξ±iββ€xiββ€Ξ²iβ,i=1,β¦,n}.
Prove that projection is a nonexpansive operator, i.e. prove, that if SβRn is nonempty, closed and convex set, then for any (x1β,x2β)βRnΓRn
β₯ΟSβ(x2β)βΟSβ(x1β)β₯2ββ€β₯x2ββx1ββ₯2β