By definition of covariance,
[tex]\mathrm{Cov}(U,V)=E[(U-E[U])(V-E[V])]=E[UV-E[U]V-UE[V]+E[U]E[V]]=E[UV]-E[U]E[V][/tex]
Since [tex]U=2X+Y-1[/tex] and [tex]V=2X-Y+1[/tex], we have
[tex]E[U]=2E[X]+E[Y]-1[/tex]
[tex]E[V]=2E[X]-E[Y]+1[/tex]
[tex]\implies E[U]E[V]=(2E[X]+E[Y]-1)(2E[X]-(E[Y]-1))=4E[X]^2-(E[Y]-1)^2=4E[X]^2-E[Y]^2+2E[Y]-1[/tex]
and
[tex]UV=(2X+Y-1)(2X-(Y-1))=4X^2-(Y-1)^2=4X^2-Y^2+2Y-1[/tex]
[tex]\implies E[UV]=4E[X^2]-E[Y^2]+2E[Y]-1[/tex]
Putting everything together, we have
[tex]\mathrm{Cov}(U,V)=(4E[X^2]-E[Y^2]+2E[Y]-1)-(4E[X]^2-E[Y]^2+2E[Y]-1)[/tex]
[tex]\mathrm{Cov}(U,V)=4(E[X^2]-E[X]^2)-(E[Y^2]-E[Y]^2)[/tex]
[tex]\mathrm{Cov}(U,V)=4V[X]-V[Y]=4a-a=\boxed{3a}[/tex]