Consider a linear programming problem of the form\

minc^(')x+d^(')y\ s.t. Ax+Dy=0.

\ a. Suppose that we have access to a very fast subroutine for solving problems of the form\

minh^(')x\ s.t. Fx_(x)f,

\ for arbitrary cost vectors

h

. How would you go about decomposing the problem?\ b. Suppose that we have access to a very fast subroutine for solving problems of the form\

min,d^(')y\ s.t. Dy =0,

\ for arbitrary right-hand side vectors

h

. How would you go about decomposing the problem?

Consider a linear programming problem of the form mins.t.cx+dyAx+DybFxfy0. a. Suppose that we have access to a very fast subroutine for solving problems of the form mins.t.hxFxf, for arbitrary cost vectors h. How would you go about decomposing the problem? b. Suppose that we have access to a very fast subroutine for solving problems of the form mins.t.dyDyhy0, for arbitrary right-hand side vectors h. How would you go about decomposing the