It seems that no operation is defined for multiplying a DiagonalMatrix by a scalar, nor is it possible to add a DiagonalMatrix to a SparseMatrix. Is this true? Also, is it possible to define a diagonal matrix containing a single constant scalar value? I see that one can accomplish the same thing by defining a vector v and applying v.setConstant(scalar), but adding v.asDiagonal() to a sparse matrix doesn't work. Any suggestions, anyone?

hm you are right on all points. Regarding scalar multiplication this is very easy to do and in the mean time you can do (D.diagonal()*scalar).asDiagonal() without overhead (expect the writing )

To make a constant diagonal matrix you can do: VectorXd::Constant(v,size).asDiagonal()

To add a diagonal matrix to an existing sparse matrix you can currently do:

VectorXd diag(size);
diag = ...;
for(int j=0; j<size;++j)
M.coeffRef(j,j) += diag(j);