This paper studies the problem of computing labeled orthogonal drawings. A label is modeled as a rectangle of prescribed size and it can be associated with either a vertex or an edge. Several optimization goals are taken into account. Namely, the labeled drawing can be required to have minimum total edge length, minimum width, minimum height, or minimum area. We present ILP models to compute optimal drawings with respect to the first three objectives and an algorithm exploiting these models which computes a drawing of minimum area (the compaction problem is known to be NP-complete in general).