Because public systems problems are frequently ill-defined and have fuzzy constraints and vague multiple objectives, their solution by means of formal optimization models is not widely accepted. This paper will explore the modes (both useful and otherwise) in which optimization has been applied to such problems. The need for optimization models which somehow “fit” the decision-makers' methods and backgrounds will be discussed. Several other aspects of public systems optimization will be considered, in a somewhat random fashion. This paper will raise more questions than answers.