FOUR GENERAL TYPES OF POLICY OPTIMIZING MODELS (OR SETS OF MATHEMATICAL MODELS) ARE DISCUSSED - THE LEGITIMATION MODEL, THE PROBABLISTIC MODEL, OPTIMUM LEVEL MODELS, AND OPTIMUM MIX MODES. THE LEGITIMATION MODEL INVOLVES ANSWERING THE QUESTION AS TO WHICH OF VARIOUS ALTERNATIVE DECISIONS THAT LACK INHERENT NUMERICAL ORDER AND DIVISIBILITY SHOULD BE CHOSEN OR LEGITIMIZED IN LIGHT OF THEIR RELATIONS TO CERTAIN WEIGHTED OR UNWEIGHTED GOALS. THE PROBABILISTIC MODEL IS LIKE THE LEGITIMATION MODEL EXCEPT THE RELATIONS BETWEEN THE ALTERNATIVE DECISIONS AND THE GOALS ARE CONTINGENT ON ONE OF MORE EVENTS WHICH SOMETIMES OCCUR AND SOMETIMES DO NOT OCCUR, OPTIMUM LEVEL MODELS INVOLVE REACHING A DECISION ALONG A SCALE OF NUMERICAL ALTERNATIVES WHERE GOING TOO HIGH OR TOO LOW ON THE SCALE PRODUCES UNDESIRABLE EFFECTS. THE OBJECT IS THUS TO FIND THE OPTIMUM LEVEL WHICH MINIMIZES THOSE UNDESIRABLE EFFECTS OR COSTS AND MAXIMIZES THE DESIRABLE EFFECTS OR BENEFITS. THE OPTIMUM MIX MODELS INVOLVE REACHING A DECISION AS TO HOW MUCH OF A SCARCE RESOURCE TO ALLOCATE TO VARIOUS ACTIVITIES OR GEOGRAPHICAL PLACES. IN ADDITION, RELATIONS BETWEEN THE MODELS ARE EMPHASIZED, INCLUDING A DISCUSSION OF PROBLEMS THAT CAN BE VIEWED AS EITHER OPTIMUM LEVEL OR OPTIMUM MIX PROBLEMS.