The concept of ,'best compromise" is not yet well-defined. Some
methods attempt to give a precise a priori definition of what is the best
compromise. This can be done by defining a hierarchy (the objective
functions are strictly ordered; the feasible region is restricted to solutions
optimising the first objective function, then the second and so on)
or by methods by "utility" (each objective functions CJ is given a
weight p/, then a unique objective function C = Ejd/Y is maximised).
Other methods are heuristic: the exploration of a feasible region is
conducted step by step toward the solution that will be finally chosen.
Such methods are: parameterisation (the most important of the CJ is
maximised; the other CJ are transformed into constraints, and the right
and side is parameterised), or STEM.