A statistical approach to separate the probability model (W, ?, P(?)) from the given parametric family {(W, ?, P
u (?)): u ? U } has first been proposed in [1, 2]. In some cases, the separated probability space (W, ?, P(?)) will be an adequate model for the given statistically stable experiment Э or controlling system [3?7]. In other cases, where it is required to synthesize the controlling system Э by some criterion, the probabilistic model (W, ?, P(?)) is declared to be its target or optimization model. It has been proposed in [1, 2] to use a class of permutation and quasi-permutation solution strategies for the construction of (W, ?, P(?)) from the family {(W, ?, P
u (?)): u ? U } of permutation distributions. This work is a direct extension of studies [1, 2] and it considers the properties of a class of extreme strategies.
|
