Conditional highs and suboptimal control are denoted by Z »* (g) and Wi * (|).
5. The complexity of the solution of the DP is mainly determined by the dimension of the problem defined by numbers 5 and g (or more parameters at each step and the number of control variables at this point.) These numbers are related, ie, the problem-dimensional and multi-5 to r can be transformed into one-dimensional and multi-dimensional in r 5.
The idea and methods of PD are more adapted to the consideration of discrete tasks. Only in rare cases it is possible to obtain an analytical solution of the Bellman. However, such an analytical solution is useful in theoretical analysis, in particular, in identifying the structure of the optimal solution.
3. The main advantage of this method is its applicability DP in any way of defining the objective function (analytic or tabular) and any admissible set of values of g and u (continuous or discrete). Deprived of the benefits of this classical optimization methods and other computational methods of mathematical programming.
4. Although computational methods DP in the discrete case and the tabulation of the functions associated with
and Uk (l) for all possible values of g, but the volume of calculations for these methods is significantly lower than the brute-force options. This is due to the fact that the conditional optimization options failed immediately discarded and retained only conditionally optimal in this step.