In terms of volume calculations both schemes are the same, but with some additional research is preferable to one or the other. For example, the addition of new steps are best carried out by using direct flow calculations to study the sensitivity to variations in the go more convenient reversing, to variations gn - direct flow, etc.
This general description of the computational scheme DP is somewhat abstract. We conclude with some general remarks.
1. In essence, both versions of the computational process is a concrete instance of a more general interpretation of incremental optimization. Instead of starting a particular task with a certain number of steps in advance of n and the fixed value of the initial state go we consider the sequence of similar problems for any n and any g: a one-step, two-step, ..., n-step task. Initially, one-step optimization problem is solved and get Maximum performance indicator Zx * (l). Then solved two-step task, which is obtained by adding to the previous one more step [get Z2 * (g)], and so steps can be increased in this sequence of tasks in to the beginning of the process and at the end. Save the index k for the numbering of steps from the initial to the final state of the system and introduce the index i, indicating the number of steps in {-step task, regardless of the computing scheme forward stroke (where i = \ k) or schema reverse (here i = n- k + \). That index i will number the newly added step in the transition from the (i-D)-by-step tasks to /-step. Appropriate management at this stage is denoted by u. We further state at the end (t-1) silo problem through g, and in the end i-steps - a g '. Then the state equation can be written as u).

Copyright ©