
In formulating the problem of optimal allocation of resources, we have assumed that the income at each step of all enterprises, and the maximum income Zfe * (g), since then the kstep until the end of the planning period, dependent only on the state of the system for £ kuy step and the Control Uh = (uk 1, uk2, Uks) at this point, but do not depend on how the funds are distributed among enterprises in the previous steps. However, in many problems of optimal allocation of income earned by ku step, and may be dependent on what means and in what amounts allocated to each of the companies in the previous steps, ie, the previous history of the process.
Thus violated one of the conditions imposed on the optimization problems, in order that they can be described by the model DP (condition 4, § 5, Ch. I). To take into account the background of the process of resource allocation, it is possible, as mentioned in § 1 of this chapter, to increase the number of state parameters at each step, it is artificial to include the number of phase coordinates of all control parameters preceding steps that define aftereffect. If the number of parameters is large, the scheme DP is so complicated that it becomes practically inapplicable. If the dimension of the artificial phase space is less than 4.3, then the problem can be solved by hand or (for a large number of steps, n) on the machine.
Consider the model of the optimal allocation of resourceseffect similar to task 2. 