|
(5.1) |
Let ,, denote a Markov process generated by, where is a small parameter and , is an (m+1) x (m+1) matrix such that for and for . We let represent the machine capacity state at time t.
Definition 5.1 A production control process is admissible, if
Definition 5.2 A function defined on is called an admissible feedback control or simply a feedback control, if
has a unique solution;
|
(5.2) |
Here we assume that the production cost function and the surplus cost function satisfy Assumption 3.1, and the machine capacity process satisfies Assumptions 3.2 and 3.3. Furthermore, similar to Assumption 4.3, we also assume
As in Fleming and Zhang (1998), the positive attrition rate implies a uniform bound for . In view of the fact that the control is bounded between 0 and m, this implies that any solution to (5.1) must satisfy
= | |||
(5.3) |
The average cost optimality equation associated with the average-cost optimal control problem in , as shown in Sethi, Zhang, and Zhang (1997), takes the form
= | |||
(5.4) |
Theorem 5.1 The minimum average expected cost of is bounded in, i.e., there exists a constant C>0 such that
In order to construct open-loop and feedback hierarchical controls for the system, one derives the limiting control problem as. As in Sethi, Zhang, and Zhou (1994), consider the enlarged control space
The average-cost optimality equation associated with the limiting control problem is
(5.5) |
Armed with Theorem 5.1, one can derive the convergence of the minimum average expected cost as goes to zero, and establish the convergence rate.
Theorem 5.2 There exists a constant C such that for all ,
This implies in particular that .
Here we only outline the major steps in the proof. For the detailed proof, the reader is referred to Sethi, Zhang, and Zhang (1997). The first step is to prove by constructing an admissible control of from the optimal control of the limiting problem , and by estimating the difference between the state trajectories corresponding to these two controls. Then one establishes the opposite inequality, namely,, by constructing a control of the limiting problem from a near-optimal control of and then using Assumptions 2.1.
The following theorem concerning open-loop controls is proved in Sethi, Zhang, and Zhang (1997).
Theorem 5.3 (Open-loop control) Letbe an optimal control for, and let
for some positive constant C.
We next consider feedback controls. We begin with an optimal feedback control for , which with a slight abuse of notation is denoted as . This is obtained by minimizing the right-hand side of (5.5), i.e.,
|
(5.6) |
|
(5.7) |
|
(5.8) |
Theorem 5.4 (Feedback control) Let n=1. Assume (A.2.1) and (A.9.3) and that the feedback control of the limiting problem is locally Lipschitz in x. Furthermore, suppose that for each, the equation
|
(5.9) |
|
(5.10) |
|
(5.11) |
where.
Since there are several hypotheses, namely (5.9)-(5.11), in Theorem 5.4, it is important to provide at least an example for which these hypotheses hold. Below we provide such an example and at the same time illustrate the ideas of constructing the asymptotically optimal controls.
Example 5.1. Consider the problem
with and the generator for to be
This is clearly a special case of the problem formulated in this section. In particular, Assumptions 3.3 and 3.4 hold and. The limiting problem is
where we use . Let us set the function
|
(5.12) |
is an asymptotically optimal feedback control for .
Remark 5.1 It is possible to obtain other optimal feedback controls for Example 5.1. It is also possible to provide examples with nonzero production cost, for which Lipschitz feedback controls satisfying (5.9)-(5.11) can be obtained, and their optimality asserted by a verification theorem similar to Theorem 4.3 in Sethi, Suo, Taksar, and Yan (1998).
Remark 5.2 A similar averaging approach is introduced in Altman and Gaitsgory (1993) and Altman and Gaitsgory (1997), Nguyen and Gaitsgory (1997), Shi, Altman and Gaitsgory (1998), Nguyen (1999) and references there in. They consider a class of nonlinear hybrid systems in which the parameters of the dynamics of the system may jump at discrete moments of time, according to a controlled Markov chain with finite states and action spaces. They assume that the unit of the length of intervals between the jumps is small. They prove that the optimal solution of the hybrid systems governed by the controlled Markov chain can be approximated by the solution of a limiting deterministic optimal control problem.