Anda belum login :: 27 Nov 2024 18:13 WIB
Home
|
Logon
Hidden
»
Administration
»
Collection Detail
Detail
Optimal scheduling for multi-item single machine with stochastic demand
Bibliografi
Author:
Ghaussi-Mujtaba, Homa
;
Menaldi, Jose Luis
(Advisor)
Topik:
MATHEMATICS|ENGINEERING
;
SYSTEM SCIENCE
Bahasa:
(EN )
ISBN:
0-599-70766-6
Penerbit:
Wayne State University Press
Tahun Terbit:
2000
Jenis:
Theses - Dissertation
Fulltext:
9966141.pdf
(0.0B;
10 download
)
Abstract
Our work is inspired by the unbounded solution of Hamilton Jacobi Bellman equations in connection with optimal switching control problems with state constraints. We consider the formulation of a production system when the demand is subjected to exponential white noise. The problem is the control of a stochastic differential equation, where the dynamics of this stochastic differential equation can be changed into n-different settings, and the state of this dynamics system is constrained (reflected diffusions) to a compact set. Our first task is to formulate the problem so that the controlled process stays in the set. The second question is to find an optimal way to control the dynamics so that the expectations of a certain functional of the controlled process, the cost is a minimum. In the formulation of the problem we construct the mathematical model componentwise, using results from one dimension, by means of reflected diffusion processes. By allowing some singularity for the cost function we write the model so that the value function becomes unbounded in one part of the boundary as the initial condition approaches that part of the boundary. We consider the dynamic programming solution to this problem. As the Hamilton Jacobi Bellman equation have only a formal meaning, we approximate our controlled Markov diffusion by a controlled Markov chain, and study the approximating discrete-time Markov chain. The process thus constructed approximates
y
(·) in the weak sense; thereby, the cost functional of this approximating process
J
h
(
y
) converges to the cost
J
(
y
). We consider an application with numerical approximation in value space and in policy space.
Opini Anda
Klik untuk menuliskan opini Anda tentang koleksi ini!
Lihat Sejarah Pengadaan
Konversi Metadata
Kembali
Process time: 0.15625 second(s)