Electronics, Год журнала: 2024, Номер 13(21), С. 4217 - 4217

Опубликована: Окт. 27, 2024

Distributed parameter systems (DPSs) frequently appear in industrial manufacturing processes, with complex characteristics such as time–space coupling, nonlinearity, infinite dimension, uncertainty and so on, which is full of challenges to the modeling system. At present, most DPS methods are offline. When internal parameters or external environment change, offline model incapable accurately representing dynamic attributes real Establishing an online for that reflects real-time dynamics system very important. In this paper, idea reinforcement learning creatively integrated into three-dimensional (3D) fuzzy a learning-based 3D method proposed. The agent improves strategy by continuously interacting environment, can adaptively establish from scratch. Specifically, paper combines deterministic gradient algorithm based on actor critic framework function represented two updated alternately. uses TD (0) target via semi-gradient method; using chain derivation rule behavior value established model. Since continuous problem, proposes average reward, effectively realize modeling. suggested methodology implemented three-zone rapid thermal chemical vapor deposition reactor simulation results demonstrate efficacy methodology.

Язык: Английский