Multi-Agent Dam Management Model Based on Improved Reinforcement Learning Technology

Run Ying Wang; Lin Xu

doi:10.4028/www.scientific.net/AMM.198-199.922

Paper Titles

Research on Semantic Business Process Model in Logistics Distribution Field
p.899

Research on Warship Maintenance Support Simulation System Based on the Method of Equipment Maintenance Support Simulation Experiment
p.905

Application Specific Architecture Template Reuse for SoC Transaction Level Modeling
p.911

OONS: An Object Oriented Neuronal Simulator
p.917

Multi-Agent Dam Management Model Based on Improved Reinforcement Learning Technology
p.922

Analyzing the Distribution of Energy Consumption of Coagulation Technological Using the Fractal Dimension
p.927

Research of Driver’s Visual Perception Model
p.932

Research on Credibility Evaluation Method of Conjoint Simulation System
p.936

Model and Simulation of the Brain Scalp Potential Analysis
p.942

HomeApplied Mechanics and MaterialsApplied Mechanics and Materials Vols. 198-199Multi-Agent Dam Management Model Based on Improved...

Multi-Agent Dam Management Model Based on Improved Reinforcement Learning Technology

Abstract:

In order to achieve efficient management of the dam, the new algorithms such as reinforcement learning, Synergetic, Structural Risk Minimization and Particle Swarm Optimization are used to establish a Cooperative Wavelet Least Squares Support Vector Machine Model. To improve the convergence rate and make full use of knowledge and advice of mechanics and hydraulics of the dam, WLS-SVRM and WLS-SVCM models are used cooperatively. Before the training online, mapping provides training samples for WLS-SVCM. During the course of training online, the numerical simulation and WLS-SVCM will provide knowledge and advices for WLS-SVRM. Case study shows that the model can provide timely information of gate opening and management information of the dam so as to provide decision support for engineering management.