| Summary: | 碩士 === 國立彰化師範大學 === 資訊管理學系 === 105 === There are many resource management problems in the real world. It is worthy to study how to make the best decisions about resource management through machine learning, so that the overall benefits can be optimized.
Reinforcement learning is one of the popular methods in machine learning because the learning process of reinforcement learning is like human. Learn by relying on your own experience and the incentives that interact with the environment.
The purpose of this study was to investigate the application of reinforcement learning to resource management, by builds the simulation environment platform, Fish Banks, simulates the dynamic resource of the renewable resource. This study designs four experimental cases, respectively, random players, simple learning players, heuristic learning Players and reinforcement learning players in the simulation environment to resource management, to understand the different types of players with two different goal’s learning performance and resource management results. The results show that the reinforcement learning player has better learning effectiveness and better performance to resource management.
keywords: Machine Learning, Reinforcement Learning, Q-Learning
|
|---|
