Balancing Multiple Sources of Reward in Reinforcement Learning.
(05/01/2014)
For many problems which would be natural for reinforcement learning, the reward signal is not a single scalar value but has multiple scalar components. Examples of such problems include agents with multiple goals and agents with multiple users. Creating a single reward value by combining the multiple components can throw away vital information and can lead to incorrect solutions. We describe the multiple reward source problem and discuss the problems with applying traditional reinforcement learning. We then present an new algorithm for finding a solution and results on simulated environments....
Tác giả: Shelton, C. R. |
Số trang: 8 |
Lĩnh vực: CNTT |
Năm XB: 2006 |
Loại tài liệu: Khác
Tài liệu cần xác thực trước khi tải
Tiêu đề | Tải về |
Balancing Multiple Sources of Reward in Reinforcement Learning. | Số trang: 8
| Loại file:
For many problems which would be natural for reinforcement learning, the reward signal is not a single scalar value but has multiple scalar components. Examples of such problems include agents with multiple goals and agents with multiple users. Creating a single reward value by combining the multiple components can throw away vital information and can lead to incorrect solutions. We describe the multiple reward source problem and discuss the problems with applying traditional reinforcement learning. We then present an new algorithm for finding a solution and results on simulated environments.
|
miễn phí
|
© Copyright 2012 Trung tâm Thông tin Khoa học và Công nghệ - Sở Khoa học & Công nghệ TP. Cần Thơ
Địa chỉ: 118/3 Trần Phú - P.Cái Khế - Q.Ninh Kiều - TPCT
Điện thoại: 0292 3824031 Fax: 0292 3812352
|
|
Lượt truy cập:
(Website trong thời gian thử nghiệm)