闫 乐,董 宇.基于关系网络的新型学习控制[J].轻工机械,2014,32(4): |
基于关系网络的新型学习控制 |
Improved Algorithm of Pure Pursuit Model |
|
DOI: |
中文关键词: 智能算法 强化学习 关系网络 权值矩阵 稳定控制 |
英文关键词:smart algorithm reinforcement learning network of relationships weight matrix stability control |
基金项目: |
|
摘要点击次数: 1472 |
全文下载次数: 1233 |
中文摘要: |
设计了一种不同于传统智能算法如自适应控制与模糊控制等的智能控制算法,该算法基于目标输入关系网络,
借鉴强化学习的思想,采用输出试探与评价进行学习,是一种模型无关的控制算法,该算法可以与随机优化算法结合,通
过实验,验证了其在控制过程中有效性,并通过与遗传算法结合,验证了与遗传算法相结合的该智能算法的控制效果。 |
英文摘要: |
A intelligent control algorithm was presented.which was different from traditional intelligent algorithms like
adaptive control and fuzzy control. This intelligent control algorithm was based on the target-input relationship network,
which learned from reinforcement learning, and used try and evaluation to learn. It was a model-independent control
algorithm. This control algorithm could be combined with stochastic optimization algorithms. Experiments verify the
effectiveness of this algorithm by combination with genetic algorithms. |
查看全文 查看/发表评论 下载PDF阅读器 |
关闭 |