FD-Autotune: Domain-Specific Configuration using Fast Downward
FD-Autotune: Domain-Specific Configuration using Fast Downward
复制标题
FD-Autotune：使用快速下行的特定于域的配置
DOI：
--
复制
发表时间：
2011
期刊：
影响因子：
--
通讯作者：
Jendrik Seipp
中科院分区：
文献类型：
--
作者：
Chris Fawcett;M. Helmert;H. Hoos;E. Karpas;Gabriele R¨oger;Jendrik Seipp
FD-Autotune学习规划系统基于最新的、高度参数化的快速向下规划框架的领域特定配置理念，通过通用自动算法配置程序进行配置。我们描述了Fast Downward的极大配置空间如何被限制到一个子空间，尽管该子空间仍然非常大，但可以通过最先进的自动配置程序进行管理。FD-Autotune使用著名的ParamILS配置器，并使用最近开发的HAL实验环境实现。
The FD-Autotune learning planning system is based on the idea of domain-specific configuration of the latest, highly parametric version of the Fast Downward Planning Framework by means of a generic automated algorithm configura-tion procedure. We describe how the extremely large con-figuration space of Fast Downward was restricted to a sub-space that, although still very large, can be managed by state-of-the-art automated configuration procedures. FD-Autotune uses the well-known ParamILS configurator and was realised using the recently developed HAL experimentation environment.