Ricochet Robots: A Transverse ASP Benchmark
Ricochet Robots: A Transverse ASP Benchmark
复制标题
Ricochet Robots：横向 ASP 基准
DOI：
10.1007/978-3-642-40564-8_35
复制
发表时间：
2013
期刊：
影响因子：
--
通讯作者：
M. Lindauer
中科院分区：
文献类型：
--
作者：
M. Gebser;H. Jost;Roland Kaminski;Philipp Obermeier;Orkunt Sabuncu;Torsten Schaub;M. Lindauer
答案集编程的一个显着特点是它的多功能性。除了可满足性测试，它还提供了各种形式的模型枚举、交集或联合以及优化。此外，有一个增量和反应式求解由于其适用于动态域的兴趣越来越大。然而，到目前为止，还没有进行比较研究，对比各自的建模能力和计算的影响。为了评估各种不同形式的ASP解决，我们提出亚历克斯兰多夫的棋盘游戏跳弹机器人作为一个横向基准问题，使我们能够比较各种方法在一个统一的设置。开始，我们考虑编码ASP规划问题的替代方法，并讨论底层建模技术。反过来，我们进行了实证分析对比传统的解决方案，优化，增量和反应式的方法。此外，我们还研究了一些提升技术在我们的案例研究领域的影响。
A distinguishing feature of Answer Set Programming is its versatility. In addition to satisfiability testing, it offers various forms of model enumeration, intersection or unioning, as well as optimization. Moreover, there is an increasing interest in incremental and reactive solving due to their applicability to dynamic domains. However, so far no comparative studies have been conducted, contrasting the respective modeling capacities and their computational impact. To assess the variety of different forms of ASP solving, we propose Alex Randolph's board game Ricochet Robots as a transverse benchmark problem that allows us to compare various approaches in a uniform setting. To begin with, we consider alternative ways of encoding ASP planning problems and discuss the underlying modeling techniques. In turn, we conduct an empirical analysis contrasting traditional solving, optimization, incremental, and reactive approaches. In addition, we study the impact of some boosting techniques in the realm of our case study.