Meta-Level Incomplete Information
作者： Reid Simmons , Richard Goodwin
DOI:
关键词: Complete information 、 Representation (mathematics) 、 Information retrieval 、 Markov model 、 Markov chain 、 Statistical model 、 Process (engineering) 、 Computer science 、 Data mining 、 Partially observable Markov decision process 、 Markov process
摘要: Typically, when we discuss planning with incomplete info rmation, are referring to information about the state of world or effects actions. One common approach dealing this type is use a probabi listie model reason partial that available. Tu make reasoning probabili st ic models tractable, assumptions and approximations made necessarily an representation available on . In extended abstract, I will disc uss concerning robot. Since information, wiJl call it metalevel illustrate so urces meta-level describe how pianner can information. parti cuiar probabilistic model, partially obsenable Markov deci si process, examples from work Xavier robot di sc ussion more concrete.
参考文章(1)
Sven Koenig, Reid Simmons, Probabilistic robot navigation in partially observable environments international joint conference on artificial intelligence. pp. 1080- 1087 ,(1995)