过度规划、部分可满足规划及简单偏好综述

Survey on over-subscription planning,partial satisfaction planning and simple preferences

智能规划是人工智能的重要分支,它从某个特定问题的初始状态出发,寻找达到解决该问题的目标状态的动作序列,但是在许多真实问题中,往往没有足够的资源来实现所有目标。因此,智能规划中对软目标的研究包括过度规划、部分可满足规划和简单偏好,已成为近年来的热点。将对过度规划、部分可满足规划和简单偏好的主要研究成果进行综述,包括这几类问题的差异比较、主要解决方案以及对将来的展望。

Automated planning is an important sub-area of artificial intelligence.A planning system starts from an initial state of a specified problem and tries to find an action sequence to achieve a desired goal state.But in a real world,an agent does not have enough resources to achieve all goals.So soft goals,including over-subscription planning（OSP） and partial satisfaction planning（PSP） and simple preferences（SP）,have become a hot-spot in planning community in recent years.This paper gave a survey about major researches on these issues,including the comparison of these related fields,solutions and future works.