论文部分内容阅读
将状态空间的问题求解过程变换为逐步缩小与目标状态的差异过程是一种问题的分解方式．求解差异的顺序可通过分析算符对状态的影响而作出规划，规划的原则是最大限度地在不改变最近已实现子目标的条件下实现下一子目标．为此，在问题分解时各层子目标选择的依据是让各算符有最大的可利用率，即以状态对算符最小约束传播的原则选择各层子目标；最后生成一个子目标规划层次集．问题求解过程就表现为从初始状态开始实现层次集中的某一子目标序列，其间可能涉及子目标回溯．
Transforming the problem solving process in the state space into the process of gradually narrowing the process difference with the target state is a way to decompose the problem. The order in which differences are solved can be made by analyzing the influence of operators on the state, and the principle of planning is to maximize the next sub-goal without changing the recently achieved sub-goals. Therefore, when the problem is decomposed, the sub-goals of each level are selected based on maximizing the availability of each operator, that is, the sub-goals of each level are selected based on the principle of minimizing the state constraints on operators. Finally, a sub- set. The solution process of the problem manifests itself as a certain sub-target sequence from the initial state, which may involve the sub-target backtracking.