智能规划是人工智能理论与实际应用结合最为紧密的研究课题之一。路标是对规划问题结构的开发，近年来广泛应用于智能规划领域。本课题将深入研究路标在启发式搜索规划方法中的指导作用，利用路标信息完善启发式评估过程，并重点挖掘灵活多变的路标排序关系所隐含的信息，为搜索算法提供更丰富的启发式策略，主要包括：1)从简化规划问题的角度使用命题路标，进一步约束路标切割启发式评估过程；2)利用路标排序关系改进放松规划图的层次结构，引导放松规划问题的增量求解；3)设计以路标为导向的前瞻搜索方法，针对路标构造前瞻路径，加强对搜索空间的探索力度；4)在随机游走过程中应用路标剪枝方法，检验路标对搜索方向的控制能力；5)利用路标排序关系刻画规划解的执行语义，在路标排序的约束下计算多样化的规划解集。本课题的研究将有助于充分发挥路标信息的启发作用，扩展规划系统的实际应用能力，为其向复杂的现实世界问题拓展奠定理论基础。

Planning has been one of the closest topics that relate the theories of Artificial Intelligence to practical applications. As an exploration of the planning problem structure, landmark is widely used in the research field of planning recently. This project will focus on the guidance power of landmarks for heuristic search planning method, improve the heuristic evaluation with the features of landmark and provide kinds of effective heuristic strategies for planning algorithms by extracting implicit information from flexible orders between landmarks. The research contexts mainly include: 1)to make use of propositional landmarks for simplifying the planning problem to propose a heuristic evaluation method based on the constraint of propositional landmarks; 2)to update the layers of relax planning graph and solve the relaxed problem incrementally under guidance of orders between landmarks; 3)to design a landmark-directed lookahead search method, construct lookahead plans for the respective propositional landmarks, which are used to explore the search space more strongly; 4)to apply a landmark-based pruning method in the random-walk procedure, checking the control power of landmarks to the search direction; 5)to describe the semantics of the plans using the orders in which the plan achieves the landmarks and generate diverse plans according to the constraint of landmark orders. The research will promote to explore the heuristic ability of landmark information deeply and develop the practical applications of planning systems, which will provide theoretical foundations for the extension of planning research to sophisticated real world problems.