针对智能规划和知识22程技术在材料设计中的结合应用问题,从材料设计规划问题的建模开始,根据经典智能规划方法在解决这类问题时存在的不足之处,提出了一种基于领域约束的材料设计规划方法。该方法通过把本体知识库中的领域知识引入到扩展的图规划系统中,使其成为基于领域约束的智能规划,既充分发挥了通用图规划方法的长处,又利用领域约束对搜索过程的限制大大缩小了问题的解空间,并进一步通过引入目标偏好对求解策略进行了更为实用的改进,从而显著提高了规划系统的求解速度和求解质量,使其能够真正达到实用水平。由此实现的系统原型已成功应用于山东莱钢一类重要齿轮钢产品的研发过程,并取得了良好的试验效果,大大缩短了产品研发周期并降低了研发成本。
To address the issue of the combination application of the intelligent planning and the knowledge engineering technology to material design, this paper starts with modeling of the material design planning problem, and presents the domain constraint integrated planning method for material design, which can effectively make up the shortcomings of conventional intelligent planning methods in solving the problem. By introducing the domain knowledge in the ontology-basod knowledge base into an extended Graphplan system, this method can not only make full use of the advantages of classical Graphplan methods, but also greatly reduce the solution space of given problems through the restriction of domain constraints in the searching process. Moreover, the practicability of the planning strategy can be further improved by introducing the notion of goal preference. As a result, the solving speed and the solution quality of the planning system can be considerably improved. The implemented system prototype successfully accomplished the optimal design task of carburized gear steel in the Laiwu Steel, and obtained good experimental results.