基于向量树搜索的1-1/2维配料算法与实现
An Algorithm for Piecing Plates based on Searching Vector Trees and Its Implementation
-
摘要: 针对工程项目中经常遇到的长大板材构件的一个方向和两个互相垂直方向的拼接问题,提出了一种基于向量树搜索的智能化处理算法。将拼接区间抽象为向量空间,求解的过程是对由向量空间内的节点构成的向量树进行深度优先搜索,直至搜索到向量空间树的叶节点(解节点)。向量空间搜索树的每一个节点是一个区段向量表,区段向量表是一个禁区边界线间距离值的有序排列。这种方法不但可以得出多种可供选择的方案,而且也有效地避免了利用组合方法解决拼接问题而产生的组合爆炸。本文利用PROLOG人工智能语言实现所提出的算法,并可以用图形表示拼接结果,计算板料利用率。Abstract: This paper provides an approach to solve piecing plates by one dimension and two vertical dimensions faced in engineering projects. This approach is based on searching vector tree which comes from the field of artificial intelligence. Piecing can be abstracted as a set of vector spaces. These vector spaces can make up of a tree of vectors, Ddpth-first search which is one of the blind search technologies can be used to search the above tree of vectors, and one of leaf nodes would be the solution. This approach can avoid explosion of combiation resulted form a huge number of nodes. Prolog, a artificial intelligence computer language, is used to implement this algorithm.