基于子结构匹配方法多规则L系统反演的研究.docVIP

基于子结构匹配方法的多规则L系统反演研究 Research on the inverse problem of mutil-rules Lindenmayer System ?based on substructure matching 朱元伟 摘　要:由已知的L系统的结果字符串寻找出原来的初始公理和产生式规则集是L系统反演的主要工作。本文以L系统的自相似特点为基础, 通过子结构的分层匹配算法实现带有两个限定条件的L系统字符串的反演，能够简便准确获取L系统初始公理和产生式规则集。算法主要包含4个步骤:分割最终字符串并建立子结构池;对子结构池中包含其他子结构的字符串进行进一步分割；通过与各层次回溯得到的子结构匹配，取得子结构与产生式前驱之间的对应关系；通过回溯迭代过程得到其最简形式，获得初始字符串；试验中，所有被测试的字符串都可以被反演到等价L系统规则集。 关键词: L系统;虚拟植物; L系统反演;子结构匹配 Abstract—Finding out the?set of original?rules is the main work of ?the inversion?of ?Lindermayer system?.The algorithm of this paper obtain a set of rules from a long Lindenmayer string with two conditions by matching sub-structures. The algorithm consists of five steps：divide the Lindenmayer string into substrings recursively and push them into the sub-structure buffer pool;further divide the sub-structure which contains others in the pool;replace the sub-structures in Lindenmayer by the identifier in the pool then do repeat as in the previous step.Using the newly obtained sub-structure match structures in buffer pool to get the map relationship between sub-structures and predecessors; replacing repeated string with a predecessor char that marking the same string in the pool, and continuing this process recursively until get the original axiom. In the tests carried out, all the Lindenmayer strings could be reverted to a set of Lindenmayer rules that identical to the original rules used to generate the string. Key-words：L-system; virtualplant; Inverse problems; substructure matching Problems