格基约化算法是密码分析的重要工具。该文借鉴遗传算法的基本策略,通过对初始格基的调整变换,提出了一种新的格基约化算法,新算法总能得到给定格中长度更短的向量和质量更高的一组基。利用该算法,针对最短向量问题(SVP)挑战的部分数据进行了测试,新算法的输出结果达到或超过了挑战的公开记录,约化效果良好。
Lattice reduction algorithms play an important role in the field of cryptanalysis.In this paper,based on the strategies of genetic algorithm,a new lattice reduction algorithm is proposed through the transformation of the initial lattice basis.The new algorithm always can obtain a shorter vector and a higher quality basis compared with the original algorithms.By the new algorithm,some lattice bases of the Shortest Vector Problem(SVP) challenge are experienced and the outputs of the new algorithm can always reach or break the records on the internet which illustrates that the new algorithm behaves well.