基于遗传禁忌混合算法的低相关区序列集搜索方法

提出一种基于遗传-禁忌混合算法(genetic-taboo hybrid algorithm)的低相关区(low correlation zone,LCZ)序列集的搜索方法.建立搜索LCZ序列集的优化模型并设定了模型的目标函数,将禁忌算法嵌入到遗传算法的变异操作中,提出一种遗传-禁忌混合算法,该算法利用遗传算法的全局搜索能力和禁忌算法的局部搜索能力从而实现高效搜索.迭代过程中,种群在进入新一轮进化之前,首先剔除移位等价序列(shift equivalent sequence),从而保持种群的多样性,防止算法过早收敛.根据优化算法收敛后目标函数保持不变的特点,设定了算法结束准则,当超过一定进化代数后目标函数值仍保持不变,则认为进化结束.对非周期四元和周期二元序列集进行数值仿真,结果证明了该方法的有效性.

Searching method of LCZ sequence set based on genetic-taboo hybrid algorithm

A searching method for low correlation zone (LCZ) sequence set based on genetic-taboo hybrid algorithm is derived in this paper. Firstly, the optimal model of LCZ sequence set and the target function are established. Then, the genetic-taboo hybrid algorithm is derived by embedding taboo algorithm (TA) in the mutation operation of the genetic algorithm (GA) , which employs the global searching capability of the GA and the local searching capability of the TA to achieve high efficient search. Shift equivalent sequences of the population are rejected before the new round of evolution to keep the population's diversity and get rid of premature convergence. According to the feature of the stability of the target function after the algorithm convergence, the stop criterion is set; the evolution is assumed to be stopped when the target function keeps invariant for a certain evolution period. Finally, simulation for aperiodic quaternary and periodic binary sequence sets shows that the proposed method is effective.