CAJ | 학술논문

混沌迭代序列是复杂系统动力学研究的一个分支,其序列值在不同参数条件下时会出现分又及混沌现象.已有的方法不能同时挖掘拟合迭代序列的迭代函数的结构及其相应条件参量.文章则旨在同时挖掘出二者,主要工作包括:(1)提出了基于协同进化的异构种群挖掘模型,能融合不同种群的优势;(2)提出了新的适合挖掘迭代序列的适应度计算方式;(3)从理论上证明了多种群协同挖掘的进化难度远大于单种群进化难度,通过实验证实了在有效协同策略下,多种群进化得到的结果远优于单种群的进化结果;(4)提出3种协同进化策略,在对迭代序列的函数拟合以及参数拟合两方面,多路并行式结合策略能达到相对较优效果;(5)在合成数据和真实数据上进行了实验,证实了算法的正确性和有效性.
혼돈질대서렬시복잡계통동역학연구적일개분지,기서렬치재불동삼수조건하시회출현분우급혼돈현상.이유적방법불능동시알굴의합질대서렬적질대함수적결구급기상응조건삼량.문장칙지재동시알굴출이자,주요공작포괄:(1)제출료기우협동진화적이구충군알굴모형,능융합불동충군적우세;(2)제출료신적괄합알굴질대서렬적괄응도계산방식;(3)종이론상증명료다충군협동알굴적진화난도원대우단충군진화난도,통과실험증실료재유효협동책략하,다충군진화득도적결과원우우단충군적진화결과;(4)제출3충협동진화책략,재대질대서렬적함수의합이급삼수의합량방면,다로병행식결합책략능체도상대교우효과;(5)재합성수거화진실수거상진행료실험,증실료산법적정학성화유효성.
Chaotic iterative sequence is a research direction in complex system kinetics research. The sequence may incarnate bifurcate or chaotic phenomenon under different parameter conditions. The existing methods cannot discover the iterative structure and parameters simultaneously. This study aims at mining the iterative functions and conditional parameters parallel. The main contributions include:(1)Proposes co-evolution model based on heterogeneous populationsto integrate the advantages of those populations.(2)Proposes a new fitness function to mine the sequence in iterative style.(3)Theoretically proves that heterogeneous populations' co-evolution is more difficult than a single population's evolution. Experimentally proves that given effective co-evolution strategy, heterogeneous populations can obtain much better results than single population.(4) Proposes three co-evolution strategies. The cooperating strategy can archive relatively good results in terms of fitting the sequence's mathematic equation and the equation's parameters.(5)Conducts extensive experiments on both synthesized and real data to validate the correctness and efficiency of the algorithm.