CAJ | 학술논문

为了使布谷鸟搜索算法(Cuckoo Search, CS)在保持强大的全局搜索能力的同时，尽可能地提高局部搜索能力，在深入分析CS算法机理的基础上，将CS算法中影响布谷鸟搜索路径步长的参数β和布谷鸟蛋被发现(淘汰)的概率pa由固定值改为随搜索过程自适应变化的动态参数，将越界的鸟窝折返回边界内、在当前代最优鸟窝附近的区域随机建立1个新的鸟窝、而非折返回边界上重新建立鸟窝，以提高算法的局部搜索能力和收敛速度。改进后的CS算法称为自适应布谷鸟搜索算法(Adaptive Cuckoo Search, ACS)。通过8个标准测试函数分别测试了CS算法和ACS算法的性能，结果表明，无论是简单的单峰函数还是复杂的多峰函数，无论是小型的低维函数还是大型的高维函数，ACS算法的寻优性能均超过CS算法。
위료사포곡조수색산법(Cuckoo Search, CS)재보지강대적전국수색능력적동시，진가능지제고국부수색능력，재심입분석CS산법궤리적기출상，장CS산법중영향포곡조수색로경보장적삼수β화포곡조단피발현(도태)적개솔pa유고정치개위수수색과정자괄응변화적동태삼수，장월계적조와절반회변계내、재당전대최우조와부근적구역수궤건립1개신적조와、이비절반회변계상중신건립조와，이제고산법적국부수색능력화수렴속도。개진후적CS산법칭위자괄응포곡조수색산법(Adaptive Cuckoo Search, ACS)。통과8개표준측시함수분별측시료CS산법화ACS산법적성능，결과표명，무론시간단적단봉함수환시복잡적다봉함수，무론시소형적저유함수환시대형적고유함수，ACS산법적심우성능균초과CS산법。
In order to make the Cuckoo Search algorithm maintaining the strong global Search ability, meanwhile improving the local Search ability as much as possible, the parameter ofβwhich affecting Step length of the search path and the fraction (pa) of the egg laid by a cuckoo abandoned change from fixed value to dynamic parameters which adaptive changing along with the search process on the basis of in-depth analysis of mechanism of CS algorithm. In addtion the nests of cuckoo that crossing the border are made to build new nests around the best nest instead of building new nests in the border to improve the local search ability and convergence speed of the algorithm.The improved CS algorithm called adaptive cuckoo search algorithm (shortened form ACS algorithm).The ACS algorithm and CS algorithm are tested by eight standard test functions. The conclusions indicate that whether a simple unimodal function, or complex multimodal functions, whether the small low-dimensional function or large high-dimensional function,the optimization performance of ACS algorithm are better than CS algorithm.