摘要
内点算法的宽邻域长步算法比窄邻域小步算法理论复杂度差,但实际计算效果优于窄邻域小步算法.为缩小内点算法的这种理论与实践间的差距,针对单调非线性互补问题,给出一个宽邻域齐次内点算法,并估计算法的复杂度.计算结果表明,该宽邻域长步算法的理论复杂度与现阶段计算效果最好的窄邻域小步算法的理论复杂度一致.数值实验也验证了该算法的有效性.
Wide neighborhood(long-step)algorithm has the less theoretical complexity than narrow neighborhood(small-step)algorithm,but it has the better practical performance than narrow neighborhood(small-step)algorithm.For reducing the gap of the two algorithms between theory and practice,a wide neighborhood homogenous algorithm for monotone nonlinear complementarity problems is presented,and the complexity of the algorithm is estimated.The results show that the theoretical complexity of the proposed algorithm is the same as that of the best narrow neighborhood algorithm.The numerical results show that the algorithm is efficient and reliable.
出处
《纺织高校基础科学学报》
CAS
2017年第3期372-378,共7页
Basic Sciences Journal of Textile Universities
基金
2016年国家级大学生创新创业训练计划项目(201610722015)
2016年陕西省省级大学生创新创业训练计划项目(2451)
2016年咸阳师范学院大学生创新创业训练计划项目(2016004)
关键词
齐次算法
单调非线性互补问题
宽邻域
内点算法
homogenous algorithm
monotone nonlinear complementarity problem
wide neigh-borhood
interior point method
