摘要
针对贷款组合优化决策模型的求解问题,以模拟退火算法为基础,利用设置记忆器和在算法后链接一个局部搜索过程的方法,对原有算法进行了改进。该算法可在求解大规模组合优化问题的迭代过程中实现快速调整,以兼顾解的质量和运行时间,快速找到最优解,克服了原有算法的随机性。数值计算结果表明,该算法具有很强的适用性。
The simulated annealing algorithm is extended by setting a m emorial de vice and linking a local search process in the algorithm. The improved algorithm is used to solve optimal decision problems of loan's portfolio. During the iter ation process for solving large combinatorial optimization, the proposed algorit hm can be quickly adjusted, so that it can pay attention to the quality of solut ion and running time, find the optimal solution fast, and overcome the randomnes s of the original algorithm. Numerical results show robustness of the proposed algorithm for loan's portfolio.
出处
《吉林大学学报(信息科学版)》
CAS
2003年第2期143-147,共5页
Journal of Jilin University(Information Science Edition)
基金
国家自然科学基金资助项目(70171004)
天津市自然科学基金资助项目(013602611)
关键词
贷数组合
NP问题
模拟退火
全局优化
随机捷索
Loan's portfolio
NP problem
Simulated anne aling algorithm
Global optimization
Random search