摘要
动态约束规划问题求解的困难在于如何处理问题的约束及时间(环境)变量.本文给出了求解一类定义在自然数集上的动态约束规划问题的差分进化算法,该方法借助于问题的约束条件设计了一种新的适应度函数及选择算子、同时给出了一种带一维不精确局部搜索的变异算子极大地增强了群体的多样性、提高了算法跳出局部最优的能力.数值试验表明,该算法性能稳定性较好,收敛速度较快,全局搜索能力较强,其对动态非线性约束规划问题求解是有效的.
The difficult to solve dynamic constrainted programming problems is how to do with the constraint and the time(invironment) variance.In this paper,a new differentinal evolution algorithm for solving a class of constrained programming problem defined in natural numbers set is proposed.First,a new fitness fuction and selection operator based on the constraint conditions of dynamic constrainted programming problem is given.Futhermore,a new mutation operator with one-dimensional inexact local search is designed.Based on these,the diversity of population is improved and enabling the algorithm to jump over any local minimum trap.The simulations show that the algorithm is reliable,fast and robust in global optimization,and the proposed algorithm is effective for solving the dynamic constrainted programming problems.
出处
《昆明理工大学学报(理工版)》
CAS
北大核心
2010年第6期114-118,共5页
Journal of Kunming University of Science and Technology(Natural Science Edition)
基金
陕西省教育厅科学研究计划项目(09JK329)
陕西省自然科学基础研究计划项目(2009JM1013)
宝鸡文理学院重点科研计划项目(ZK0840)
关键词
动态规划
约束规划
差分进化
约束处理
dynamic programming
constrained programming
differentinal evolution
deal with the constraints