摘要
结合非线性规划的约束条件构造了一个新的极大熵函数,利用该函数将问题转化成了两个目标的多目标优化问题。通过对违反约束动态的进行惩罚,提出了一种新的极大熵多目标粒子群算法。该方法能有效的保持群体中不可行解的一定比例,从而增加了群体的多样性,而且避免了传统的过度惩罚缺陷,使群体更好地向最优解逼近。计算机仿真表明,该算法对非线性规划问题求解是非常有效的。
A new maximum entropy function based on the constraint conditions of nonlinear programming problems is given. Then using the new maximum entropy function, the nonlinear programming problem is transformed into a bi-objective optimization problem. By dynamically penalty to the constraint violations so as to keep a ratio of infeasible solutions in swarm, a new maximum entropy multiobjective particle swarm algorithm is presented. This method can not only increase the diversity of population but also avoid the defect of over-penalization. So it can make the group approach optimal solution easily. The computer simulations demonstrate the proposed algorithm is effectiveness to solve nonlinear programming problems.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第4期914-916,共3页
Computer Engineering and Design
基金
国家自然科学基金项目(60374063)
陕西省自然科学基础研究计划基金项目(2006A12)
陕西省教育厅科学技术研究计划基金项目(07JK180)
宝鸡文理学院重点科研基金项目(ZK0619)
关键词
非线性规划
约束规划
多目标优化
粒子群算法
动态惩罚
极大熵
nonlinear programming
constrained programming
multi-objective optimization
particle swarm algorithm
dynamic penalty
maximum entropy