摘要
提出了一种有约束的变测度积分一水平集的算法,对不同的箱子采用不同的测度,结合确定性数论方法选取一致分布佳点集来代替Monte—Carlo随机投点,使水平值充分地下降,更快地到达全局最小,从而提高算法的计算效率.给出了算法的收敛性证明,并通过数值算例验证了它的有效性.
A variable measure algorithm for global optimization problem with constraints is proposed. Taking different measure in different sub-box and choosing a good point set of uniform with the deterministic number theory instead of Monte-Carlo method, the level value can be reduced enough to reach the global optimization and improve the efficiency of the algorithm. Then the global convergence of this algorithm is proven, and the simulation examples show the validity of the algorithm.
出处
《系统科学与数学》
CSCD
北大核心
2008年第2期232-242,共11页
Journal of Systems Science and Mathematical Sciences
基金
苏州科技学院校重点学科基金
上海教委重点学科支助项目
关键词
积分-水平集
变测度
约束最优化
收敛性
Integral-level set, variable measure, constrained problem, convergence.
