摘要
基于指数函数的性质,提出简易罚函数法(SPFM),用于有效求解难约束优化问题(COP),并屏蔽选取罚因子的困难性。将SPFM和差分演化相结合,给出一种求解难COP的改进差分演化算法(MDE)。利用MDE求解Bump问题可以得出该问题的多个新的最优解,证明MDE在求解难COP时的高效性。
For solving hard Constrained Optimization Problem(COP), this paper proposes a new method named Simple Penalty Function Method (SPFM) based on properties of exponent function. SPFM avoids the difficulty of choosing the penalty factors. Modified Differential Evolution algorithm(MDE) is advanced, which combines SPFM with Differential Evolution(DE). By using MDE to solve Bump problem, more better optimization solutions gained by MDE shows that MDE is effective.
出处
《计算机工程》
CAS
CSCD
北大核心
2008年第13期193-194,217,共3页
Computer Engineering
基金
国家自然科学基金资助重点项目(60473045)
关键词
差分演化
约束优化
罚函数法
Bump问题
Differential Evolution(DE)
constrained optimization
Penalty Function Method(PFM)
Bump problem