MIMIC算法在非线性多约束机械优化问题中的应用

将双变量相关的分布估计算法与惩罚函数法相结合,用于解决机械工程上非线性多约束优化设计问题.算法对每一次迭代寻优后的变量进行保留,建立概率模型,通过正态分布发生函数对变量重新取值,并采用惩罚函数来处理变量约束条件.仿真结果表明,该算法能有效防止早熟收敛,提高算法的全局搜索能力,具有较好的应用前景.

Application of MIMIC Algorithm in the Nonlinear Multi-constrained Mechanical Optimization Problem

The Bivariate Correlation Estimation of distribution algorithms （EDAs）is combined with the penalty func- tion method, which is a optimization way to solve the nonlinear and multi-constrained mechanical desgin problems. After recording the every optimization development value and constraining the variable with the penalty function, a probability model can be generated, and a value though the normal distribution function should be recorded once a- gain. The simulation results show that the proposed algorithm can effect the prevention of the premature conver- gence, improve the overall search ability and have a good application outlook.