An Integrated Evolutionary Algorithm for Expensive Global Optimization

We propose an integrated algorithm named low dimensional simplex evolution extension (LDSEE) for expensive global optimization in which only a very limited number of function evaluations is allowed. The new algorithm accelerates an existing global optimization, low dimensional simplex evolution (LDSE), by using radial basis function (RBF) interpolation and tabu search. Different from other expensive global optimization methods, LDSEE integrates the RBF interpolation and tabu search with the LDSE algorithm rather than just calling existing global optimization algorithms as subroutines. As a result, it can keep a good balance between the model approximation and the global search. Meanwhile it is self-contained. It does not rely on other GO algorithms and is very easy to use. Numerical results show that it is a competitive alternative for expensive global optimization.

基于模拟的企业过程模型自动优化技术研究

企业过程模型参数自动优化是一个多参数多目标的系统优化问题.采用线性加权法将其转换成多参数单目标问题的求解,提出基于计算机模拟的企业过程模型参数的自动寻优方法.它将传统的共轭梯度法FR（fletcher reeves）和禁忌搜索算法TS（tabu search）结合起来,采用FR法进行局部寻优,由TS法实现从当前局部最优点向全域范围内的更优区域转移,循环往复达到求出全域范围最优点的目的.改进了FR法与TS法,克服了各自的缺陷,并提出禁忌区域表的概念,从而加速寻优过程.它适用于任意多维曲面的多极值问题最优求解,对企业进行BPR（business process reengineering）和实施ERP（enterprise resource planning）管理有较大的指导意义.