计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (1): 100-100.

• 产品、研发、测试 • 上一篇    下一篇

基于进化算法的优化平台设计

刘红星,肇莹,朱博,高敦堂   

  1. 南京大学
  • 收稿日期:2005-12-02 修回日期:1900-01-01 出版日期:2007-01-01 发布日期:2007-01-01
  • 通讯作者: 刘红星 lhx lhx

Design of optimization platform using evolutionary algorithm

Hongxing Liu,,,   

  1. 南京大学
  • Received:2005-12-02 Revised:1900-01-01 Online:2007-01-01 Published:2007-01-01
  • Contact: Hongxing Liu

摘要: 线性规划非线性规划等优化软件在社会、经济、工程等领域应用潜力巨大。现有优化软件大都采用的是经典的局部优化技术或者简单的全局优化技术。本文将进化算法引入称为优化平台的优化软件设计。对平台的关键技术进行了分析,提出了相应的平台方案,并予以了实现。本平台方案的特点是:界面动态调整增广目标函数中的惩罚因子,使用两个特别的进化算子,采用了特别的并行计算机制和退回机制。经测试,按所提方案实现的平台,操作方便,求解精度高而稳定,有显著的优越性。所提的优化平台方案是令人满意的

关键词: 平台, 进化算法, 遗传算法, 优化, 数学规划

Abstract: The optimization software, on linear programming and nonlinear programming, is playing more and more important role in social, economic and engineering areas. The current optimization software usually either employs classic local optimization techniques or adopts some simple global optimization techniques. In his paper, the evolutionary algorithm is introduced to the computation of optimization software, called optimization platform. The key procedure of the platform design is analyzed, and the scheme of the platform is proposed, and finally the platform is implemented. The features of the designed platform includes: dynamically adjusting the penalty factor related to the extended objective function at interface, using only two specially designed evolution operators, having a particular mechanism of parallel computing, and a going-back mechanism. With several tests, it was indicated that the implemented platform is easy to use, and can achieve precise solutions stably, having outstanding advantages, and so, the proposed scheme is satisfactory.

Key words: platform, evolutionary algorithm, genetic algorithm, optimization, mathematical programming