A new genetic algorithm is presented based on the musical performance. The novelty of this algorithm is that a new genetic algorithm, mimicking the musical process of searching for a perfect state of harmony, which in...A new genetic algorithm is presented based on the musical performance. The novelty of this algorithm is that a new genetic algorithm, mimicking the musical process of searching for a perfect state of harmony, which increases the robustness of it greatly and gives a new meaning of it in the meantime, has been developed, Combining the advantages of the new genetic algorithm, simplex algorithm and tabu search, a hybrid algorithm is proposed. In order to verify the effectiveness of the hybrid algorithm, it is applied to solving some typical numerical function optimization problems which are poorly solved by traditional genetic algorithms. The experimental results show that the hybrid algorithm is fast and reliable.展开更多
料筒温度是影响注塑成型产品质量的关键参数之一,比例积分微分(proportional integral differential,PID)控制器在料筒温度控制系统中广泛应用。然而,PID控制器参数优化依赖操作者经验,存在成本高、效率低、精度差的问题。为解决上述问...料筒温度是影响注塑成型产品质量的关键参数之一,比例积分微分(proportional integral differential,PID)控制器在料筒温度控制系统中广泛应用。然而,PID控制器参数优化依赖操作者经验,存在成本高、效率低、精度差的问题。为解决上述问题,围绕料筒温度PID控制器参数优化展开研究。首先,针对料筒温度控制系统性能优化,提出性能评价指标,设计PID控制器参数优化框架;然后,基于知识指引型数据驱动优化策略的思想,对单纯形搜索算法进行改进,提出了基于历史梯度近似的知识指引型单纯形搜索算法,以提高PID控制器参数优化效率。实验结果表明,与单纯形搜索算法相比,改进后的算法以牺牲少量优化精度为代价,显著提升了优化效率,降低了PID控制器参数优化成本。展开更多
基金Supported by the National Natural Science Foun-dation of China (60273075)
文摘A new genetic algorithm is presented based on the musical performance. The novelty of this algorithm is that a new genetic algorithm, mimicking the musical process of searching for a perfect state of harmony, which increases the robustness of it greatly and gives a new meaning of it in the meantime, has been developed, Combining the advantages of the new genetic algorithm, simplex algorithm and tabu search, a hybrid algorithm is proposed. In order to verify the effectiveness of the hybrid algorithm, it is applied to solving some typical numerical function optimization problems which are poorly solved by traditional genetic algorithms. The experimental results show that the hybrid algorithm is fast and reliable.
文摘料筒温度是影响注塑成型产品质量的关键参数之一,比例积分微分(proportional integral differential,PID)控制器在料筒温度控制系统中广泛应用。然而,PID控制器参数优化依赖操作者经验,存在成本高、效率低、精度差的问题。为解决上述问题,围绕料筒温度PID控制器参数优化展开研究。首先,针对料筒温度控制系统性能优化,提出性能评价指标,设计PID控制器参数优化框架;然后,基于知识指引型数据驱动优化策略的思想,对单纯形搜索算法进行改进,提出了基于历史梯度近似的知识指引型单纯形搜索算法,以提高PID控制器参数优化效率。实验结果表明,与单纯形搜索算法相比,改进后的算法以牺牲少量优化精度为代价,显著提升了优化效率,降低了PID控制器参数优化成本。
基金国家自然科学基金(the National Natural Science Foundation of China under Grant No.5027150)湖南省教育厅一般项目(the Common Project of Bureau of Education of Hunan Province under Grant No.05C410)
