模型参数优化中的混合方法及其应用

归纳了现代典型混合软计算法的构成,指出其或松或紧融合多种单一要素来提供具有互补性的推理与搜索方法,是解决知识驱动推理或纯数据驱动推理问题的有效变通手段.通过知识域与经验数据的结合增加了模型可靠度和问题解决能力.为了解决最广泛和高复杂性的工程或社会问题,应开发有效整合各类推理搜索技术的智能工具箱,以获得低成本解决方案,并举例说明了混合软计算的应用.

Hybrid Idea of Parameters Optimization in Models and Their Applications

Some hybrid idea of soft computing and their difference and features is briefly discussed. It is pointed out that modern hybrid systems are usually based on a loose or tight integration of constituent technologies. This type of integration provides complementary reasoning and searching methods that allow us to combine domain knowledge and empirical data to develop flexible computing tools and solve complex problems.