论文部分内容阅读
为了增强和声搜索算法在求解高维多模态问题时的空间全局探索能力和求解精度,通过定义的4种反向计算方法和高斯分布估计算法,提出一种动态自适应高维和声搜索算法.该算法采用正交试验初始化和声记忆库;利用多维动态自适应算法进行和声创作;采用动态反向选择算法更新和声记忆库,并改进和声音调微调调解步长,从而增强算法的空间探索能力,避免陷入局部搜索.通过6个标准的高维Benchmark函数测试表明,本文算法在全局搜索能力、收敛速度和求解精度等方面都有明显改进.
In order to enhance the spatial exploration ability and accuracy of the harmonic search algorithm in solving high-dimensional multi-modal problems, a dynamic adaptive high-dimensional harmonic search algorithm is proposed based on the four inverse calculation methods and the Gaussian distribution estimation algorithm The algorithm uses orthogonal test to initialize the sound memory bank, makes use of multidimensional dynamic adaptive algorithm for harmony creation, uses the dynamic reverse selection algorithm to update the harmony memory bank and improves the harmonics tuning step of the harmonics, so as to enhance the performance of the algorithm Space exploration ability to avoid falling into local search.Through six standard high-dimensional Benchmark function tests, it shows that the proposed algorithm has obvious improvement in global search ability, convergence speed and solution precision.