一种求解高维复杂优化问题的动态自适应和声搜索算法  

Dynamic Self-adaptive Harmony Search Algorithm for Solving High-dimensional Complex Optimization Problems

作　　者：拓守恒[1] 邓方安[1] 

机构地区：[1]陕西理工学院数学与计算机科学学院,汉中723000

出　　处：《计算机科学》2012年第9期240-243,246,共5页Computer Science

基　　金：国家高技术研究发展计划(863计划)(2008AA01A303);国家自然科学基金(81160183);宁夏自然科学基金项目(NZ11105);宁夏卫生厅科研项目(2011033);陕西理工学院"汉水文化"省级重点学科课题(SLGH1226)资助

摘　　要：为了更好地提高求解高维复杂优化问题的能力,提出一种动态自适应和声搜索(DSHS)算法。该算法采用正交试验来设计算法的初始化和声记忆库;利用多维动态自适应调整算子和单维和声微调算子相结合的策略进行和声创作;改进和声音调调解步长,从而增强算法的扰动能力,避免其陷入局部搜索。通过6个标准Benchmark函数测试表明,该算法在全局搜索能力、收敛速度和稳定性方面都有明显提高。This study presented a dynamic self-adaptive harmony search（DSHS）algorithm to solve high-dimensional optimization problems.In the proposed DSHS algorithm,the orthogonal experimental design algorithm was used to initialize population;two new harmony adjustment operators,multi-dimensional dynamic adaptive adjustment operator and one-dimensional tones fine-tuning operator,were integrated to the improvisation scheme.For avoiding the search being trapped in local optimum,an improved band width adjustment algorithm was employed to enhance the disturbance performance.6benchmark function experiments show that the proposed algorithm has strong convergence velocity,stabilization and capacity of space exploration on solving high-dimensional complex optimization problems,compared with most other approaches.

关 键 词：高维优化问题 动态自适应 和声搜索算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]