《Tab.1 Tests on typical mathematical problems》
In order to test the performance of MSEGO algorithm,four classical test functions are selected to solve the optimized values by MSEGO.The test functions are all nonlinear problems with multiple peak values.The results of the tests are summarized in Table 1,where EGO is the traditional algorithm.In addition,Table2 compares the total compute samples and iterations between MSEGO and EGO algorithms.In all tests,the MSEGO algorithm can achieve the global optimum,which is closer to analytical values than EGO with a maximum deviation of 0.7%.As for EGO,its deviations are usually larger than MSEGO and for the Shekel function,it failed to find the global optimum.Although the total number of samples are higher for MSEGO,the number of iterations reduces significantly.Due to parallel computation,it saves computation time.Thus,the MSEGO algorithm has a global search ability at low computational costs.
图表编号 | XD0017592100 严禁用于非法目的 |
---|---|
绘制时间 | 2018.10.26 |
作者 | 李琛玺、武晓龙、祝培源、宋立明、李军、丰镇平 |
绘制单位 | Institute of Turbomachinery, Xi'an Jiaotong University、Navy comsumer representative office of engine in shenyang area、Institute of Turbomachinery, Xi'an Jiaotong University、Institute of Turbomachinery, Xi'an Jiaotong University、Institute of Turbomachin |
更多格式 | 高清、无水印(增值服务) |