Xiang Y, Gong XG
Institute of Solid State Physics, Academia Sinica, 230031-Hefei, People's Republic of China.
Phys Rev E Stat Phys Plasmas Fluids Relat Interdiscip Topics. 2000 Sep;62(3 Pt B):4473-6. doi: 10.1103/physreve.62.4473.
We have explored the evolution of the efficiency of generalized simulated annealing (GSA) through a comparative study with classical simulated annealing (CSA) and fast simulated annealing (FSA). Our calculations on the Thomson model and nickel clusters show that the relative efficiency of GSA compared to CSA and FSA increases with the number of variables of the objective function. Thus, relative to CSA and FSA, the more complex the system, the more efficient is the GSA method.
我们通过与经典模拟退火算法(CSA)和快速模拟退火算法(FSA)的对比研究,探索了广义模拟退火算法(GSA)效率的演变。我们对汤姆逊模型和镍团簇的计算表明,与CSA和FSA相比,GSA的相对效率随着目标函数变量数量的增加而提高。因此,相对于CSA和FSA,系统越复杂,GSA方法的效率就越高。