模拟退火算法在连续变量全局优化问题中应用  

Application of Simulated Annealing to Global Optimization Problems with Continuous Variables

作　　者：江加和[1] 宋子善[1] 沈为群[1] 邱力为[1] 

机构地区：[1]北京航空航天大学自动控制系

出　　处：《北京航空航天大学学报》2001年第5期556-559,共4页Journal of Beijing University of Aeronautics and Astronautics

摘　　要：研究了基于模拟退火算法应用于连续变量全局优化问题 ,并给出了实现步骤 .介绍了控制参数实用选择方法 .针对连续变量的特殊性 ,给出了新解产生的实用方法 .最后以计算机视觉领域中的基本矩阵的求解作为一个实例 ,来说明它在连续变量中的应用 .大量数字仿真结果表明该算法能有效地解决连续变量全局优化问题 .Global optimization problems with continuous variables were discussed based on simulated annealing. Implementation steps were given. Practical methods of determining control parameters were introduced. According to characteristics of continuous variables, practical methods of generating new points from a current point were given. Finally, an example of computing the fundamental matrix in computer vision domain was given to demonstrate the application to continuous variables. The simulation results showed that the algorithm can efficiently solve the global optimization problems with continuous variables.

关 键 词：金属退火 最优化算法 随机搜索 基本矩阵 模拟退火算法 连续变量 

分 类 号：TG156.2[金属学及工艺—热处理] O242.1[金属学及工艺—金属学]