文档介绍：摘 要
基于距离测度的实数编码遗传模拟退火算法是 2004 年蔡良伟提出的一种具有
全局搜索能力的新算法。本文对该算法进行了改进，定义了基于距离测度和适应度
的自适应变异概率，采用改进的交叉策略和模拟退火操作，并在群体趋于一致时保
留最优个体，重新产生其它新个体。对带边界约束函数优化问题进行仿真计算，结
果表明改进后算法收敛概率较高，收敛速度较快。针对 TSP 问题，本文定义了基于
距离测度和适应度的的自适应交叉概率，采用改进的交换启发交叉算子，产生不差
于父代的个体，根据最优和次优个体的差异，采用 2 变换法产生新个体或者进行模
拟退火操作。仿真实验结果表明新算法收敛概率较高，得到最优解的平均代数要远
远小于其他所比较的算法。
关键词：遗传算法，模拟退火算法，距离测度，适应度，函数优化
ABSTRACT
Real-coded genetic annealing algorithm based on distance measurement is a new
algorithm for global-searching which proposed by Cai Liang Wei in 2004. The algorithm has
been improved in this paper, it defined the mutation probability according to distance
measurement and fitness, adopted improved crossover and simulated annealing operation. In
addition, when population tended to uniform, it reserved the best individual and reproduced
other individuals. Applying it to the function optimization problem with boundary constraints,
the simulated results show that the improved algorithm is more effective in realizing the high
convergence probability and rapid evolving speed. In order to solve TSP problem, it defined
the crossing probability based on distance measurement and fitness, adopted improved
exchange heuristic crossover operator, which the new individual is no worse than the old one.
According to the difference between the best individual and the better individual, it adopted
the method that use double exchange produce new individuals or do simulated annealing
operation. The simulated results show that the new algorithm has high convergence
probability and the generation numbers is far less than other comparative algorithms .
Shen Honglian (Department of Mathematics and Physics)
Directed by prof. Zhang Gu