论文部分内容阅读
用基于实数 (浮点 )编码的遗传算法 (FGA )求解连续参数优化问题是遗传算法的重要应用领域 ,但一直缺少其完整的收敛性结果 .分析了在绝大多数实际应用中使用的 FGA的收敛性问题 ,在采用最优个体保留策略的前提下得到了保证收敛性的一般条件 ,并以之检验了采用常用的变异与交叉算子时 FGA的收敛性 .这些结论不仅给出了有用的收敛性结果 ,而且对进一步认识算子的本质 ,指导自适应变异算子的设计等都提供了帮助 .同时所用的方法和结论也适用于进化规划和进化策略的收敛性问题 .
It is an important application field of genetic algorithm to solve continuous parameter optimization problem with genetic algorithm (FGA) based on real number (floating point) coding, but it has been lacking its complete convergence result.An analysis of the FGA used in the vast majority of practical applications Convergence problem, the general condition of guaranteed convergence is obtained under the premise of using the optimal individual retentive strategy, and the convergence of FGA when using the common mutation and crossover operator is tested.These conclusions not only provide useful But also help to further understand the nature of operators and guide the design of adaptive mutation operators.The methods and conclusions also apply to the convergence of evolutionary programming and evolutionary strategies.