归一化实数编码的多维并行遗传算法

给出了归一化多维实数编码的基本定义,并在此基础上提出了基于归一化实数编码的多维并行遗传算法;对归一化实数编码多维并行交叉算子、多维并行变异算子进行了详细的研究;提出了多维优化问题归一化实数编码长度计算公式;对遗传算法的控制参数确定进行了阐述;对归一化实数编码的多维并行遗传算法适应度函数的确定方法进行了研究.实验表明,归一化实数编码多维并行遗传算法可以大大提高多维优化问题的收敛速度,从而进一步提高算法的性能,这些特点对于计算复杂的非线性多维优化问题具有重要的意义.

Multi-Dimension Parallel Genetic Algorithms Based on Normalization Real Number Encoding

This paper gives the definition of multi-dimension normalization real number encoding,then proposes parallel genetic algorithms based on normalization real number encoding,introduces the operator mechanism of multi-dimension parallel crossover and mutation,presents the calculating method of code length,describes the confirming method of control parameters,and studies fitness function for parallel genetic algorithms.The result of multi-dimension function optimization test shows that the algorithm can obtain global optimization with fast speed,it can further enhance algorithm's performance,it has important meaning for calculating complex problem of nonlinear multi-dimension optimization.