论文部分内容阅读
在现有文献研究的基础上,对传统实数遗传算法的进化策略又作了进一步研究,提出了一种改进的进化策略.进化策略克服了传统实数遗传算法中交叉得到的优秀个体有可能在变异过程中遭到破坏而不能生存的不足,并取消了交叉概率,使交叉产生的个体数增多,这样可增大产生更优秀个体的可能性,因而可使实数遗传算法的性能得到更好的改善.另外,给出了一种计算种群中个体适应度的计算公式和计算方法.该方法不但使得遗传算法具有较强的局部搜索能力,而且具有较强的广域搜索能力和较好的种群多样性,不易陷入局部最优解,从而可快速收敛到全局最优解.5个测试函数的计算结果表明,给出的实数遗传算法的改进进化策略比传统实数遗传算法进化策略的运算速度明显提高,迭代次数明显减少,从而验证了提出的实数遗传算法改进进化策略的有效性.
Based on the existing literatures, the evolutionary strategy of traditional real-number genetic algorithm is further studied and an improved evolutionary strategy is proposed.The evolutionary strategy overcomes the possibility of crossover The process of being destroyed and can not survive, and to eliminate the crossover probability, so that the number of cross-generated individuals, which can increase the probability of producing better individuals, which can make the real genetic algorithm to better performance In addition, a calculation formula and calculation method of individual fitness in population are given, which not only makes genetic algorithm have strong local search ability, but also has strong ability of wide area search and good population diversity Which is not easy to fall into the local optimal solution so that it can converge to the global optimal solution quickly.The calculation results of the five test functions show that the improved evolutionary strategy of real genetic algorithm is faster than that of the traditional real genetic algorithm , The number of iterations decreased obviously, which verified that the proposed real-valued genetic algorithm can improve the effectiveness of evolutionary strategy.