驱动主义学派下的一类新的遗传算法

New genetic algorithm based on drivenism school

针对人工智能传统的三大学派的各自优缺点总结了驱动主义学派的人工智能思想。通过对遗传学中若干机理的分析及与经典意义上的遗传算法的对比,对传统的遗传算法进行了改进。引入了性别编码、同性竞争、交配评价和人工酶催化驱动技术。以求Rosenbrock函数的极大值为例,对修正后的遗传算法进行了实例分析。分析结果表明,该算法较好地拟合真实的遗传规律,更好地解决传统求解过程中容易陷入局部极值的情形,具有更高的求解效率。

Drivenism artificial intelligence is summarized based on advantages and disadvantages of three traditional academic circles in artificial intelligence. Through discussing genetic mechanism and comparing with traditional genetic algorithm, this paper proposed the improvement for traditional genetic algorithm. Corresponding technology plans including sex code, intro- duced same-sex competition, evaluation for mating and artificial enzymes catalytic driving. A problem for maximum value of Rosenbrock function was dealt with by the improved genetic algorithm as an example. The analysis result states that the algorithm can simulate true genetic rules and deal with the local value status in computing process better. The efficiency for dealing with optimal problem is higher.