基于混合编码遗传算法的证券组合投资问题的求解

Solving the Problem of Portfolio Investment Based on Mixed-Coding Genetic Algorithms

针对文献[1]提出的模型,设计了一种二进制编码和实数编码相结合的混合编码遗传算法(MCGA),仿真结果表明,混合编码遗传算法对该模型求解能极大地缩短算法的进化代数及保证收敛到最优解,并得到了比二次规划算法更好的结果;证明了这种混合编码的方法能有效克服二进制编码和实数编码各自的缺陷,快速提高遗传算法的收敛性能。

In order to solving the problem of portfolio investment, the paper design the Mixed-Coding Genetic Algorithms（MCGA）, which combined binary coding with float eoding. The simulation experiment is proved that MCGA can cut the evolution of Genetic Algorithms greatly and assure obtain the optimization, and better result is obtained than two programming. By the comparison, the advantage which combined binary coding with float coding and increasing convergence is proved.