摘要
在传统遗传规划中引入多目标优化原理,探索新的经费分配方法和管理模式,建立了一种多目标优化的非线性遗传规划模型,提出了一种先进的基于正交试验的新型混合遗传算法来求解该问题。对求解过程中的选择算子、交叉算子和变异算子等进行正交试验,得到的种群个体明显优于基本遗传算法的个体。这种基于多目标优化的遗传规划模型能产生精度更高的最优解,通过对经费分配问题的实验验证,得到了较好的结果。
The multiobjective optimization principle is applied to genetic programming （GP）, an assignment model equipment fund＇s management and its resolving method based on multiobjective optimization are presented. A new GP is put forward, which covers merits ofGP and orthogonal design method. Then the modified ＂Elitist Model＂, the ＂self-crossover operator＂ and ＂increase ＆ decrease mutation operator＂ are designed. The coding method, fitness function and initial population which are fit for the problem are discussed. This GP model based on multiobjective optimization is presented as a tool for optimal results with higher precision, tests of fund＇s assignment problem show this model is better than traditional GP.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第7期1620-1623,共4页
Computer Engineering and Design
基金
湖南省杰出中青年专家科技基金项目(02JJYB012)
教育部重点科研基金项目(02A056)。
关键词
遗传规划
多目标优化
正交试验
混合遗传算法
最优解
genetic programming
multiobjective optimization
orthogonal design method
hybrid genetic algorithms
optimizing resolution
