基于混合遗传算法的经费分配多目标规划研究

为了探索新的经费分配方法和管理模式,建立了一种新的多目标非线性规划优化模型,提出了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,具体是通过交叉、变异来产生后代,后代与父代组成一个扩大群体,然后对这个群体的个体单独进行正交试验,并以正交试验的结果代替试验前个体,再通过排序选出最好的群体组成下一次遗传的父代,得到的种群个体明显优于基本遗传算法的个体。仿真结果表明,该算法收敛寻优能力强,并能产生很多次优解,是一种高效稳定的方法。

The Study of Multi-goal Layout on Outlay Distribution Based on HGA

In order to explore new idea of outlay distribution and management model, a new multi|goal nonlinear optimized model was founded, and a hybrid genetic algorithm based on direct cross experimentation was proposed so as to solve the problem. The direct cross experimentation is done to choice operator, cross operator and aberrance operator in the process of solving problems. Concretely, by the idea of cross and aberrance, the got offspring, and with the forefather, an expended colony was formed, then the direct cross experimentation is used to the lonely unit of the colony, and the results of the experimentation instead of the former units. Then by ordering the best pack of colony is chosen to form the forefather which would be mentioned in the next heredity, and the got units are obviously super to the ones which come from the radical genetic algorithm. The result of simulation shows that this algorithm is good at the ability of constringency and researching super solution, and it will produce super solution several times. In a word, it is an effective and stable idea.