基于数据仓库的多目标优化遗传算法

基于数据仓库的多目标优化遗传算法为解决多目标优化问题提供了有效的途径。其基本思想是:为求Pareto最优解的多目标优化遗传算法建立一个数据仓库,将进化过程中所产生的每一代Pareto最优解放入数据仓库中,在每一代先对数据仓库中的所有个体进行求Pareto最优解运算,淘汰掉劣解,再进行个体间的欧氏距离运算,将小于指定值的其中一个个体作为劣解处理。大量的计算机仿真计算表明,这种算法不仅能够有效地避免交叉或变异操作对Pareto最优解产生的破坏,而且进化速度极快,算法稳定,一般只需20￣40代的运算,即可得到分布广泛的Pareto最优解。

The Multiobjective Optimization Genetic Algorithm Based on Data Warehouse

A multiobjective optimization genetic algorithm based on Data Warehouse is an effective way to solve the problem of multiobjective optimization. The basic idea of the algorithm is: Firstly, set up a Data Warehouse for the algorithm, put Pareto optimum solutions of every generation into it. Secondly calculate Pareto optimum solutions of every individual in the Data Warehouse of every generation, extinguish the non-optimum solutions. Finally calculate Euclidean distance among the individuals, extinguish the one between the individuals as non-optimum solution if its value is less than the designated value. A large number of calculations indicate that not only the algorithm effectively avoids damaging Pareto optimum solutions because of crossover operations or mutation operations, but also the evolution rate is extremely quick and the algorithm is stable, it can obtain extensive Pareto optimum solutions easily but only need 30~50 generations or so.