基于改进遗传算法的RTM注射速率优化

针对目前遗传算法初始种群大多数为随机产生,注射速率优化过程容易早熟或不收敛问题,提出了基于注射速率规则的改进遗传算法。在大量注射速率历史数据基础上,建立了注射速率影响因素决策表,提出了规则相似度计算模型。由基于规则的种群生成算子生成初始种群,以填充质量最优为目标,并构造适应度函数,然后进行遗传操作,最后采用面向对象编程语言实现该算法。实例表明该算法比标准遗传算法收敛更快,而且在用该算法优化得到的注射速率下的充填质量比在用标准遗传算法优化得到的注射速率下的充填质量更好,说明采用改进遗传算法优化注射速率更为合理和可靠。

RTM Injection Rate Optimization Based on Improved Genetic Algorithm

Generating initial population of GA randomly in optimization process of injection rate always leads to the premature convergence or misconvergence of the algorithm. Concerning the problem, an improved GA was proposed based on injection rate rules. A decision table about influence factors on injection rate was established according to the generous RTM numerical simulation historical data, and the similarity calculation model of injection rate rules was present in the article. After similarity calculation, the initial population was obtained by the rule-based population generating operator. The paper chooses filling quality as the evaluation criteria, constructs a new fitness function, then conducts genetic operation. Finally, the new algorithm was programed with OOPL. The case showed the convergence speed of improved GA is faster than that of standard GA and the filling quality at the optimized injection rate by improved GA is better than by standard GA.