摘要
畜禽饲养对饲料养分的需求越来越高,饲料配方需要处理的原料种类不断扩大,配方算法需要处理的数据非常巨大,因此对传统遗传算法的计算效率提出更高的挑战和要求。文中分析了基于实数编码的遗传算法在种群初始化和交叉、变异操作过程中存在的缺陷,在此基础上,提出了基于经验值引导和及时检查修正联合作用的算法优化改进措施,并以特定猪饲料配方为例,通过仿真实验,验证了改进后的遗传算法在执行效率和解的质量两方面都有显著提升。
The increasing demand of feed nutrient in livestock and poultry is higher and higher,feed formulation needs to deal with the growing variety of raw materials,the data needed to be processed is very large,it brings the computational efficiency of traditional genetic algorithm more challenges and requirements. This paper analyzes the defects,that exists in the process of population initialization,crossover and mutation operation of genetic algorithm based on real-coding,on this basis,it puts forward measures for improvement based on joint action of guidance of empirical value and timely detection and correction. Then,using feed formulation as an example,an emulate experiment is presented. That shows the modified alrorithm has a clear promotion in efficiency and quality of the solution.
出处
《信息技术》
2017年第3期34-36,41,共4页
Information Technology
基金
国家自然科学基金((61502287)
山西省高校科技创新项目(201505)
榆林市科技计划项目(2014CXY-02)
关键词
实数编码
遗传算法
饲料配方
real-coding
genetic alrorithm
feed formulation