摘要
为解决蚁群算法在初始阶段执行效率低、信息素随机分布、路径杂乱无章的缺点,提出将正交设计方法引入初始优化中。创建正交离散过程,形成正交优化的路径设置;优化初始化过程,形成初始解;以动态概率转移规则来构造路径;精练的选路策略等4项改进措施的初始路径优化模型。该模型提高了算法的执行效率,模拟算例成功应用于连续域问题的饲料配方设计方面,表明该算法有效且可行,开辟了一条饲料配方设计的新途径,同时对蚁群算法解决连续域问题提供可参考的模型和求解方法。
To solve the problems of low efficiency, randomly distributed pheromone, scrambled paths in the initial stage of the ant colony algorithm, several methods are adopted to solve the optimization model that orthogonal design is proposed to create orthogonal discrete points and continuous optimization course is formulated as discrete optimization and is designed to optimize the initial route set in order to get initial solution. Dynamical transfer rules are used to construct the probably paths and other refined path set strategies to enhance the implementation of the algorithm efficiency. And then a feed formula design simulation is successfully presented to prove feasibility and validity of the proposed method. Consequently, the study offered referenced model opens up a new way to feed formula design meanwhile providing a useful idea to solve the continual domain problem.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第5期1031-1034,共4页
Computer Engineering and Design
基金
国家自然科学基金项目(60743010)
山东省自然科学基金项目(Z2008G02)
关键词
蚁群算法
初始路径优化
正交设计
正交离散
连续优化
饲料配方设计
ant colony algorithm
initial route optimization
orthogonal design
orthogonal discrete
continuous optimization
feed formula design
