欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
为解决后验式场景下的多目标生产调度问题，提出一种基于自组织映射神经网络的策略来生成近似Pareto边界。该方法首先使用拉格朗日松弛法获得若干Pareto解，从而将搜索范围划分为若干区域。对于每一个区域，构造两个并发运行的自组织映射神经网络搜索区域中的Pareto解，在不增加求解时间的情况下提高了求解精度。另外，根据多目标调度问题的特点，改变了神经网络训练过程中邻域的定义，从而加快了求解速度。仿真实验验证了该算法的可行性与有效性。
英文摘要：
To solve the posterior multi-criteria scheduling problems, a new method based on Self-Organization feature Map neural network （SOM） was presented to generate approximate Pareto boundaries. Firstly, Lagrange relaxation algorithm was applied to obtain several Pareto optima which were exploited to divide the search space into several domains. For each domain, two simultaneously running SOM were constructed to explore the remaining Pareto optima. According to the char- acteristics of multi-criteria scheduling, a new definition of neighbor for neural network training was proposed. Numerical experiments demonstrated the feasibility and the effectiveness of this algorithm.
同期刊论文项目
同项目期刊论文
Constrained robust feedback model predictive control for uncertain systems with polytopic descriptio
Sewage Flow Optimization Algorithm for Large-scale Urban Sewer Networks Based on Network Community D
Comment on "Competitive analysis of a better online algorithm to minimize total completion time on a
An Online Algorithm and Its Competitive Analysis for a Single Machine Scheduling Problem with Bounde
期刊信息
位置：