依据已经报道的相关研究,对当前用于烧结配料的各种优化算法进行了对比分析.结果发现,对于关系简单、规模较小的模型,线性规划法方便易用,求解效率高;对于大规模复杂问题,遗传算法、粒子群算法和蚊群算法能灵活有效解决问题;神经网络能够基于大数据对生产结果进行预测;基于神经网络而建立的专家系统能够很好地指导生产实践.未来优化配料应以铁水成本最低及保证烧结矿冶金性能为目标进行多目标优化.
Based on the published results,the methods and algorithms for the optimization of the raw materials blending in sinter process were comparatively analyzed.The results confirmed that linear programming was a convenient and efficient algorithm to solve the common models with small-scale calculations,while genetic algorithm,particle swarm optimization and ant colony algorithm can handle with the complicated models with large-scale calculations.Neural net algorithm was a power method to predict production results with big data.The expert system coupled with neutral net can solve many complicated problems in industrial operation process.The goal of raw materials blending in sinter process should be the lowest cost of hot metal and maintain the quality of sinter in the future.
参考文献
- 下载量()
- 访问量()
- 您的评分：
-
10%
-
20%
-
30%
-
40%
-
50%