摘要: 针对烧结配矿中传统试错法难以求出最优解的问题,综合考虑品位、成本、库存、国外矿比例等多种影响因素,建立其优化配比的线性规划模型,用带罚函数的一阶梯度法完成模型求解。并在构建知识库和模型库的基础上,引入专家系统的解释器和推理机制,建立一套基于专家系统的多角色决策支持系统。实验结果表明,该系统在满足烧结配矿工艺要求的基础上,能降低生产成本7.97%。
关键词:
多角色,
决策支持系统,
配料优化,
专家系统,
一阶梯度法
Abstract: Aiming to the difficulties to obtain optimal solutions to sintering ore blending problem by using the traditional trial-fault-trial method, this paper takes several influencing factors into account, such as the component, cost, storage, overseas ore proportion and so on, builds an optimal model for sintering ore blending, and solves the model by using the first-order gradient method with penalty function. An expert system based multi-role decision support system is constructed by introducing the interpreter and reasoning mechanism of expert system based on the foundation of knowledge base and model base. Comparison results with practical running data show that the DSS can effectively reduce 7.97 percent of the production cost on the basis of meeting the requirement of sintering ore blending.
Key words:
multi-role,
decision support system,
blending optimization,
expert system,
first-order gradient method
中图分类号:
张瑞军, 黄彦. 基于专家系统的多角色烧结配矿决策支持系统[J]. 计算机工程, 2011, 37(19): 233-235,238.
ZHANG Rui-Jun, HUANG Pan. Multi-role Decision Support System for Sintering Ore Blending Based on Expert System[J]. Computer Engineering, 2011, 37(19): 233-235,238.