基于Multi-Agent技术的露天矿山生产调度系统框架的构建  

Framework of Open-pit Mine Production Scheduling System Based on Multi-Agent Technology

作　　者：霍晓宇[1] 杨仕教[1] 戴剑勇[1] 吴长振[1] 

机构地区：[1]南华大学核资源与核燃料工程学院,湖南衡阳421001

出　　处：《南华大学学报（自然科学版）》2012年第3期33-38,共6页Journal of University of South China：Science and Technology

基　　金：国家自然科学基金资助项目(50974076);湖南省自然科学基金资助项目(10JJ2041)

摘　　要：针对于露天矿山生产调度系统的复杂非线性、工艺点多面广等特性,引进具有分布式管理特性的多智能体技术,建立基于多智能体技术的露天矿山生产调度系统.在系统中将生产调度分为任务Agent、生产调度Agent、爆破Agent、运输Agent、破碎Agent以及资源Agent六个单智能体.运用具有较强非线性拟合能力的自适应神经模糊推理系统(ANFIS)进行局部与整体的建模,并利用粒子群优化算法对模型进行优化,得出最优化生产调度方案.对矿山实际生产调度方案的制定具有一定的参考价值.Considering the characteristics of complex nonlinear and processes over a broad area of the open-pit mine production scheduling system, the multi-agent technology with distributed management features is introduced to establish the production scheduling system of open-pit mine, which is based on multi-agent technology. The production scheduling sys- tem is composed of tasks agent, production scheduling agent, blasting agent, transportation agent, crushing agent and resource agent. The ANFIS with the stronger nonlinear curve fit- ting capability is used to model the local and global system. The model is optimized by par- ticle swarm optimization algorithm, and then the optimal production scheduling program is obtained. The program provides certain reference value for the enactment of the actual mineproduction scheduling program.

关 键 词：多智能体技术 生产调度系统 自适应神经模糊推理系统 粒子群优化算法 

分 类 号：TD87[矿业工程—非金属矿开采]