选矿专家系统环境的总体结构

介绍选矿专家系统环境MPESE的总体结构。综合应用专家系统、人工神经元网络、遗传技术、多媒体技术、软件工程、计算数学、工艺矿物学及矿物加工学，建立了选矿专家系统的开发环境，可用于建造新型选矿专家系统，用以完成矿石可选性预测、流程及设备的选择、生产过程操作指导及机械设备的故障诊断。MPESE系统环境主要由五部分组成：知识库、多媒体库、推理机制、用户界面和应用工具。系统以Microsoft Windows为操作平台，采用32位编程技术。知识库由统一的基本元素－对象组成。系统的推理机制封装在各知识类中，通过属性继承和消息传递来执行推理求解。各推理机制相互独立，提高系统的灵活性。

GENERAL STRUCTURE OF MINERAL PROCESSING EXPERT SYSTEM ENVIRONMENT

The general structure of Mineral Processing Expert System Environment(MPESE)is introduced in this paper By synthesizing expert system theory,artificial neural networks,genetic technology,multimedia technology,software engineering,numerical analysis,processing mineralogy and mineral processing theory,the development theory and methods for mineral processing expert system is studied,and MPESE is established as a tool to build new expert system for ore beneficiability prediction,flow-sheet and equipment selection,production process control,and machines faults diagnosis The MPESE system consists of five parts:knowledge base,multimedia base,inference engine,user interface and application tools By adopting 32 bits programming technology on Microsoft Windows platform,the system is very effective The knowledge base consists of universal basic units The inference mechanisms of the system are encapsulated in each knowledge class,being executed through characteristic inheritance and message transferring Each mechanism is independent by each other due to improved system's flexibility