现代电力 ›› 2020, Vol. 37 ›› Issue (4): 399-407.doi: 10.19725/j.cnki.1007-2322.2020.0094

• 智能电网 • 上一篇    下一篇

基于智能体的非玩家角色智能模拟技术及在调控仿真培训中的应用

齐晓琳, 吕闫, 韩昳, 李媛媛, 李志宏, 宋旭日, 狄方春   

  1. 1.中国电力科学研究院有限公司,北京市 海淀区 100192;
    2.电力调度自动化技术研究与系统评价北京市重点实验室(中国电力科学研究院有限公司),北京市 海淀区 100192
  • 收稿日期:2020-02-12 出版日期:2020-08-10 发布日期:2020-08-07
  • 作者简介:齐晓琳(1992),女,硕士,工程师,研究方向:电网调控仿真,E-mail: qixiaolin1992@126.com ;吕闫(1991),女,硕士,工程师,研究方向:电力系统及其自动化,E-mail:lvyan@epri.sgcc.com.cn;韩昳(1994),女,硕士,工程师,研究方向:电网调控仿真,E-mail:929810632@qq.com;李媛媛(1991),女,硕士,工程师,研究方向:电力系统自动化,E-mail:yylnff@sina.com;李志宏(1981),男,工程师,研究方向:电力系统自动化,E-mail:18600119127@163.com;宋旭日(1987),男,高级工程师,研究方向:电力系统机器自动化,E-mail:songxuri@epri.sgcc.com.cn;狄方春(1978),男,高级工程师,研究方向:计算机技术及应用,E-mail:difangchun@epri.sgcc.com.cn
  • 基金资助:
    国家重点研发计划项目 (2018YFB0904501);国家电网有限公司科学技术项目(以数据为中心的调控云广域服务平台及典型应用服务研究5442DZ180018)

BDI Agent Based Non-Player-Character Intelligent Simulation Technology and Its Application in Dispatcher Training Simulator

QI Xiaolin, LV Yan, HAN Yi, LI Yuanyuan, LI Zhihong, SONG Xuri, DI Fangchun   

  1. 1. China Electric Power Research Institute, Haidian Ditrict, Beijing 100192, China;
    2. Beijing Key Laboratory of Research and System Evaluation of Power Dispatching Automation Technology(China Electric Power Research Institute), Haidian District, Beijing 100192, China
  • Received:2020-02-12 Online:2020-08-10 Published:2020-08-07
  • Supported by:
    National Key R&D Program of China (2018YFB0904501); Science and Technology Program of State Grid Corporation of China (5442DZ180018)

摘要: 针对多级调控联合仿真培训系统(Multi-Level Dispatcher Training Simulator, DTS)中跨域广、多用户多角色沟通冲突及无法模拟调度员下令过程等问题,提出了基于BDIAgent的非游戏玩家(Non-Player-Character, NPC)智能模拟处理方法。通过分析游戏中的NPC特点,将其与DTS系统进行融合,构建基于BDIAgent的NPC模型,设计调控仿真培训的行为树与行为子树,运用消息事件触发技术实现多个NPC间的消息交互及电网事件自动触发。通过与传统仿真系统的流程作对比,验证了非游戏玩家智能模拟技术在多级调控系统联合仿真培训中的可行性和实用性,实现了调控仿真培训从人工、统一培训方式过渡到智能、差异化培训方式。

关键词: 调控培训仿真系统, 非游戏玩家, 行为树, 智能体, 消息事件触发

Abstract: To remedy such imperfections in multi-level dispatcher training simulator (DTS) as wide cross-domain, communication conflicts among multi-users and multi roles and unable to simulate the instructing process of the dispatcher, a non-player-character (NPC) intelligent simulation method based on BDI Agent was proposed. By means of analyzing the NPC features in the game and integrating it with DTS, a BDI Agent based NPC model was constructed, the behavior tree and sub-behavior tree for dispatcher training simulator was designed and the message event triggering technology was utilized to implement message interaction among multi-NPCs and automatic triggering of grid events. Comparing the proposed method with the process of traditional simulation system, both feasibility and practicality of NPC intelligent simulation technology in multi-level DTS are verified, thus the traditional artificial and unified training mode of dispatchers can be replaced by intelligent and differentiation training mode.

Key words: Dispatcher Training Simulator(DTS), Non-Player-Character, Behavior Tree, Agent, Message Event Trigger

中图分类号: 

  • TM73