English version
您所在的位置: 爱游戏app / 学术动态

基于性能退化的最优检测与任务终止决策

爱游戏app

主(zhu)讲人(ren): 赵先(xian)

时间: 11月15日 上午 9:00-10:00

地点(dian): 腾讯会议780628970

主(zhu)讲(jiang)人(ren)简介: 赵先,北京理工(gong)大(da)学(xue)管理与经济(ji)学(xue)院教授(shou)、博士生导师。长期(qi)从事(shi)系统可靠性相关(guan)领(ling)域的研究,在可靠性领(ling)域旗舰期(qi)刊和主(zhu)流期(qi)刊发表或录用SCI/SSCI论(lun)文50余篇。主(zhu)持(chi)国(guo)家自然科(ke)学(xue)基(ji)金重点项(xiang)(xiang)目(mu)1项(xiang)(xiang),面(mian)上项(xiang)(xiang)目(mu)和青年(nian)项(xiang)(xiang)目(mu)4项(xiang)(xiang)。兼任中国(guo)优(you)选法统筹(chou)法与经济(ji)数学(xue)研究会(hui)(hui)工(gong)业工(gong)程分(fen)会(hui)(hui)副理事(shi)长,中国(guo)运筹(chou)学(xue)会(hui)(hui)可靠性分(fen)会(hui)(hui)常(chang)务理事(shi),中国(guo)系统工(gong)程学(xue)会(hui)(hui)系统可靠性工(gong)程专业委员会(hui)(hui)常(chang)务理事(shi)。先后入(ru)选北京市优(you)秀(xiu)人(ren)才支(zhi)持(chi)计划,教育部新(xin)世(shi)纪优(you)秀(xiu)人(ren)才支(zhi)持(chi)计划,获(huo)第十四(si)届霍(huo)英东教育基(ji)金会(hui)(hui)高等院校青年(nian)教师奖。

内容摘要: 安全(quan)关(guan)键系(xi)统(tong)(tong)(tong)被广泛应用于(yu)实际工程(cheng)领域(yu)中(zhong),其(qi)失效(xiao)通常会导致灾难性的(de)(de)(de)后果。因(yin)此(ci),当系(xi)统(tong)(tong)(tong)的(de)(de)(de)故(gu)障风险较高时(shi)(shi),需要终(zhong)(zhong)(zhong)(zhong)止(zhi)任务(wu)(wu)并对系(xi)统(tong)(tong)(tong)展开(kai)救援来降(jiang)(jiang)低失效(xiao)风险。通过(guo)终(zhong)(zhong)(zhong)(zhong)止(zhi)任务(wu)(wu)来提高安全(quan)关(guan)键系(xi)统(tong)(tong)(tong)的(de)(de)(de)生(sheng)存(cun)能力是(shi)以(yi)降(jiang)(jiang)低任务(wu)(wu)可(ke)靠性为代价的(de)(de)(de),因(yin)此(ci)在系(xi)统(tong)(tong)(tong)失效(xiao)特征分(fen)析的(de)(de)(de)基础上,开(kai)展任务(wu)(wu)终(zhong)(zhong)(zhong)(zhong)止(zhi)策略(lve)的(de)(de)(de)优(you)(you)化(hua)设计以(yi)平(ping)衡任务(wu)(wu)可(ke)靠性和(he)系(xi)统(tong)(tong)(tong)生(sheng)存(cun)性具有十分(fen)重要的(de)(de)(de)应用价值(zhi)(zhi)。本(ben)报告提出了基于(yu)性能退化(hua)的(de)(de)(de)周期性检测(ce)和(he)任务(wu)(wu)终(zhong)(zhong)(zhong)(zhong)止(zhi)策略(lve),构建了检测(ce)和(he)任务(wu)(wu)终(zhong)(zhong)(zhong)(zhong)止(zhi)决策的(de)(de)(de)联合优(you)(you)化(hua)模型。基于(yu)系(xi)统(tong)(tong)(tong)的(de)(de)(de)退化(hua)水平(ping)和(he)任务(wu)(wu)执行时(shi)(shi)间,考虑(lv)了动(dong)态(tai)的(de)(de)(de)任务(wu)(wu)终(zhong)(zhong)(zhong)(zhong)止(zhi)决策。利(li)用马(ma)尔(er)可(ke)夫决策过(guo)程(cheng)对该优(you)(you)化(hua)问题(ti)进行了建模与分(fen)析,通过(guo)选择最优(you)(you)决策(终(zhong)(zhong)(zhong)(zhong)止(zhi)或继续执行任务(wu)(wu))以(yi)最小化(hua)检测(ce)、任务(wu)(wu)失败和(he)系(xi)统(tong)(tong)(tong)失效(xiao)的(de)(de)(de)期望总(zong)成本(ben)。推导了最优(you)(you)策略(lve)的(de)(de)(de)解析性质,同时(shi)(shi)对几种启发式策略(lve)进行了数值(zhi)(zhi)评估(gu),给出了在不同策略(lve)下的(de)(de)(de)任务(wu)(wu)可(ke)靠度(du)和(he)系(xi)统(tong)(tong)(tong)生(sheng)存(cun)度(du)。

爱游戏app
    下一篇: 没有了
爱游戏app