为促进居民用户柔性负荷高效参与需求响应,帮助用户从被动角色转变为主动角色,实现需求侧最大效益。该文在智能电网环境下,根据用电设备的特性,以概率论的角度对家电设备状态进行描述定义,基于异步深度强化学习(asynchronous deep reinf...为促进居民用户柔性负荷高效参与需求响应,帮助用户从被动角色转变为主动角色,实现需求侧最大效益。该文在智能电网环境下,根据用电设备的特性,以概率论的角度对家电设备状态进行描述定义,基于异步深度强化学习(asynchronous deep reinforcement learning,ADRL)进行家庭能源管理系统调度的在线优化。学习过程采用异步优势演员–评判家(asynchronous advantage actor-critic,A3C)方法,联合用户历史用电设备运行状态的概率分布,通过多智能体利用CPU多线程功能同时执行多个动作的决策。该方法在包括光伏发电、电动汽车和居民住宅电器设备信息的某高维数据库上进行仿真验证。最后通过不同住宅情境下的优化决策效果对比分析可知,所提在线能耗调度策略可用于向电力用户提供实时反馈,以实现用户用电经济性目标。展开更多
In order to optimize the transitional time during the successive exposure scans for a step-and-scan lithography and improve the productivity in a wafer production process, an investigation of the motion trajectory pla...In order to optimize the transitional time during the successive exposure scans for a step-and-scan lithography and improve the productivity in a wafer production process, an investigation of the motion trajectory planning along the scanning direction for wafer stage was carried out. The motions of wafer stage were divided into two respective logical moves （i. e. step-move and scan-move） and the multi-motionoverlap algorithms （MMOA） were presented for optimizing the transitional time between the successive exposure scans. The conventional motion planning method, the Hazelton method and the MMOA were analyzed theoretically and simulated using MATLAB under four different exposure field sizes. The results show that the total time between two successive scans consumed by MMOA is reduced by 4.82%, 2.62%, 3.06% and 3.96%, compared with those of the conventional motion planning method; and reduced by 2.58%, 0.76%, 1.63% and 2.92%, compared with those of the Hazehon method respectively. The theoretical analyses and simulation results illuminate that the MMOA can effectively minimize the transitional step time between successive exposure scans and therefore increase the wafer fabricating productivity.展开更多
基金the National Basic Research Program of China(No.2003CB716206)the National Natural Science Foundation of China(No.50605025)
文摘In order to optimize the transitional time during the successive exposure scans for a step-and-scan lithography and improve the productivity in a wafer production process, an investigation of the motion trajectory planning along the scanning direction for wafer stage was carried out. The motions of wafer stage were divided into two respective logical moves （i. e. step-move and scan-move） and the multi-motionoverlap algorithms （MMOA） were presented for optimizing the transitional time between the successive exposure scans. The conventional motion planning method, the Hazelton method and the MMOA were analyzed theoretically and simulated using MATLAB under four different exposure field sizes. The results show that the total time between two successive scans consumed by MMOA is reduced by 4.82%, 2.62%, 3.06% and 3.96%, compared with those of the conventional motion planning method; and reduced by 2.58%, 0.76%, 1.63% and 2.92%, compared with those of the Hazehon method respectively. The theoretical analyses and simulation results illuminate that the MMOA can effectively minimize the transitional step time between successive exposure scans and therefore increase the wafer fabricating productivity.
