This work proposes a reinforcement learning(RL)approach to tackle the control problem of branch overload relief in large power systems.Accordingly,a control agent is trained to change generators'real power output ...This work proposes a reinforcement learning(RL)approach to tackle the control problem of branch overload relief in large power systems.Accordingly,a control agent is trained to change generators'real power output in order to relieve the stressed branches.For large power systems,this control problem becomes one whose decision space(i.e.,the action space)is both highly-dimensioned and continuous.This makes it extremely difficult to have successful training for RL-based agents.To improve the effectiveness,a data-driven and model-based hybrid approach is proposed to optimize the control by combining RL-agent actions and generator shifting factor-driven actions.Accordingly,with the proposed approach the RL-agent successfully trains on large power systems.The proposed design is tested on both the IEEE 118-bus testing system and a 2749-bus real system.The obtained results show that the proposed hybrid approach outperforms the data-driven training approach.展开更多
Inspecting and testing code changes typically require a significant amount of developer effort.As a system evolves,developers often create composite changes by mixing multiple development issues,as opposed to addressi...Inspecting and testing code changes typically require a significant amount of developer effort.As a system evolves,developers often create composite changes by mixing multiple development issues,as opposed to addressing one independent issue—an atomic change.Inspecting composite changes often becomes time-consuming and error-prone.To test unrelated edits on composite changes,rerunning all regression tests may require excessive time.To address the problem,we present an interactive technique for change decomposition to support code reviews and regression test selection,called ChgCutter.When a developer specifies code change within a diff patch,ChgCutter partitions composite changes into a set of related atomic changes,which is more cohesive and self-contained regarding the issue being addressed.For composite change inspection,it generates an intermediate program version that only includes a related change subset using program dependence relationships.For cost reduction during regression testing,it safely selects only affected tests responsible for changes to an intermediate version.In the evaluation,we apply ChgCutter to 28 composite changes in four open source projects.ChgCutter partitions these changes with 95.7% accuracy,while selecting affected tests with 89.0% accuracy.We conduct a user study with professional software engineers at PayPal and find that ChgCutter is helpful in understanding and validating composite changes,scaling to industry projects.展开更多
电子商务是一个正在快速增长的领域,如今的消费者和企业已逐渐摒弃现金交易,开始拥抱数字支付所带来的便利。放眼亚太地区,中国的数字支付市场遥遥领先。根据Pay Pal 2017年发布的白皮书《数字支付:超越交易的思考》显示:86%的受访中...电子商务是一个正在快速增长的领域,如今的消费者和企业已逐渐摒弃现金交易,开始拥抱数字支付所带来的便利。放眼亚太地区,中国的数字支付市场遥遥领先。根据Pay Pal 2017年发布的白皮书《数字支付:超越交易的思考》显示:86%的受访中国消费者已经开始使用数字支付方式,比例远高于亚洲平均水平(58%),在受访的7个市场中位列首位。展开更多
基金This work was supported by the Science and Technology Project of State Grid Corporation of China(No.5100-201958522A-0-0-00).
文摘This work proposes a reinforcement learning(RL)approach to tackle the control problem of branch overload relief in large power systems.Accordingly,a control agent is trained to change generators'real power output in order to relieve the stressed branches.For large power systems,this control problem becomes one whose decision space(i.e.,the action space)is both highly-dimensioned and continuous.This makes it extremely difficult to have successful training for RL-based agents.To improve the effectiveness,a data-driven and model-based hybrid approach is proposed to optimize the control by combining RL-agent actions and generator shifting factor-driven actions.Accordingly,with the proposed approach the RL-agent successfully trains on large power systems.The proposed design is tested on both the IEEE 118-bus testing system and a 2749-bus real system.The obtained results show that the proposed hybrid approach outperforms the data-driven training approach.
文摘Inspecting and testing code changes typically require a significant amount of developer effort.As a system evolves,developers often create composite changes by mixing multiple development issues,as opposed to addressing one independent issue—an atomic change.Inspecting composite changes often becomes time-consuming and error-prone.To test unrelated edits on composite changes,rerunning all regression tests may require excessive time.To address the problem,we present an interactive technique for change decomposition to support code reviews and regression test selection,called ChgCutter.When a developer specifies code change within a diff patch,ChgCutter partitions composite changes into a set of related atomic changes,which is more cohesive and self-contained regarding the issue being addressed.For composite change inspection,it generates an intermediate program version that only includes a related change subset using program dependence relationships.For cost reduction during regression testing,it safely selects only affected tests responsible for changes to an intermediate version.In the evaluation,we apply ChgCutter to 28 composite changes in four open source projects.ChgCutter partitions these changes with 95.7% accuracy,while selecting affected tests with 89.0% accuracy.We conduct a user study with professional software engineers at PayPal and find that ChgCutter is helpful in understanding and validating composite changes,scaling to industry projects.
文摘电子商务是一个正在快速增长的领域,如今的消费者和企业已逐渐摒弃现金交易,开始拥抱数字支付所带来的便利。放眼亚太地区,中国的数字支付市场遥遥领先。根据Pay Pal 2017年发布的白皮书《数字支付:超越交易的思考》显示:86%的受访中国消费者已经开始使用数字支付方式,比例远高于亚洲平均水平(58%),在受访的7个市场中位列首位。