Aim To investigate the model free multi step average reward reinforcement learning algorithm. Methods By combining the R learning algorithms with the temporal difference learning (TD( λ ) learning) algorithm...Aim To investigate the model free multi step average reward reinforcement learning algorithm. Methods By combining the R learning algorithms with the temporal difference learning (TD( λ ) learning) algorithms for average reward problems, a novel incremental algorithm, called R( λ ) learning, was proposed. Results and Conclusion The proposed algorithm is a natural extension of the Q( λ) learning, the multi step discounted reward reinforcement learning algorithm, to the average reward cases. Simulation results show that the R( λ ) learning with intermediate λ values makes significant performance improvement over the simple R learning.展开更多
In spite of the good performance of the steel plate shear wall(SPSW)in recent earthquakes and experimental studies,the need for huge columns to surround the infill plate is a major shortcoming of the system.This short...In spite of the good performance of the steel plate shear wall(SPSW)in recent earthquakes and experimental studies,the need for huge columns to surround the infill plate is a major shortcoming of the system.This shortcoming can be resolved by using semi-supported SPSW.The semi-supported SPSW has secondary columns that prevent the transfer of stress from the infill plate to the main columns.In spite of extensive experimental and numerical investigations on SPSWs,there are many ambiguities regarding the behavior of the semi-supported SPSW.Although stress in the columns is reduced,incomplete diagonal tension field action is formed in the infill plate that creates new problems.In this paper,a new type of semi-supported SPSW is presented in which the steel plate and the secondary columns are angled.The creation of the angle of the plate and the secondary column makes it possible to use the full capacity of the steel plate as well as the capacity of the secondary columns.Numerical results showed that the wall with a 60°angle has a favorable performance relative to the semi-supported wall.Moreover,with the 60°angle,stiffness,strength and energy absorption is increased.The angle of the secondary columns has little effect on the non-elastic stiffness.Nevertheless,using a wall with an angle of more than 90°can neutralize the wall’s behavior relative to conventional walls.Therefore,the wall with a 60°angle as an optimal angle is recommended.展开更多
文摘Aim To investigate the model free multi step average reward reinforcement learning algorithm. Methods By combining the R learning algorithms with the temporal difference learning (TD( λ ) learning) algorithms for average reward problems, a novel incremental algorithm, called R( λ ) learning, was proposed. Results and Conclusion The proposed algorithm is a natural extension of the Q( λ) learning, the multi step discounted reward reinforcement learning algorithm, to the average reward cases. Simulation results show that the R( λ ) learning with intermediate λ values makes significant performance improvement over the simple R learning.
文摘In spite of the good performance of the steel plate shear wall(SPSW)in recent earthquakes and experimental studies,the need for huge columns to surround the infill plate is a major shortcoming of the system.This shortcoming can be resolved by using semi-supported SPSW.The semi-supported SPSW has secondary columns that prevent the transfer of stress from the infill plate to the main columns.In spite of extensive experimental and numerical investigations on SPSWs,there are many ambiguities regarding the behavior of the semi-supported SPSW.Although stress in the columns is reduced,incomplete diagonal tension field action is formed in the infill plate that creates new problems.In this paper,a new type of semi-supported SPSW is presented in which the steel plate and the secondary columns are angled.The creation of the angle of the plate and the secondary column makes it possible to use the full capacity of the steel plate as well as the capacity of the secondary columns.Numerical results showed that the wall with a 60°angle has a favorable performance relative to the semi-supported wall.Moreover,with the 60°angle,stiffness,strength and energy absorption is increased.The angle of the secondary columns has little effect on the non-elastic stiffness.Nevertheless,using a wall with an angle of more than 90°can neutralize the wall’s behavior relative to conventional walls.Therefore,the wall with a 60°angle as an optimal angle is recommended.
