In multiagent reinforcement learning, with different assumptions of the opponents’ policies, an agent adopts quite different learning rules, and gets different learning performances. We prove that, in multiagent doma...In multiagent reinforcement learning, with different assumptions of the opponents’ policies, an agent adopts quite different learning rules, and gets different learning performances. We prove that, in multiagent domains, convergence of the Q values is guaranteed only when an agent behaves optimally and its opponents’ strategies satisfy certain conditions, and an agent can get best learning performances when it adopts the same learning algorithm as that of its opponents.展开更多
Jinping Underground laboratory for Nuclear Astrophysics(JUNA) will take the advantage of the ultra-low background of CJPL lab and high current accelerator based on an ECR source and a highly sensitive detector to dire...Jinping Underground laboratory for Nuclear Astrophysics(JUNA) will take the advantage of the ultra-low background of CJPL lab and high current accelerator based on an ECR source and a highly sensitive detector to directly study for the first time a number of crucial reactions occurring at their relevant stellar energies during the evolution of hydrostatic stars. In its first phase, JUNA aims at the direct measurements of^(25)Mg(p,γ)^(26)Al,^(19)F(p,α)^(16)O,^(13)C(α,n)^(16)O and ^(12)C(α,γ)^(16)O reactions. The experimental setup,which includes an accelerator system with high stability and high intensity, a detector system, and a shielding material with low background, will be established during the above research. The current progress of JUNA will be given.展开更多
文摘In multiagent reinforcement learning, with different assumptions of the opponents’ policies, an agent adopts quite different learning rules, and gets different learning performances. We prove that, in multiagent domains, convergence of the Q values is guaranteed only when an agent behaves optimally and its opponents’ strategies satisfy certain conditions, and an agent can get best learning performances when it adopts the same learning algorithm as that of its opponents.
基金supported by the National Natural Science Foundation of China(Grant Nos.11490560 and 11321064)the National Basic Research Program of China(Grant No.2013CB834406)
文摘Jinping Underground laboratory for Nuclear Astrophysics(JUNA) will take the advantage of the ultra-low background of CJPL lab and high current accelerator based on an ECR source and a highly sensitive detector to directly study for the first time a number of crucial reactions occurring at their relevant stellar energies during the evolution of hydrostatic stars. In its first phase, JUNA aims at the direct measurements of^(25)Mg(p,γ)^(26)Al,^(19)F(p,α)^(16)O,^(13)C(α,n)^(16)O and ^(12)C(α,γ)^(16)O reactions. The experimental setup,which includes an accelerator system with high stability and high intensity, a detector system, and a shielding material with low background, will be established during the above research. The current progress of JUNA will be given.