AIT Associated Repository of Academic Resources (Aichi Institute of Technology) / 愛知工業大学学術情報リポジトリ
Not a member yet
3244 research outputs found
Sort by
Water quality of the Ohtakigawa River influenced by the Makio Dam and its effects on the water resource management
Get PDFThe water quality of the Ohtakigawa River has been influenced by volcanic activities of Mount Ontake in Japan. In this study, in order to analyze Ohtakigawa water quality affected by artificial infrastructures such as the Makio Dam, we mainly observed suspended solids (SS), Aluminum (Al) and Chlorophyll-a (Chl.a) in the Ohtakigawa river, where SS and Al prevent algae from growing, and Chl.a is contained in algae. In usual discharge cases from the Makio Dam through water pipes, the concentrations of Al and SS were decreased 0.1 times and 0.2 times, respectively, and the concentration of Chl.a was increased 3.3 times. In heavy rain cases, where overflow discharge through flood sluice gates were carried out, their concentrations showed the tendency similar to the above usual cases. And, laboratory experiments showed 80% of SS was deposited. Thereby, it is implied that both sedimentation and dilution effects caused algae inhibitors such as SS and Al to decrease and hence algae grew in the Makio Dam. In conclusion, the Makio Dam is capable of rehabilitation of the Ohtakigawa river water quality
Go Algorithm Based on Deep Learning and Playout
Get PDFThis paper describes a go algorithm based on deep learning and playout. The algorithm runs on a small resource environment which consists of one CPU and one GPU. The best next move can be obtained by using a Value-Monte-Carlo tree search method. It is one of the best-first search methods. The proposed method omits the process of tree policy which has been proposed by AlphaGo. Instead of tree policy, the method adds the top 20 candidates with the highest probability in synchronization with SL policy network as leaves of the node when expanding a leaf node. The win/loss function according to the rollout policy advocated by AlphaGo is substituted by playout, which is commonly used in ordinary Monte-Carlo tree search. As a node evaluation value, not an ordinary UCB1 value but an action value advocated by AlphaGo is adopted. Numerical experiments confirmed the statistical significance of the proposed method and clarified both the best mixing parameter value and the node expansion threshold
