記述言語：日本語   出版者・発行元：一般社団法人 人工知能学会  

The diffusion of new technologies such as operating systems can be analyzed using game-theoretic frameworks. While previous studies have primarily focused on the spread of a single technology, in practice, multiple technologies―such as Windows and Unix-based systems (including macOS and Linux)―coexist. In this study, we develop a tractable model based on the stochastic block model to investigate the conditions under which such coexistence emerges, taking into account community structures within a network. First, we show that when community structure is strong, the Nash equilibrium shifts from single-technology dominance to coexistence of multiple technologies. Second, we observe critical slowing down near the transition point in the time it takes for the system to reach equilibrium, suggesting the presence of a phase transition (bifurcation phenomenon). Third, we analytically confirm the existence of such bifurcations using dynamical systems theory.

DOI： 10.1527/tjsai.41-2_fn26-d

Scopus

CiNii Research

記述言語：英語   出版者・発行元：一般社団法人 日本航空宇宙学会  

The discharge current and potential difference between cathode and ground of a Hall thruster were predicted sequentially by Recurrent Neural Network (RNN) in order to optimize operating conditions. The prediction accuracy and calculation cost for three RNN models, the standard Echo State Network (simpleESN), cycle-groupedESN, and Long Short-Term Memory (LSTM) were compared. The ESN model structures were chosen using Bayesian optimization. We calculated the normalized root mean square error (NRMSE) of the model output against the experimental results of a 200 W class Hall thruster developed at Kyushu University. The NRMSE of the simpleESN model output against discharge current was 0.0407, about 1/10 that of the LSTM. The NRMSE of the simpleESN model against the potential difference between cathode and ground was 0.0981, about 1/7 that of the LSTM. Moreover, cycle-groupedESN reduced the calculation time for the optimization process to 24 seconds, compared to 303 seconds with simpleESN, though the NRMSE against discharge current of the cycle-grouped ESN was 0.1179. These results show that the simpleESN and cycle-groupedESN models are superior to the LSTM in both prediction accuracy and calculation costs for prediction of discharge current and voltage between ground and cathode in Hall thrusters in the considered settings.

DOI： 10.2322/tjsass.67.1

Scopus

CiNii Research

記述言語：英語   出版者・発行元：宇宙技術および科学の国際シンポジウム  

<p>Artificial neural networks are used to predict discharge current, a series of the performance of Hall thrusters. To reduce the calculation cost of predictions, we used an Echo State Network (ESN), a model of reservoir computing, for the prediction of the one-step-ahead and ten-step-ahead discharge current dependency on magnetic field strength using operation parameters and plume images as an input data set. The mean absolute percentage error was 1.70% and its calculation time was about 1.0 second, which was about 1/20,000 in comparison with our previous study using Feedforward Neural Network and Convolutional Neural Network.</p>

DOI： 10.57350/jesa.3

CiNii Research