Advanced direct torque control based on neural tree controllers for induction motor drives.

This paper introduces a novel direct torque control approach based on the decision tree (T-DTC), employing artificial neural networks that are effectively trained to enhance accuracy and robustness. The main objective of T-DTC is the substantial reduction of flux and torque ripples inherent in the conventional DTC, ensuring effective control of the induction motor. The conventional hysteresis controllers for stator flux and electromagnetic torque are replaced by two advanced controllers named M5 Prime model trees. Additionally, the traditional switching table is substituted with a novel decision tree table utilizing the classifier algorithm 4.5. The effectiveness of the proposed T-DTC strategy is demonstrated through simulation in MATLAB/Simulink and validated in real-time using an HIL platform based on OPAL-RT OP 5600 and Virtex 6 FPGA ML605. The results obtained demonstrate a notable improvement compared to existing techniques in the literature.

Reinforcement Learning for Constrained Energy Trading Games With Incomplete Information.

This paper considers the problem of designing adaptive learning algorithms to seek the Nash equilibrium (NE) of the constrained energy trading game among individually strategic players with incomplete information. In this game, each player uses the learning automaton scheme to generate the action probability distribution based on his/her private information for maximizing his own averaged utility. It is shown that if one of admissible mixed-strategies converges to the NE with probability one, then the averaged utility and trading quantity almost surely converge to their expected ones, respectively. For the given discontinuous pricing function, the utility function has already been proved to be upper semicontinuous and payoff secure which guarantee the existence of the mixed-strategy NE. By the strict diagonal concavity of the regularized Lagrange function, the uniqueness of NE is also guaranteed. Finally, an adaptive learning algorithm is provided to generate the strategy probability distribution for seeking the mixed-strategy NE.