Optimal Q-Learning Approach for Tuning the Cavity Filters

摘要

The current research paper elucidates the optimal approach for tuning the Microwave Cavity Filters. The proposed solution uses Q-learning approach which is a special case of Temporal Difference used in Reinforcement Learning. The results are optimized using Lagrangian Multiplier. The proposed algorithm is tested on a commercially used filter. In this research work, only four screws were used for training the algorithm and for testing it. The algorithm could understand the strategies and could tune the reflection characteristics of the considered filter in 43 steps which proves the effectiveness of the algorithm to assist in the tuning process.