This paper presents a model-free reinforcement learning–based control strategy for a three-stage multi-cellular DC–DC power converter (chopper). A Deep Q-Network (DQN) agent is trained to regulate the flying capacitor voltages and load current directly from interaction with the system, without requiring explicit knowledge of its mathematical model. The framework incorporates tailored state preprocessing and reward shaping to enhance learning stability and performance. The proposed approach is benchmarked against a conventional model-based Sliding Mode Controller (SMC) under step reference scenario. Simulation results show that the DQN agent achieved comparable dynamic behavior and superior steady-state accuracy while generalizing effectively across operating conditions. Furthermore, the learned control policy produced smoother switching actions, resulting in reduced oscillations and lower switching stress on power devices. These findings highlight the potential of model-free reinforcement learning as a robust and adaptive alternative to classical control methods for power electronic converters.

\begin{abstract} This study aims to address the effect of date kernel powder (DKP) as reinforcement obtained from Touggourt oasis, Algeria, on the elastic properties of biocomposites based on two prepared DKPs with grain size 300 $\mu$m and 500 $\mu$m mixed on epoxy (ER) matrix. The weight percentage of powders with 5\%, 10\%, and 15\% was used to obtain epoxy matrix (ER)/date kernel powder (DKP) biocomposites. The effects of DKP size on the elastic properties of ER/DKP biocomposites, such as ultrasonic wave velocities (longitudinal and shear), longitudinal modulus, shear modulus, bulk modulus, Young’s modulus of elasticity, ultrasonic microhardness, Poisson’s ratio, and acoustic impedance, were determined using the ultrasonic through-transmission method. In addition, the two biocomposites prepared to analyze the chemical changes in the functional groups and their morphology were studied using X-ray diffraction and optical microscopy, respectively. The results of ultrasonic characterization of the ER/DKP biocomposites showed that there is a significant relationship between the sizes of DKP and elastic constant values. In addition, the experimental results illustrated that the optimum weight percent of DKP reinforcement in neat ER for excellent mechanical behavior of ER/DKP biocomposites is 5\% and 10\% for 300 $\mu$m and 500 $\mu$m, respectively. \noindent\textbf{Keywords:} Date kernel powder (DKP), Bio-sourced materials, Ecofriendly materials, Ultrasonic characterization, Elastic constants \end{abstract}