Reinforcement Learning (RL) seeks to develop systems capable of autonomous decision-making by learning through interaction with their environment. Central to this process are reward engineering and reward shaping, which are essential for enhancing the efficiency and effectiveness of RL algorithms. These techniques guide agents toward desired behaviors, improve learning stability, and accelerate convergence by addressing challenges such as sparse and delayed rewards. However, the complexity of real-world environments and the computational demands of RL algorithms remain significant obstacles to broader adoption. Recent advancements in deep learning have enabled RL to handle high-dimensional state and action spaces, facilitating applications in robotics, autonomous driving, and complex decision-making tasks. In response to these developments, this paper provides one of the first comprehensive reviews of reward design in RL, with a focus on the methodologies and techniques underpinning reward engineering and shaping. By introducing a detailed taxonomy, critically analyzing current approaches, and highlighting their limitations, this work fills an important gap in the literature, offering insights into how reward structures can be optimized to meet the growing demands of modern AI systems.

The optimization of vanadium redox flow batteries (VRFBs) is closely related to the flow rate control: a proper regulation of the electrolyte flow rate reduces losses and prolongs battery lifetime. To this end, a flow factor control strategy in VRFBs was proposed in the literature provided with numerical/ experimental validations. Yet, a theoretical justification of this approach was lacking. The respective control law is a generalization of Faraday’s law of electrolysis since it employs a special scaling parameter referred to as the flow factor . In this paper, we show that this coefficient is directly related to the conversion rate of electrolyte in the cell. Furthermore, we pose an optimal control problem with maximization of total battery power integrated over time to determine an optimal flow factor. To this end, we use a simple stochastic policy gradient algorithm. The case studies illustrate the application of the computed optimal controller under various load currents and demonstrate that there is no single flow factor for all modes, the optimal performance of the battery can only be guaranteed by different values of this parameter. As a result, the proposed control strategy can be used for advanced control and monitoring tools for industrial VRFB systems.