Reinforcement learning is one of the paradigms and methodologies of machine learning developed in the computational intelligence community. It is used to describe and solve the problem in which agents maximize returns or achieve specific goals through learning strategies in the process of interaction with complex environments. The goal of reinforcement learning is to get the best solution to the current problem through reward and punishment: by rewarding good strategy and punishing bad strategy, continuously strengthening the process, to finally arrive at the best solution. To some extent, it has close connections to both adaptive control and optimisation.

More general scenarios for reinforcement learning and adaptive optimisation present a major challenge in complex dynamic systems. The process of controlling complex dynamic systems and industrial plants, or parts of such, involves a variety of challenging aspects that reinforcement learning algorithms need to tackle. Dealing with the complexity of such industrial process can involve computer communication, complex networks, continuous state and action spaces, high-dimensional dynamics, partially observable state spaces, randomness induced by the heteroscedastic sensor noise and latent variables, delayed characteristics, and nonstationary in the optimal steering, i.e. the optimal policy will not approach a fixed operation point.

The aim of this Special Issue is to bring together work on reinforcement learning and adaptive optimisation of complex dynamic systems and industrial applications. We invite authors to contribute original research articles as well as review articles related to all aspects of reinforcement learning algorithms, complex dynamic modelling, optimisation theory, optimal control methods, signal processing, and practical applications. Of particular interest are papers devoted to the development of complex industrial applications. Papers presenting computational issues, search strategies, and modelling and solution techniques to practical industrial problems are also welcome.

Potential topics include but are not limited to the following:

• Reinforcement learning algorithms in complex dynamics
• Iterative learning and adaptive optimisation of complex systems
• Decision optimisation in complex processes
• Unmanned system control and computer communication
• Multi-agent reinforcement learning and control
• Neural network system and adaptive optimisation
• Fuzzy dynamic systems and adaptive optimisation
• Data driven modelling, control and optimisation
• Signal processing and optimisation
• Complex process control and optimisation
• Complex industrial process and applications