A Survey on Smart Optimisation Techniques for 6G-oriented Integrated Circuits Design.

With the rapid development of next-generation wireless communications, there is a growing demand for high-quality integrated circuits (ICs), particularly analog ICs, which play a pivotal role for the full roll-out sixth-generation (6G) technology. So far, the IC design has been performed through manual approaches which sometimes results in time-consuming turnaround, especially the sizing phase of analog IC design. In order to make the IC design process much faster, recently automated methods for optimizing IC design has gained a lot of attention. From this perspective, this paper aims at providing a survey of the most recent works on optimization strategies for analog IC sizing, as well as a related categorization into two main categories: analytical methods and simulation-based methods. A further sub-classification within the realm of simulation-based methods is also provided by dividing the core mathematical principles into three major sub-methods: Bayesian-based, metaheuristic-based, and reinforcement-learning-based techniques. In addition, with the main aim of providing insights on the utilization of optimization algorithms for the IC sizing process, we present a case study involving the utilization of various metaheuristic algorithms in the design of a bandgap reference circuit - an essential analog IC component. The paper is concluded by highlighting potential future research directions in the field of analog IC design optimization and automation, which include exploration of multi-agent reinforcement learning, integration of quantum computing, and further development of full-flow automated tools for analog IC design. [ABSTRACT FROM AUTHOR]