The field of robotics and optimization is witnessing significant developments, with a growing focus on zero-order optimization techniques and their applications in trajectory optimization and policy optimization. These techniques are particularly useful for handling non-differentiable functions and escaping local minima, and are being explored in various contexts, including industrial process stabilization and robotic reconfiguration. Machine learning is also being leveraged to improve the stability and quality of industrial processes, with promising results. Noteworthy papers include:

• A mathematical tutorial on random search that provides a unifying perspective for understanding a wide range of algorithms commonly used in robotics.
• A pipeline consisting of two neural networks that successfully improves stability in terms of temperature control by about 3 times compared to ordinary solvers.
• An implementation and evaluation of different methods for reconfiguring a connected arrangement of tiles into a desired target shape using a single active robot.