Analog implementation of optimization algorithms: a distributed optimization view

Digital implementations of optimization algorithms for real time applications are often not fast enough for satisfying performance requirements. Analog computing has the potential to address this challenge, being able to provide an answer in milli- to nano-seconds. In this paper we show how we can leverage the distributed optimization framework to generate analog implementations of optimization algorithms for arbitrary functions. We present mappings of distributed algorithms to both electrical and mechanical domain, and discuss practical aspects of the physical realization using electronic circuits.