Adaptively Learning Memory Incorporating PSO

TL;DR

A new PSO-inspired algorithm is introduced that incorporates the positive experiences of the swarm to learn the geometry of the search space,thus obtaining the ability to consistently reach global optimum and is especially suitable for nonsmooth semiconvex functions optimization.

Abstract

Selection of perefect parameters for low-pass filters can sometimes be an expensive problem with no analytical solution or differentiability of cost function. In this paper, we introduce a new PSO-inspired algorithm, that incorporates the positive experiences of the swarm to learn the geometry of the search space,thus obtaining the ability to consistently reach global optimum and is especially suitable for nonsmooth semiconvex functions optimization. We compare it to a set of other algorithms on test functions of choice to prove it's suitability to a certain range of problems, and then apply it to the problem of finding perfect parameters for exponential smoothing algorithm.