Unexplained Fluctuations in Particle Swarm Optimisation Performance with Increasing Problem Dimensionality

We study the behaviour of particle swarm optimisation (PSO) with increasing problem dimension for the Alpine 1 function as an exploratory and preliminary case study. Performance trends are analysed and the tuned population size for PSO across dimensions is considered. While performance generally decreases monotonically with scale, there is an unexpected improvement in performance part way along the trend. This also appears to coincide with a counter-intuitive transition from large to small populations being preferred, and underlines the challenge, and importance of, selecting the right algorithm and configuration for the problem at each increase in dimensionality