Sonification of population behavior in particle swarm optimization
Tiago Fernandes Tavares, Alan Godoy · 2013
High-dimensional optimization problems may be addressed using populational meta-heuristics, whose statistical properties may indicate important characteristics of the optimization process. There is a great number of these properties, which means their joint visualization may become impractical. We developed a method for sonically displaying characteristics of population dynamics in particle swarm optimization processes as soundscape parameters. This process allows jointly analyzing several dimensions of the population's dynamics. Moreover, design decisions aimed at generating aesthetically appealing soundscapes, which allows the proposed system to be used as an automated music composition environment.