In this paper, we propose an alternative strategy of adapting the inertia weight parameter during the course of particle swarm optimization. Three oscillating inertia weight functions of time are presented and their influence on the performance and success of the algorithm is investigated by optimizing six standard benchmark functions. Instances of the algorithm that make use of inertia weight adaptation schemes from the literature are also tested for comparison purposes. The results show that the proposed functions are competitive and, in some cases, outperform the established inertia weight functions, in terms of consistency and speed of convergence.
|Publication status||Published - 2009|
|Event||11th Annual Conference on Genetic and Evolutionary Computation - Montreal, Canada|
Duration: 8 Jul 2009 → 12 Jul 2009
|Conference||11th Annual Conference on Genetic and Evolutionary Computation|
|Abbreviated title||GECCO '09|
|Period||8/07/09 → 12/07/09|