In this report, I present my personal method to design a more accurate version of PSO, assuming we know what kind of problems we will have to solve. To illustrate the method, I consider a class of problems that are moderately multimodal and not of very high dimensionality (typically smaller than 30). The starting version is Standard PSO 2007 (SPSO 07). I use a modified velocity update equation is used, specific to the considered class of problems. Then, in order to improve the robustness of the algorithm, more general modifications are added (in particular, better initialisations, and the use of two kind of particles). On the whole, the resulting algorithm is indeed an improvement of SPSO 07, though the improvement is not always very impressive. In passing, I give a precise definition of "better than", and explain why the classical "mean best" performance criterion may easily be completely meaningless.
[1]
E. Sandgren,et al.
Nonlinear Integer and Discrete Programming in Mechanical Design Optimization
,
1990
.
[2]
E. Sandgren,et al.
Nonlinear Integer and Discrete Programming for Topological Decision Making in Engineering Design
,
1990
.
[3]
Tien-Tsin Wong,et al.
Sampling with Hammersley and Halton Points
,
1997,
J. Graphics, GPU, & Game Tools.
[4]
Louis Gacôgne.
Steady state evolutionary algorithm with an operator family
,
2002
.
[5]
Godfrey C. Onwubolu,et al.
New optimization techniques in engineering
,
2004,
Studies in Fuzziness and Soft Computing.
[6]
M. Clerc,et al.
Particle Swarm Optimization
,
2006
.