Effect of the Pseudorandom Number Generators on the Standard Particle Swarm Optimization on a GPU

A key component of particle swarm optimization algorithms is pseudorandom number generators (PRNGs) which provide random numbers to drive the stochastic search process. In this paper, we implemented ten PRNGs on CPUs and graphics processing units (GPUs). We present the effect of PRNGs on a parallel implementation of the standard particle swarm optimization on a GPU. The performance of SPSO algorithms is influenced by the quality of the PRNGs running on a GPU. By using the combined Tausworthe PRNG, the proposed parallel implementation of SPSO provides up to 307 times speedup compared to a serial CPU SPSO implementation. Speedup is greatly accelerated for high dimension, large particles and complex benchmark functions. Here, the experiments were conducted on well-known six benchmark functions. Consequently, this implementation can be widely used in real optimizing problems.

[1]  Jaspreet Kaur,et al.  Parallel Implementation of PSO Algorithm Using GPGPU , 2016, 2016 Second International Conference on Computational Intelligence & Communication Technology (CICT).

[2]  Donald E. Knuth,et al.  The art of computer programming. Vol.2: Seminumerical algorithms , 1981 .

[3]  Asadollah Shahbahrami,et al.  High performance implementation of APSO algorithm using GPU platform , 2015, 2015 The International Symposium on Artificial Intelligence and Signal Processing (AISP).

[4]  Zhongwen Luo,et al.  Artificial neural network computation on graphic process unit , 2005, Proceedings. 2005 IEEE International Joint Conference on Neural Networks, 2005..

[5]  James Kennedy,et al.  Defining a Standard for Particle Swarm Optimization , 2007, 2007 IEEE Swarm Intelligence Symposium.

[6]  Hiroshi Hattori,et al.  A CUDA Implementation of the Standard Particle Swarm Optimization , 2016, 2016 18th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC).

[7]  Riccardo Poli,et al.  Particle swarm optimization , 1995, Swarm Intelligence.

[8]  Pierre L'Ecuyer,et al.  Tables of maximally equidistributed combined LFSR generators , 1999, Math. Comput..

[9]  Lee W. Howes.  Efficient Random Number Generation and Application Using , 2007 .

[10]  Dinesh Manocha,et al.  Fast computation of generalized Voronoi diagrams using graphics hardware , 1999, SIGGRAPH.