Accelerated Numerical Optimization with Explicit Consideration of Model Constraints
Population based metaheuristics can benefit from parallelization in order to address complex numerical optimization problems. Typical realistic problems usually involve non-linear functions, integer variables and many constraints, making the identification of optimal solutions mathematically challenging and computationally expensive. In this work, a parallelized version of the Particle Swarm Optimization technique is proposed, whose main contribution is the explicit consideration of constraints. The implementation is tested on a classic set of optimization problems. Speedups up to 101x were obtained using a single GPU on a standard PC using the Py-Cuda technology.
KeywordsNumerical optimization Particle swarm optimization GPU
This research was partially supported by grants from Consejo Nacional de InvestigacionesCientíficas y Técnicas (CONICET) and Universidad Tecnológica Nacional (UTN) of Argentina. The authors also gratefully acknowledge the support of NVIDIA Corporation with the donation of the TITAN X GPU used in this research.
- 8.NumPy Homepage. http://www.numpy.org/
- 9.PyCuda reference. https://developer.nvidia.com/pycuda
- 10.Liang, J.J., Runarsson, T.P., Mezura-Montes, E., Clerc, M., Suganthan, P.N., Coello, C.A.C., Deb, K.: Problem definitions and evaluation criteria for the CEC 2006 special session on constrained real-parameter optimization. Technical report (2006)Google Scholar