DEMCMC-GPU: An Efficient Multi-Objective Optimization Method with GPU Acceleration on the Fermi Architecture

In this paper, we present an efficient method implemented on Graphics Processing Unit (GPU), DEMCMC-GPU, for multi-objective continuous optimization problems. The DEMCMC-GPU kernel is the DEMCMC algorithm, which combines the attractive features of Differential Evolution (DE) and Markov Chain Monte Carlo (MCMC) to evolve a population of Markov chains toward a diversified set of solutions at the Pareto optimal front in the multi-objective search space. With parallel evolution of a population of Markov chains, the DEMCMC algorithm is a natural fit for the GPU architecture. The implementation of DEMCMC-GPU on the pre-Fermi architecture can lead to a ~25 speedup on a set of multi-objective benchmark function problems, compare to the CPU-only implementation of DEMCMC. By taking advantage of new cache mechanism in the emerging NVIDIA Fermi GPU architecture, efficient sorting algorithm on GPU, and efficient parallel pseudorandom number generators, the speedup of DEMCMC-GPU can be aggressively improved to ~100.

[1]  J. D. de Pablo,et al.  Optimal allocation of replicas in parallel tempering simulations. , 2005, The Journal of chemical physics.

[2]  Daan Frenkel,et al.  Configurational bias Monte Carlo: a new sampling scheme for flexible chains , 1992 .

[3]  Marco Laumanns,et al.  Scalable Test Problems for Evolutionary Multiobjective Optimization , 2005, Evolutionary Multiobjective Optimization.

[4]  Marco Laumanns,et al.  SPEA2: Improving the strength pareto evolutionary algorithm , 2001 .

[5]  Yaohang Li,et al.  GPU-accelerated multi-scoring functions protein loop structure sampling , 2010, 2010 IEEE International Symposium on Parallel & Distributed Processing, Workshops and Phd Forum (IPDPSW).

[6]  D. Kofke,et al.  Selection of temperature intervals for parallel-tempering simulations. , 2005, The Journal of chemical physics.

[7]  DebK.,et al.  A fast and elitist multiobjective genetic algorithm , 2002 .

[8]  Lothar Thiele,et al.  Comparison of Multiobjective Evolutionary Algorithms: Empirical Results , 2000, Evolutionary Computation.

[9]  Marco Laumanns,et al.  PISA: A Platform and Programming Language Independent Interface for Search Algorithms , 2003, EMO.

[10]  Man Leung Wong,et al.  Parallel multi-objective evolutionary algorithms on graphics processing units , 2009, GECCO '09.

[11]  Patrick M. Reed,et al.  How effective and efficient are multiobjective evolutionary algorithms at hydrologic model calibration , 2005 .

[12]  Weihang Zhu,et al.  Nonlinear optimization with a massively parallel Evolution Strategy-Pattern Search algorithm on graphics hardware , 2011, Appl. Soft Comput..

[13]  Michael Garland,et al.  Designing efficient sorting algorithms for manycore GPUs , 2009, 2009 IEEE International Symposium on Parallel & Distributed Processing.

[14]  C. D. Gelatt,et al.  Optimization by Simulated Annealing , 1983, Science.

[15]  R. K. Ursem.  Multi-objective Optimization using Evolutionary Algorithms , 2009 .

[16]  Y. Okamoto,et al.  GENERALIZED-ENSEMBLE MONTE CARLO METHOD FOR SYSTEMS WITH ROUGH ENERGY LANDSCAPE , 1997, cond-mat/9710306.

[17]  Weihang Zhu,et al.  Massively parallel differential evolution—pattern search optimization with graphics hardware acceleration: an investigation on bound constrained optimization problems , 2011, J. Glob. Optim..

[18]  Yaohang Li,et al.  GPU-accelerated differential evolutionary Markov Chain Monte Carlo method for multi-objective optimization over continuous space , 2010, BADS '10.

[19]  Jun S. Liu,et al.  The Multiple-Try Method and Local Optimization in Metropolis Sampling , 2000 .

[20]  Matt Pharr,et al.  Gpu gems 2: programming techniques for high-performance graphics and general-purpose computation , 2005 .

[21]  Rainer Storn,et al.  Differential Evolution – A Simple and Efficient Heuristic for global Optimization over Continuous Spaces , 1997, J. Glob. Optim..

[22]  Jason Wittenberg,et al.  Clarify: Software for Interpreting and Presenting Statistical Results , 2003 .

[23]  Kalyanmoy Deb,et al.  A fast and elitist multiobjective genetic algorithm: NSGA-II , 2002, IEEE Trans. Evol. Comput..

[24]  Yaohang Li,et al.  Integrating multiple scoring functions to improve protein loop structure conformation space sampling , 2010, 2010 IEEE Symposium on Computational Intelligence in Bioinformatics and Computational Biology.

[25]  C. Geyer,et al.  Annealing Markov chain Monte Carlo with applications to ancestral inference , 1995 .

[26]  V. Chaudhary,et al.  Applying graphics processor units to Monte Carlo dose calculation in radiation therapy , 2010, Journal of medical physics.

[27]  G. Parisi,et al.  Simulated tempering: a new Monte Carlo scheme , 1992, hep-lat/9205018.

[28]  S. Sorooshian,et al.  Effective and efficient algorithm for multiobjective optimization of hydrologic models , 2003 .

[29]  N. Metropolis,et al.  Equation of State Calculations by Fast Computing Machines , 1953, Resonance.

[30]  Yaohang Li,et al.  Hybrid parallel tempering and simulated annealing method , 2009, Appl. Math. Comput..

[31]  Nicola Beume,et al.  On the Complexity of Computing the Hypervolume Indicator , 2009, IEEE Transactions on Evolutionary Computation.

[32]  Ying Tan,et al.  GPU-based parallel particle swarm optimization , 2009, 2009 IEEE Congress on Evolutionary Computation.

[33]  W. K. Hastings,et al.  Monte Carlo Sampling Methods Using Markov Chains and Their Applications , 1970 .