Sungjoo Ha
ABSTRACT
We tackle the problem of knowledge discovery in time series data using genetic programming and GPGPUs. Using genetic programming, various precursor patterns that have certain attractive qualities are evolved to predict the events of interest. Unfortunately, evolving a set of diverse patterns typically takes huge execution time, sometimes longer than one month for this case. In this paper, we address this problem by proposing a parallel GP framework using GPGPUs, particularly in the context of big financial data. By maximally exploiting the structure of the nVidia GPGPU platform on stock market time series data, we were able see more than 250-fold reduction in the running time.
- E. Cantú-Paz. A survey of parallel genetic algorithms. Calculateurs Paralleles, Reseaux et Systems Repartis, 10, 1998.Google Scholar
- P. Collet, E. Lutton, M. Schoenauer, and J. Louchet. Take it EASEA. Parallel Problem Solving from Nature - PPSN VI, volume 1917, pages 891--901, 2000. Google ScholarDigital Library
- M. J. Gibson, E. Keedwell, and D. A. Savic. Understanding the efficient parallelisation of cellular automata on CPU and GPGPU hardware. Genetic and Evolutionary Computation Conference, pages 171--172, 2013. Google ScholarDigital Library
- M. Harris. Optimizing parallel reduction in CUDA. Technical report, nVidia, 2008.Google Scholar
- J. I. Hidalgo, J. M. Colmenar, J. L. Risco-Martín, C. Sánchez-Lacruz, J. Lanchares, O. Garnica, and J. Díaz. Solving GA-hard problems with EMMRS and GPGPUs. Genetic and Evolutionary Computation Conference, pages 1007--1014, 2014. Google ScholarDigital Library
- R. Hrbacek and L. Sekanina. Towards highly optimized cartesian genetic programming: From sequential via SIMD and thread to massive parallel implementation. Genetic and Evolutionary Computation Conference, pages 1015--1022, 2014. Google ScholarDigital Library
- J. Jaros and R. Tyrala. GPU-accelerated evolutionary design of the complete exchange communication on wormhole networks. Genetic and Evolutionary Computation Conference, pages 1023--1030, 2014. Google ScholarDigital Library
- J. R. Koza. Genetic Programming: On the Programming of Computers by Means of Natural Selection. MIT Press, 1992. Google ScholarDigital Library
- P. Krömer, V. Snásel, J. Platos, and A. Abraham. Many-threaded implementation of differential evolution for the CUDA platform. Genetic and Evolutionary Computation Conference, pages 1595--1602, 2011. Google ScholarDigital Library
- Y.-K. Kwon and B.-R. Moon. Personalized email marketing with a genetic programming circuit model. Genetic and Evolutionary Computation Conference, pages 1352--1358, 2001.Google Scholar
- W. B. Langdon. Graphics processing units and genetic programming: An overview. Soft Computing, 15 (8): 1657--1669, 2011. Google ScholarDigital Library
- 001)Last, Klein, Kandel, and K}Last01knowledgediscoveryM. Last, Y. Klein, A. Kandel, and A. K. Knowledge discovery in time series databases. IEEE Transactions on Systems, Man, and Cybernetics, 31: 160--169, 2001. Google ScholarDigital Library
- S.-K. Lee and B. R. Moon. A new modular genetic programming for finding attractive technical patterns in stock markets. Genetic and Evolutionary Computation Conference, pages 1219--1226, 2010. Google ScholarDigital Library
- P. Lipinski. ECGA vs. BOA in discovering stock market trading experts. Genetic and Evolutionary Computation Conference, pages 531--538, 2007. Google ScholarDigital Library
- T. V. Luong, N. Melab, and E.-G. Talbi. Parallel local search on GPU. Research Report RR-6915, 2009.Google Scholar
- T. V. Luong, N. Melab, and E.-G. Talbi. GPU-based island model for evolutionary algorithms. Genetic and Evolutionary Computation Conference, pages 1089--1096, 2010. Google ScholarDigital Library
- O. Maitre, L. A. Baumes, N. Lachiche, A. Corma, and P. Collet. Coarse grain parallelization of evolutionary algorithms on GPGPU cards with EASEA. Genetic and Evolutionary Computation Conference, pages 1403--1410, 2009. Google ScholarDigital Library
- L. Mussi, Y. S. Nashed, and S. Cagnoni. GPU-based asynchronous particle swarm optimization. Genetic and Evolutionary Computation Conference, pages 1555--1562, 12--16 July 2011. Google ScholarDigital Library
- P. Pospichal, J. Jaros, and J. Schwarz. Parallel genetic algorithm on the CUDA architecture. Applications of Evolutionary Computation, pages 442--451, 2010. Google ScholarDigital Library
- J.-Y. Potvin, P. Soriano, and M. Vallée. Generating trading rules on the stock markets with genetic programming. Computers and Operations Research, 31 (7): 1033--1047, June 2004. Google ScholarDigital Library
- R. J. Povinelli. Using genetic algorithms to find temporal patterns indicative of time series events. GECCO 2000 Workshop: Data Mining with Evolutionary Algorithms, pages 80--84, 2000.Google Scholar
- P. Przymus and K. Kaczmarski. Time series queries processing with GPU support. 17th East European Conference on Advances in Databases and Information Systems, pages 53--60, 2013.Google Scholar
- A. K. Qin, F. Raimondo, F. Forbes, and Y.-S. Ong. An improved CUDA-based implementation of differential evolution on GPU. Genetic and Evolutionary Computation Conference, pages 991--998, 2012. Google ScholarDigital Library
- K. Rocki and R. Suda. Accelerating 2-opt and 3-opt local search using GPU in the travelling salesman problem. International Conference on High Performance Computing and Simulation, pages 489--495, 2012.Google Scholar
- S. Shao, X. Liu, M. Zhou, J. Zhan, X. Liu, Y. Chu, and H. Chen. A GPU-based implementation of an enhanced GEP algorithm. T. Soule and J. H. Moore, editors, Genetic and Evolutionary Computation Conference, pages 999--1006, 2012. Google ScholarDigital Library
- S. Solomon, P. Thulasiraman, and R. K. Thulasiram. Collaborative multi-swarm PSO for task matching using graphics processing units. Genetic and Evolutionary Computation Conference, pages 1563--1570, 2011. Google ScholarDigital Library
- K. C. Tan, Q. Yu, and T. H. Lee. A distributed evolutionary classifier for knowledge discovery in data mining. IEEE Transactions on Systems, Man, and Cybernetics, Part C, 35 (2): 131--142, 2005. Google ScholarDigital Library
- P. Vidal and E. Alba. A multi-GPU implementation of a cellular genetic algorithm. IEEE Congress on Evolutionary Computation, pages 1--7, 2010.Google ScholarCross Ref
- V. Volkov. Better performance at lower occupancy. GPU Technology Conference, 2010.Google Scholar
- S. Zhang and Z. He. Implementation of parallel genetic algorithm based on CUDA. Advances in Computation and Intelligence, volume 5821 of Lecture Notes in Computer Science, pages 24--30, 2009. Google ScholarDigital Library
Index Terms
- Fast Knowledge Discovery in Time Series with GPGPU on Genetic Programming
Recommendations
Fast parallel genetic programming: multi-core CPU versus many-core GPU
Genetic Programming (GP) is a computationally intensive technique which is also highly parallel in nature. In recent years, significant performance improvements have been achieved over a standard GP CPU-based approach by harnessing the parallel ...
GPGPU: general-purpose computation on graphics hardware
SC '06: Proceedings of the 2006 ACM/IEEE conference on SupercomputingThe graphics processor (GPU) on today's commodity video cards has evolved into an extremely powerful and flexible processor. Modern graphics architectures provide tremendous memory bandwidth and computational horsepower, with dozens of fully ...
Efficient Abstractions for GPGPU Programming
General purpose (GP)GPU programming demands to couple highly parallel computing units with classic CPUs to obtain a high performance. Heterogenous systems lead to complex designs combining multiple paradigms and programming languages to manage each ...
Comments