The Design and Implementation of a GPU-enabled Multi-objective Tabu-search Intended for Real World and High-dimensional Applications

Abstract Metaheuristics is a class of approximate methods based on heuristics that can effectively handle real world (usually NP-hard) problems of high-dimensionality with multiple objectives. An existing multi-objective Tabu-Search (MOTS2) has been re-designed by and ported onto Compute Unified Device Architecture (CUDA) so as to effectively deal with a scalable multi-objective problem with a range of decision variables. The high computational cost due to the problem complexity is addressed by employing Graphics Processing Units (GPUs), which alleviate the computational intensity. The main challenges of the re-implementation are the effective communication with the GPU and the transparent integration with the optimization procedures. Finally, future work is proposed towards heterogeneous applications, where improved features are accelerated by the GPUs.

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

[2]  Qingfu Zhang,et al.  Multiobjective optimization Test Instances for the CEC 2009 Special Session and Competition , 2009 .

[3]  El-Ghazali Talbi,et al.  Metaheuristics - From Design to Implementation , 2009 .

[4]  El-Ghazali Talbi,et al.  ParadisEO-MO-GPU: a framework for parallel GPU-based local search metaheuristics , 2013, GECCO '13.

[5]  Jack J. Dongarra,et al.  Towards dense linear algebra for hybrid GPU accelerated manycore systems , 2009, Parallel Comput..

[6]  El-Ghazali Talbi,et al.  GPU-Based Multi-start Local Search Algorithms , 2011, LION.

[7]  Shinn-Ying Ho,et al.  Intelligent evolutionary algorithms for large parameter optimization problems , 2004, IEEE Trans. Evol. Comput..

[8]  Christian Schulz,et al.  Efficient local search on the GPU - Investigations on the vehicle routing problem , 2013, J. Parallel Distributed Comput..

[9]  El-Ghazali Talbi,et al.  GPU Computing for Parallel Local Search Metaheuristic Algorithms , 2013, IEEE Transactions on Computers.

[10]  Martin Lilleeng Sætra,et al.  Graphics processing unit (GPU) programming strategies and trends in GPU computing , 2013, J. Parallel Distributed Comput..

[11]  Qingfu Zhang,et al.  MOEA/D: A Multiobjective Evolutionary Algorithm Based on Decomposition , 2007, IEEE Transactions on Evolutionary Computation.

[12]  Jonathan E. Fieldsend,et al.  On the effect of selection and archiving operators in many-objective particle swarm optimisation , 2013, GECCO '13.

[13]  Marco Laumanns,et al.  Scalable multi-objective optimization test problems , 2002, Proceedings of the 2002 Congress on Evolutionary Computation. CEC'02 (Cat. No.02TH8600).

[14]  R. Lyndon While,et al.  A Scalable Multi-objective Test Problem Toolkit , 2005, EMO.

[15]  Francisco Luna,et al.  On the scalability of multi-objective metaheuristics for the software scheduling problem , 2011, 2011 11th International Conference on Intelligent Systems Design and Applications.

[16]  R. Lyndon While,et al.  A review of multiobjective test problems and a scalable test problem toolkit , 2006, IEEE Transactions on Evolutionary Computation.

[17]  Timoleon Kipouros,et al.  Multi-Objective Optimization of a Fluid StructureInteraction Benchmarking , 2013 .

[18]  Carlos A. Coello Coello,et al.  A Study of Multiobjective Metaheuristics When Solving Parameter Scalable Problems , 2010, IEEE Transactions on Evolutionary Computation.

[19]  Michel Gendreau,et al.  Handbook of Metaheuristics , 2010 .

[20]  Timoleon Kipouros,et al.  Biobjective Optimisation of Preliminary Aircraft Trajectories , 2013, EMO.

[21]  Stefano Cagnoni,et al.  libCudaOptimize: an open source library of GPU-based metaheuristics , 2012, GECCO '12.

[22]  P. John Clarkson,et al.  The development of a multi-objective Tabu Search algorithm for continuous optimisation problems , 2008, Eur. J. Oper. Res..