Comparison of simulated annealing parallelization methods for quadratic assignment problems

Karesel atama problemi (KAP), NP-hard sınıfındaki en zor kombinatoryal optimizasyon problemlerinden birisidir. Problemin zorluğundan dolayı birçok araştırmacı bu tip atama problemini çalışılmaktadır. Bu çalışmada tavlama benzetimi yöntemi MATLAB platformunda paralelleştirilerek iyi bilinen bir KAP Kütüphanesi olan QAPLIB’den alınan 36 örnek problemi çözmek için kullanılmıştır. Değişik paralelleştirme yöntemlerinin performansları kullanılan problemler için karşılaştırılmıştır. Sonuç olarak seri tavlama benzetimi yöntemiyle karşılaştırıldığında, paralel yöntemlerin uygun parametreler kullanıldığında daha hızlı sonuç verdiği görülmüştür. Quadratic assignment problem (QAP) is one of the most difficult combinatorial optimization problems in the NP-hard class. Due to the difficulty of the problem, many researchers have been studying this type of assignment problem. In this work, simulated annealing method is parallelized on MATLAB platform and is used to solve 36 problems from QAPLIB which is a well-known QAP library. The performance of different parallelization methods is compared for the problems used. As a result, when compared with the serial simulated annealing method, it is seen that the parallel methods give faster results when the appropriate parameters are used.

[1]  Singiresu S. Rao.  Engineering Optimization : Theory and Practice , 2010 .

[2]  Jadranka Skorin-Kapov,et al.  Tabu Search Applied to the Quadratic Assignment Problem , 1990, INFORMS J. Comput..

[3]  Per S. Laursen.  Parallel heuristic search - Introductions and a new approach , 1996, Solving Combinatorial Optimization Problems in Parallel.

[4]  Hao Chen,et al.  Parallel Genetic Simulated Annealing: A Massively Parallel SIMD Algorithm , 1998, IEEE Trans. Parallel Distributed Syst..

[5]  Mohamad Amin Kaviani,et al.  A hybrid Tabu search-simulated annealing method to solve quadratic assignment problem , 2014 .

[6]  T. Koopmans,et al.  Assignment Problems and the Location of Economic Activities , 1957 .

[7]  Gerald Paul,et al.  Comparative performance of tabu search and simulated annealing heuristics for the quadratic assignment problem , 2010, Oper. Res. Lett..

[8]  Morteza Karami,et al.  Analysis and performance measurement of existing solution methods of quadratic assignment problem , 2013 .

[9]  Thomas Stützle,et al.  Tabu search vs. simulated annealing as a function of the size of quadratic assignment problem instances , 2014, Comput. Oper. Res..

[10]  Panos M. Pardalos,et al.  Generating quadratic assignment test problems with known optimal permutations , 1992, Comput. Optim. Appl..

[11]  Jiunn-Chin Wang,et al.  A Multistart Simulated Annealing Algorithm for the Quadratic Assignment Problem , 2012, 2012 Third International Conference on Innovations in Bio-Inspired Computing and Applications.

[12]  Thomas E. Vollmann,et al.  An Experimental Comparison of Techniques for the Assignment of Facilities to Locations , 1968, Oper. Res..

[13]  Franz Rendl,et al.  QAPLIB – A Quadratic Assignment Problem Library , 1997, J. Glob. Optim..

[14]  Esin Onbasçioglu,et al.  Parallel Simulated Annealing Algorithms in Global Optimization , 2001, J. Glob. Optim..

[15]  José A. García-Rodríguez,et al.  An efficient implementation of parallel simulated annealing algorithm in GPUs , 2012, Journal of Global Optimization.

[16]  Nair Maria Maia de Abreu,et al.  A survey for the quadratic assignment problem , 2007, Eur. J. Oper. Res..

[17]  Panos M. Pardalos,et al.  Parallelized Heuristics for Combinatorial Search , 1997 .

[18]  S. M. H. Seyedkashi,et al.  New Simulated Annealing Algorithm for Quadratic Assignment Problem , 2010 .

[19]  M. Kováč.  Solving Quadratic Assignment Problem in parallel using local search with simulated annealing elements , 2013, The International Conference on Digital Technologies 2013.

[20]  Gerald Paul,et al.  A GPU implementation of the Simulated Annealing Heuristic for the Quadratic Assignment Problem , 2012, ArXiv.