c AS : The Cunning Ant System Solving the traveling salesman problem and the quadratic assignment problem using the cunning ant system

This paper describes the cunning ant system (cAS), which can be used to optimize problems with candidate solutions represented by permutations. cAS generates new candidate solutions from a combination of the pheromone density and the archive of past candidate solutions. Performance of cAS is demonstrated on two important classes of permutation problems: (1) the traveling salesman problem and (2) the quadratic assignment problem. In both problem classes, cAS is shown to perform very well and outperform other techniques for solving these problems. The paper also discusses how the efficiency of cAS can be further improved by combining cAS with local search and by parallelizing cAS using a master-slave architecture. Finally, the paper presents a pheromone-density entropy measure, which can be used to measure the potential of cAS and other ACO algorithms to generate novel, fitter solutions.

[1]  A. Wagner.  Robustness and Evolvability in Living Systems , 2005 .

[2]  Richard F. Hartl,et al.  An improved Ant System algorithm for theVehicle Routing Problem , 1999, Ann. Oper. Res..

[3]  Shigeyoshi Tsutsui,et al.  Probabilistic Model-Building Genetic Algorithms in Permutation Representation Domain Using Edge Histogram , 2002, PPSN.

[4]  Manuel López-Ibáñez,et al.  Ant colony optimization , 2010, GECCO '10.

[5]  Seid H. Pourtakdoust,et al.  An Extension of Ant Colony System to Continuous Optimization Problems , 2004, ANTS Workshop.

[6]  Corso Elvezia.  Ant Colonies for the QAP , 1997 .

[7]  Marco Dorigo,et al.  Ant system: optimization by a colony of cooperating agents , 1996, IEEE Trans. Syst. Man Cybern. Part B.

[8]  Shigeyoshi Tsutsui.  Convergence Analysis of the Cunning Ant System with Entropy Measure , 2008, PDPTA.

[9]  Luca Maria Gambardella,et al.  Solving symmetric and asymmetric TSPs by ant colonies , 1996, Proceedings of IEEE International Conference on Evolutionary Computation.

[10]  Shigeyoshi Tsutsui.  An Enhanced Aggregation Pheromone System for Real-Parameter Optimization in the ACO Metaphor , 2006, ANTS Workshop.

[11]  Adnan Acan.  An External Memory Implementation in Ant Colony Optimization , 2004, ANTS Workshop.

[12]  Thomas Stützle,et al.  Parallel Ant Colony Optimization for the Traveling Salesman Problem , 2006, ANTS Workshop.

[13]  Luca Maria Gambardella,et al.  Ant colony system: a cooperative learning approach to the traveling salesman problem , 1997, IEEE Trans. Evol. Comput..

[14]  ATSPDavid S. JohnsonAT.  Experimental Analysis of Heuristics for the Stsp , 2001 .

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

[16]  Brian W. Kernighan,et al.  An Effective Heuristic Algorithm for the Traveling-Salesman Problem , 1973, Oper. Res..

[17]  Andrew Lewis,et al.  A Parallel Implementation of Ant Colony Optimization , 2002, J. Parallel Distributed Comput..

[18]  Vittorio Maniezzo,et al.  The Ant System Applied to the Quadratic Assignment Problem , 1999, IEEE Trans. Knowl. Data Eng..

[19]  Charles Fleurent,et al.  Genetic Hybrids for the Quadratic Assignment Problem , 1993, Quadratic Assignment and Related Problems.

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

[21]  Marco Dorigo,et al.  Ant colony optimization for continuous domains , 2008, Eur. J. Oper. Res..

[22]  David E. Goldberg,et al.  The Design of Innovation: Lessons from and for Competent Genetic Algorithms , 2002 .

[23]  Adnan Acan.  An External Partial Permutations Memory for Ant Colony Optimization , 2005, EvoCOP.

[24]  Thomas Stützle,et al.  MAX-MIN Ant System , 2000, Future Gener. Comput. Syst..

[25]  Shigeyoshi Tsutsui,et al.  Cunning Ant System for Quadratic Assignment Problem with Local Search and Parallelization , 2007, PReMI.

[26]  William J. Cook,et al.  Chained Lin-Kernighan for Large Traveling Salesman Problems , 2003, INFORMS Journal on Computing.