Do not be afraid of local minima: affine shaker and particle swarm

Stochastic local search techniques are powerful and flexible methods to optimize difficult functions. While each method is characterized by search trajectories produced through a randomized selection of the next step, a notable difference is caused by the interaction of different searchers, as exemplified by the Particle Swarm methods. In this paper we evaluate two extreme approaches, Particle Swarm Optimization, with interaction between the individual "cognitive" component and the "social" knowledge, and Repeated Affine Shaker, without any interaction between searchers but with an aggressive capability of scouting out local minima. The results, unexpected to the authors, show that Affine Shaker provides remarkably efficient and effective results when compared with PSO, while the advantage of Particle Swarm is visible only for functions with a very regular structure of the local minima leading to the global optimum and only for specific experimental conditions.

[1]  Peter J. Angeline,et al.  Evolutionary Optimization Versus Particle Swarm Optimization: Philosophy and Performance Differences , 1998, Evolutionary Programming.

[2]  P. J. Angeline,et al.  Using selection to improve particle swarm optimization , 1998, 1998 IEEE International Conference on Evolutionary Computation Proceedings. IEEE World Congress on Computational Intelligence (Cat. No.98TH8360).

[3]  David B. Fogel,et al.  A Note on the Empirical Evaluation of Intermediate Recombination , 1995, Evolutionary Computation.

[4]  Russell C. Eberhart,et al.  Parameter Selection in Particle Swarm Optimization , 1998, Evolutionary Programming.

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

[6]  Yue Shi,et al.  A modified particle swarm optimizer , 1998, 1998 IEEE International Conference on Evolutionary Computation Proceedings. IEEE World Congress on Computational Intelligence (Cat. No.98TH8360).

[7]  Roberto Battiti,et al.  The Reactive Tabu Search , 1994, INFORMS J. Comput..

[8]  R. Battiti,et al.  Simulated annealing and Tabu search in the long run: A comparison on QAP tasks☆ , 1994 .

[9]  Roberto Battiti,et al.  Learning with first, second, and no derivatives: A case study in high energy physics , 1994, Neurocomputing.

[10]  Saman K. Halgamuge,et al.  Self-organizing hierarchical particle swarm optimizer with time-varying acceleration coefficients , 2004, IEEE Transactions on Evolutionary Computation.

[11]  R. Eberhart,et al.  Comparing inertia weights and constriction factors in particle swarm optimization , 2000, Proceedings of the 2000 Congress on Evolutionary Computation. CEC00 (Cat. No.00TH8512).

[12]  Riccardo Poli,et al.  Particle swarm optimization , 1995, Swarm Intelligence.

[13]  Maurice Clerc,et al.  The particle swarm - explosion, stability, and convergence in a multidimensional complex space , 2002, IEEE Trans. Evol. Comput..

[14]  R. Eberhart,et al.  Empirical study of particle swarm optimization , 1999, Proceedings of the 1999 Congress on Evolutionary Computation-CEC99 (Cat. No. 99TH8406).