The sensitivity of single objective optimization algorithm control parameter values under different computational constraints

When solving a single objective optimization problem, a user desires an accurate solution, but may be computationally constrained in terms of the number of objective function evaluations (OFEs) that can be afforded. The OFE budget is application specific, varying depending on the time, computing resources, and the nature of the optimization problem. Control parameter value sensitivity to this OFE budget constraint is investigated for the particle swarm- and differential evolution optimization algorithms. The algorithms are tuned to selected testing problems under different OFE budget constraints, and then their performance is assessed at different OFE budgets from what they were tuned for. The results give evidence that combinations of optimization algorithm control parameter values which perform well for high OFE budgets do not perform well for low OFE budgets and vice versa. This indicates that when selecting control parameter values for these two algorithms, not only should the optimization problem characteristics be taken into account, but also the computational constraints.

[1]  Andries Petrus Engelbrecht,et al.  Fundamentals of Computational Swarm Intelligence , 2005 .

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

[3]  A. E. Eiben,et al.  Comparing parameter tuning methods for evolutionary algorithms , 2009, 2009 IEEE Congress on Evolutionary Computation.

[4]  Qingfu Zhang,et al.  Differential Evolution With Composite Trial Vector Generation Strategies and Control Parameters , 2011, IEEE Transactions on Evolutionary Computation.

[5]  Saku Kukkonen,et al.  Real-parameter optimization with differential evolution , 2005, 2005 IEEE Congress on Evolutionary Computation.

[6]  Russell C. Eberhart,et al.  A new optimizer using particle swarm theory , 1995, MHS'95. Proceedings of the Sixth International Symposium on Micro Machine and Human Science.

[7]  Andries Petrus Engelbrecht,et al.  A Cooperative approach to particle swarm optimization , 2004, IEEE Transactions on Evolutionary Computation.

[8]  Rainer Storn,et al.  Differential Evolution – A Simple and Efficient Heuristic for global Optimization over Continuous Spaces , 1997, J. Glob. Optim..

[9]  John J. Grefenstette,et al.  Optimization of Control Parameters for Genetic Algorithms , 1986, IEEE Transactions on Systems, Man, and Cybernetics.

[10]  M. E. H. Pedersen,et al.  Tuning & simplifying heuristical optimization , 2010 .

[11]  David H. Wolpert,et al.  No free lunch theorems for optimization , 1997, IEEE Trans. Evol. Comput..

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

[13]  Schalk Kok,et al.  Locating and Characterizing the Stationary Points of the Extended Rosenbrock Function , 2009, Evolutionary Computation.