Monthly Archives: June 2012

Paper “SPOT Applied to Non-Stochastic Optimization Problems” presented at GECCO

 Abstract. By default, the R version of the sequential parameter optimization (SPO) software uses a Random Forest (RF) model for the surrogate optimization. The RF model is built rather fast. It runs robustly (i.e. it does not crash) and can handle non-ordered parameters very well. However, the RF model does provide poor optimization performance for a number of problems, due to the inbuilt discontinuities. It would often be more reasonable to use Kriging models. Continue reading

Review: Experimental Methods for the Analysis of Optimization Algorithms

European Journal of Operational Research (EJOR) review of the “Experimental Methods for the Analysis of Optimization Algorithms” book (http://www.springer.com/978-3-642-02537-2) by T. Bartz-Beielstein, M. Chiarandini, L. Paquete, and M. Preuss can be found on the following WWW page: http://www.sciencedirect.com/science/article/pii/S0377221711004425

ECJ Special Issue: Automated Design and Assessment of Heuristic Search Methods

The ECJ Special Issue on Automated Design and Assessment of Heuristic Search Methods contains (besides other interesting articles) the free article “Resampling Methods for Meta-Model Validation with Recommendations for Evolutionary Computation”, which discusses resampling in SPO and related methods   and Tobias’ and Simon’s article “On the Effect of Response Transformations in Sequential Parameter Optimization”.

Automatically Improving the Anytime Behaviour of Optimisation Algorithms

The paper “Automatically Improving the Anytime Behaviour of Optimisation Algorithms” by Manuel Lopez-Ibanez and Thomas Stützle can be downloaded from http://iridia.ulb.ac.be/IridiaTrSeries/link/IridiaTr2012-012.pdf
Anytime algorithms  may be interrupted at any moment and return a solution and keep steadily improving their solutions until interrupted, eventually finding the optimal. Continue reading