Screening the Parameters Affecting Heuristic Performance Export

@inproceedings{ridge07screening, abstract = {This research screens the tuning parameters of a combinatorial optimization heuristic. Specifically, it presents a Design of Experiments (DOE) approach that uses a Fractional Factorial Design to screen the tuning parameters of Ant Colony System (ACS) for the Travelling Salesperson problem. Screening is a preliminary step to building a Response Surface Model (RSM) [20, 18]. It identifies those parameters that need not be included in a Response Surface Model, thus reducing the complexity and expense of the RSM design. 10 algorithm parameters and 2 problem characteristics are considered. Open questions on the effect of 3 parameters on performance are answered. Ant placement and choice of ant for pheromone update have no effect. However, the choice of parallel or sequential solution construction does indeed influence performance. A further parameter, sometimes assumed important, was shown to have no effect on performance. A new problem characteristic that effects performance was identified. The importance of measuring solution time was highlighted by helping identify the prohibitive cost of non-integer parameters where those parameters are exponents in the ACS algorithm's computations. All results are obtained with a publicly available algorithm and problem generator.}, author = {Ridge, Enda and Kudenko, Daniel}, booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference (GECCO '07)}, citeulike-article-id = {5781410}, citeulike-linkout-0 = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.65.1349}, keywords = {experimental-method, heuristics}, posted-at = {2009-09-14 15:07:30}, priority = {2}, title = {Screening the Parameters Affecting Heuristic Performance}, url = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.65.1349}, volume = {1}, year = {2007} }

TY - CONF ID - ridge07screening L3 - citeulike-article-id:5781410 N2 - This research screens the tuning parameters of a combinatorial optimization heuristic. Specifically, it presents a Design of Experiments (DOE) approach that uses a Fractional Factorial Design to screen the tuning parameters of Ant Colony System (ACS) for the Travelling Salesperson problem. Screening is a preliminary step to building a Response Surface Model (RSM) [20, 18]. It identifies those parameters that need not be included in a Response Surface Model, thus reducing the complexity and expense of the RSM design. 10 algorithm parameters and 2 problem characteristics are considered. Open questions on the effect of 3 parameters on performance are answered. Ant placement and choice of ant for pheromone update have no effect. However, the choice of parallel or sequential solution construction does indeed influence performance. A further parameter, sometimes assumed important, was shown to have no effect on performance. A new problem characteristic that effects performance was identified. The importance of measuring solution time was highlighted by helping identify the prohibitive cost of non-integer parameters where those parameters are exponents in the ACS algorithm’s computations. All results are obtained with a publicly available algorithm and problem generator. TI - Screening the Parameters Affecting Heuristic Performance VL - 1 T2 - Proceedings of the Genetic and Evolutionary Computation Conference (GECCO '07) KW - experimental-method KW - heuristics AU - Ridge, Enda AU - Kudenko, Daniel PY - 2007/// UR - http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.65.1349 ER -