On the impact of objective function transformations on evolutionary and black-box algorithms

Storch, Tobias

doi:10.1007/s10710-006-9004-8

Cited by 2 publications

(2 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some restricted sets of transformations of the objective function were also studied [13]. There, however, the authors were concerned with conceptual arguments on what they coined "neutral transformations".…”

Section: Genetic Algorithmsmentioning

confidence: 99%

Stochastic tunneling transformation during selection in genetic algorithm

Mayer

Hamacher

2014

Proceedings of the 2014 Annual Conference on Genetic and Evolutionary Computation

View full text Add to dashboard Cite

Genetic Algorithms (GA) combine mutational and recombination operators to then select between individuals. Thereby, competition becomes the driving force to improve solutions. Now, this naive approach to biological evolution often assumes a static fitness function, e.g., co-evolutionary effects cannot easily be leveraged.Here, we introduce a fitness landscape transformation inspired by Monte-Carlo-based optimization schemes. In the StochasticTunneling (STUN) framework fitness values are non-linearly transformed under preservation of the relative ranking of optima. The "base line" of the STUN-transformation can be set based on different memory mechanisms -from current to full history.This STUN-based GA-variant allows to include co-evolution and history into the GA. Based on analytic arguments we can show that the non-linearity of the transformation generates high population densities in areas of interest.We numerically simulated small, controllable, and well understood test instance: replicas of Ising-spin glasses. For these systems the STUN-GAs have shown significant improvements in terms of relative error for given computational effort. In addition, we introduce an empirical measure of selection to discuss the improved convergence behavior.

show abstract

Section: Genetic Algorithmsmentioning

confidence: 99%

Stochastic tunneling transformation during selection in genetic algorithm

Mayer

Hamacher

2014

Proceedings of the 2014 Annual Conference on Genetic and Evolutionary Computation

View full text Add to dashboard Cite

show abstract

“…, µ}, is chosen with probability 1/µ) is used in Step 2 of the (µ + 1) EA. How and which transformations of the objective function do not affect the optimization behavior of the (µ + 1) EA and variants thereof at all was investigated in detail by Storch (2006). There it is examined how different operators and classes of heuristics behave when the function to be optimized is transformed.…”

Section: Contributions and Further Article Structurementioning

confidence: 99%

On the Choice of the Parent Population Size

Storch

2008

Evolutionary Computation

Self Cite

View full text Add to dashboard Cite

Evolutionary algorithms (EAs) are population-based randomized search heuristics that often solve problems successfully. Here the focus is on the possible effects of changing the parent population size in a simple, but still realistic, mutation-based EA. It preserves diversity by avoiding duplicates in its population. On the one hand its behavior on well-known pseudo-Boolean example functions is investigated by means of a rigorous runtime analysis. A comparison with the expected runtime of the algorithm's variant that does not avoid duplicates demonstrates the strengths and weaknesses of maintaining diversity. On the other hand, newly developed functions are presented for which the optimizer considered that even a decrease of the population size by a single increment leads from efficient optimization to enormous runtime and overwhelming probability. This is proven for all feasible population sizes and thereby this result forms a hierarchy theorem. In order to obtain all these results new methods for the analysis of the EA are developed.

show abstract

On the impact of objective function transformations on evolutionary and black-box algorithms

Cited by 2 publications

References 10 publications

Stochastic tunneling transformation during selection in genetic algorithm

Stochastic tunneling transformation during selection in genetic algorithm

On the Choice of the Parent Population Size

Contact Info

Product

Resources

About