An Approach to Sensitivity Analysis of Inference Equations in Algebraic Bayesian Networks

Zolotin, Andrey A.; Mal’chevskaya, Ekaterina A.; Tulupyev, Alexander L.; Sirotkin, Alexander

doi:10.1007/978-3-319-68321-8_4

Cited by 2 publications

(1 citation statement)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Proposed method considers not only influence of every single parameter but also their interaction. In addition, this method is less costly in terms of computation and, thus, can be used as a part of Landscape Analysis (LA) procedure [8] before solving any optimization problem.…”

Section: Introductionmentioning

confidence: 99%

Meta-Optimization of Mind Evolutionary Computation Algorithm Using Design of Experiments

Sakharov

Karpenko

2018

Advances in Intelligent Systems and Computing

View full text Add to dashboard Cite

This paper presents a new technique for solving a meta-optimization problem for Mind Evolutionary Computation (MEC) algorithm using a full factorial designed experiment. This approach can be also generalized for other global optimization population-based algorithms. In general, design of experiments allows one to determine the influence of input factors and their interaction on the output of a process. It's proposed to use such an approach to identify the most important free parameters as well as to estimate their interaction and determine the optimal values of those parameters for specific classes of objective functions. The paper contains the description of proposed method and software implementation along with the results of numerical experiments conducted to determine optimal values of the MEC algorithm's free parameters.

show abstract

Section: Introductionmentioning

confidence: 99%