Limits...
Prediction of Cascading Failures in Spatial Networks.

Shunkun Y, Jiaquan Z, Dan L - PLoS ONE (2016)

Bottom Line: Cascading overload failures are widely found in large-scale parallel systems and remain a major threat to system reliability; therefore, they are of great concern to maintainers and managers of different systems.Fortunately, overload failures that propagate through a network exhibit certain spatial-temporal correlations, which allows the study of a group of nodes that share common spatial and temporal characteristics.We simulated overload failure propagations in a weighted lattice network that start with a center attack and predicted the failure percentages of different groups of nodes that are separated by a given distance.

View Article: PubMed Central - PubMed

Affiliation: School of Reliability and Systems Engineering, Beihang University, Beijing, China.

ABSTRACT
Cascading overload failures are widely found in large-scale parallel systems and remain a major threat to system reliability; therefore, they are of great concern to maintainers and managers of different systems. Accurate cascading failure prediction can provide useful information to help control networks. However, for a large, gradually growing network with increasing complexity, it is often impractical to explore the behavior of a single node from the perspective of failure propagation. Fortunately, overload failures that propagate through a network exhibit certain spatial-temporal correlations, which allows the study of a group of nodes that share common spatial and temporal characteristics. Therefore, in this study, we seek to predict the failure rates of nodes in a given group using machine-learning methods. We simulated overload failure propagations in a weighted lattice network that start with a center attack and predicted the failure percentages of different groups of nodes that are separated by a given distance. The experimental results of a feedforward neural network (FNN), a recurrent neural network (RNN) and support vector regression (SVR) all show that these different models can accurately predict the similar behavior of nodes in a given group during cascading overload propagation.

No MeSH data available.


Related in: MedlinePlus

Real value vs. predicted value of different models without considering α, n = 2500.(a) FNN. (b) RNN. (c) SVR. (d) Average error of each model.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4836660&req=5

pone.0153904.g007: Real value vs. predicted value of different models without considering α, n = 2500.(a) FNN. (b) RNN. (c) SVR. (d) Average error of each model.

Mentions: Second, to make the results more accurate, we each calculated five groups of cascading failures for α on the interval [0.4, 0.8, 1.2, 1.6, 2.0] and used these groups of data as the testing dataset. Because the training dataset included data from α on the interval [0.2, 0.6, 1.0, 1.4, 1.8, 2.2], which are different from the testing data, we increased the randomness of the α in the proposed experiments. Aside from adding 25 nodes to the input and changing the test dataset, we did not change the other parts of the models and repeated the experiment in a similar manner to that described in the last section. Because the results of all of the predicted values with the corresponding real values cannot be clearly presented in a single figure, we present them in different figures based on the different learning methods. Because the points primarily lie near the line y = x, the new model containing 50 inputs that was used to predict failure rates without knowing α still performs well. The similar results of the three methods demonstrate that each method reached an appropriate optimal solution, and the small deviations signify the stability of the methods (Fig 7).


Prediction of Cascading Failures in Spatial Networks.

Shunkun Y, Jiaquan Z, Dan L - PLoS ONE (2016)

Real value vs. predicted value of different models without considering α, n = 2500.(a) FNN. (b) RNN. (c) SVR. (d) Average error of each model.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4836660&req=5

pone.0153904.g007: Real value vs. predicted value of different models without considering α, n = 2500.(a) FNN. (b) RNN. (c) SVR. (d) Average error of each model.
Mentions: Second, to make the results more accurate, we each calculated five groups of cascading failures for α on the interval [0.4, 0.8, 1.2, 1.6, 2.0] and used these groups of data as the testing dataset. Because the training dataset included data from α on the interval [0.2, 0.6, 1.0, 1.4, 1.8, 2.2], which are different from the testing data, we increased the randomness of the α in the proposed experiments. Aside from adding 25 nodes to the input and changing the test dataset, we did not change the other parts of the models and repeated the experiment in a similar manner to that described in the last section. Because the results of all of the predicted values with the corresponding real values cannot be clearly presented in a single figure, we present them in different figures based on the different learning methods. Because the points primarily lie near the line y = x, the new model containing 50 inputs that was used to predict failure rates without knowing α still performs well. The similar results of the three methods demonstrate that each method reached an appropriate optimal solution, and the small deviations signify the stability of the methods (Fig 7).

Bottom Line: Cascading overload failures are widely found in large-scale parallel systems and remain a major threat to system reliability; therefore, they are of great concern to maintainers and managers of different systems.Fortunately, overload failures that propagate through a network exhibit certain spatial-temporal correlations, which allows the study of a group of nodes that share common spatial and temporal characteristics.We simulated overload failure propagations in a weighted lattice network that start with a center attack and predicted the failure percentages of different groups of nodes that are separated by a given distance.

View Article: PubMed Central - PubMed

Affiliation: School of Reliability and Systems Engineering, Beihang University, Beijing, China.

ABSTRACT
Cascading overload failures are widely found in large-scale parallel systems and remain a major threat to system reliability; therefore, they are of great concern to maintainers and managers of different systems. Accurate cascading failure prediction can provide useful information to help control networks. However, for a large, gradually growing network with increasing complexity, it is often impractical to explore the behavior of a single node from the perspective of failure propagation. Fortunately, overload failures that propagate through a network exhibit certain spatial-temporal correlations, which allows the study of a group of nodes that share common spatial and temporal characteristics. Therefore, in this study, we seek to predict the failure rates of nodes in a given group using machine-learning methods. We simulated overload failure propagations in a weighted lattice network that start with a center attack and predicted the failure percentages of different groups of nodes that are separated by a given distance. The experimental results of a feedforward neural network (FNN), a recurrent neural network (RNN) and support vector regression (SVR) all show that these different models can accurately predict the similar behavior of nodes in a given group during cascading overload propagation.

No MeSH data available.


Related in: MedlinePlus