Evolving random graph generators: A case for increased algorithmic primitive granularity

Pope, Aaron Scott; Tauritz, Daniel R.; Kent, Alexander D.

doi:10.1109/ssci.2016.7849929

Cited by 11 publications

(5 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In GP, the solutions being evolved typically take the form of programs or heuristics. GP has been shown capable of automatically generating and optimizing heuristics for problems in a variety of domains, including graph algorithm applications [2,8,25,26]. A set of primitive operations is usually constructed by observing the common and essential elements of algorithms which have been designed to solve the intended problem.…”

Section: Genetic Programmingmentioning

confidence: 99%

“…Hyperheuristics have been used to tailor heuristics to specific application domains [4], including those involving graph algorithms. Previous work investigated the use of hyper-heuristics to generate and optimize random graph generation heuristics that produce graphs with desirable characteristics, such as specific centrality distributions or community structures [2,8,26]. Customized graph partitioning heuristics have also been generated that improve upon the performance of general-purpose algorithms for targetted classes of graphs, including those representing computer networks [25].…”

Section: Related Workmentioning

confidence: 99%

“…Other selection methods considered include fitness proportional and uniform random selection. The possible values for these parameters were inspired by previous work evolving graph algorithms [25,26]. The convergence threshold was hand tuned to ensure termination in the time available.…”

Section: Hyper-heuristic Approachmentioning

confidence: 99%

“…The following categories of operations make up the set of primitives available to the GP. These operations were inspired by previous work on authentication graph analysis [15,27] and the automated design of graph-based heuristics [2,7,25,26]. Math operators: Basic addition, subtraction, multiplication, division, modulus, exponentiation, additive and multiplicative inverse.…”

Section: Primitive Operationsmentioning

confidence: 99%

“…, 1.0}) value randomly chosen once during initialization. These possible values were inspired by previous work evolving graph heuristics [25,26]. Boolean nodes: True and false constant nodes, as well as a node that randomly returns true according to an input probability.…”

Section: Primitive Operationsmentioning

confidence: 99%

See 4 more Smart Citations

Automated design of network security metrics

Pope

Morning

Tauritz

et al. 2018

Proceedings of the Genetic and Evolutionary Computation Conference Companion

Self Cite

View full text Add to dashboard Cite

Many abstract security measurements are based on characteristics of a graph that represents the network. These are typically simple and quick to compute but are often of little practical use in making real-world predictions. Practical network security is often measured using simulation or real-world exercises. These approaches better represent realistic outcomes but can be costly and time-consuming. This work aims to combine the strengths of these two approaches, developing efficient heuristics that accurately predict attack success. Hyper-heuristic machine learning techniques, trained on network attack simulation training data, are used to produce novel graphbased security metrics. These low-cost metrics serve as an approximation for simulation when measuring network security in real time. The approach is tested and verified using a simulation based on activity from an actual large enterprise network. The results demonstrate the potential of using hyper-heuristic techniques to rapidly evolve and react to emerging cybersecurity threats. CCS CONCEPTS • Security and privacy → Network security; • Software and its engineering → Genetic programming; • Mathematics of computing → Graph algorithms; Approximation algorithms;

show abstract

Section: Genetic Programmingmentioning

confidence: 99%