Results 1 
5 of
5
Countering Poisonous Inputs with Memetic Neuroevolution
"... Applied to certain problems, neuroevolution frequently gets stuck in local optima with very low fitness; in particular, this is true for some reinforcement learning problems where the input to the controller is a highdimensional and/or illchosen state description. Evidently, some controller inputs ..."
Abstract

Cited by 5 (2 self)
 Add to MetaCart
Applied to certain problems, neuroevolution frequently gets stuck in local optima with very low fitness; in particular, this is true for some reinforcement learning problems where the input to the controller is a highdimensional and/or illchosen state description. Evidently, some controller inputs are “poisonous”, and their inclusion induce such local optima. Previously, we proposed the memetic climber, which evolves neural network topology and weights at different timescales, as a solution to this problem. In this paper, we further explore the memetic climber, and introduce its populationbased counterpart: the memetic ES. We also explore which types of inputs are poisonous for two different reinforcement learning problems.
Template Method Hyperheuristics
"... The optimization literature is awash with metaphoricallyinspired metaheuristics and their subsequent variants and hybridizations. This results in a plethora of methods, with descriptions that are often polluted with the language of the metaphor which inspired them [8]. Within such a fragmented fie ..."
Abstract
 Add to MetaCart
(Show Context)
The optimization literature is awash with metaphoricallyinspired metaheuristics and their subsequent variants and hybridizations. This results in a plethora of methods, with descriptions that are often polluted with the language of the metaphor which inspired them [8]. Within such a fragmented field, the traditional approach of manual ‘operator tweaking ’ makes it difficult to establish the contribution of individual metaheuristic components to the overall success of a methodology. Irrespective of whether it happens to best the stateoftheart, such ‘tweaking ’ is so labourintensive that does relatively little to advance scientific understanding. In order to introduce further structure and rigour, it is therefore desirable to not only to be able to specify entire families of metaheuristics (rather than individual metaheuristics), but also
CostSensitive Attack Graph, Minimization Analysis
, 2009
"... To prevent an exploit, the security analyst must implement a suitable countermeasure. In this paper, we consider costsensitive attack graphs (CAGs) for network vulnerability analysis. In these attack graphs, a weight is assigned to each countermeasure to represent the cost of its implementation. Th ..."
Abstract
 Add to MetaCart
(Show Context)
To prevent an exploit, the security analyst must implement a suitable countermeasure. In this paper, we consider costsensitive attack graphs (CAGs) for network vulnerability analysis. In these attack graphs, a weight is assigned to each countermeasure to represent the cost of its implementation. There may be multiple countermeasures with different weights for preventing a single exploit. Also, a single countermeasure may prevent multiple exploits. We present a binary particle swarm optimization algorithm with a timevarying velocity clamping, called SwarmCAGTVVC, for minimization analysis of costsensitive attack graphs. The aim is to find a critical set of countermeasures with minimum weight whose implementation causes the initial nodes and the goal nodes of the graph to be completely disconnected. This problem is in fact a constrained optimization problem. A repair method is used to convert the constrained optimization problem into an unconstrained one. A local search heuristic is used to improve the overall performance of the algorithm. We compare the performance of SwarmCAGTVVC with a greedy algorithm GreedyCAG and a genetic algorithm GenNAG for minimization analysis of several largescale costsensitive attack graphs. On average, the weight of a critical set of countermeasures found by SwarmCAGTVVC is 6.15 percent less than the weight of a critical set of countermeasures found by GreedyCAG. Also, SwarmCAGTVVC performs better than GenNAG in terms of convergence speed and accuracy. The results of the experiments show that SwarmCAGTVVC can be successfully used for minimization analysis of largescale costsensitive attack graphs. c © 2010 ISC. All rights reserved. 1
An Intelligent MultiRestart Memetic Algorithm for Box Constrained Global Optimisation
"... In this paper, we propose a multirestart memetic algorithm framework for boxconstrained global continuous optimisation. In this framework, an evolutionary algorithm (EA) and a local optimizer are employed as separated building blocks. The EA is used to explore the search space for ‘very ’ promisin ..."
Abstract
 Add to MetaCart
(Show Context)
In this paper, we propose a multirestart memetic algorithm framework for boxconstrained global continuous optimisation. In this framework, an evolutionary algorithm (EA) and a local optimizer are employed as separated building blocks. The EA is used to explore the search space for ‘very ’ promising solutions (e.g. solutions in the attraction basin of the global optimum) through its exploration capability and previous EA search history, and local search is used to improve these promising solutions to local optima. An estimation of distribution algorithm (EDA) combined with a derivative free local optimizer, called NEWUOA (Powell, 2008), is developed based on this framework and empirically compared with several wellknown EAs on a set of 40 commonlyused test functions. The main components of the specific algorithm include: (1) an adaptive multivariate probability model, (2) a multiple sampling strategy, (3) decoupling of the hybridisation strategy and (4) a restart mechanism. The adaptive multivariate probability model and multiple sampling strategy are designed to enhance the exploration capability. The restart mechanism attempts to make the search escape from local optima, resorting to previous search history. Comparison results show that the algorithm is comparable with the best known EAs, including the winner of the 2005 IEEE Congress on Evolutionary Computation (CEC2005), and significantly better than the others in terms of both the solution quality and computational cost.
Particle Swarm Optimization, Attack Scenario, Countermeasure, CostSensitive Attack Graph, Minimization Analysis
, 2009
"... To prevent an exploit, the security analyst must implement a suitable countermeasure. In this paper, we consider costsensitive attack graphs (CAGs) for network vulnerability analysis. In these attack graphs, a weight is assigned to each countermeasure to represent the cost of its implementation. Th ..."
Abstract
 Add to MetaCart
(Show Context)
To prevent an exploit, the security analyst must implement a suitable countermeasure. In this paper, we consider costsensitive attack graphs (CAGs) for network vulnerability analysis. In these attack graphs, a weight is assigned to each countermeasure to represent the cost of its implementation. There may be multiple countermeasures with different weights for preventing a single exploit. Also, a single countermeasure may prevent multiple exploits. We present a binary particle swarm optimization algorithm with a timevarying velocity clamping, called SwarmCAGTVVC, for minimization analysis of costsensitive attack graphs. The aim is to find a critical set of countermeasures with minimum weight whose implementation causes the initial nodes and the goal nodes of the graph to be completely disconnected. This problem is in fact a constrained optimization problem. A repair method is used to convert the constrained optimization problem into an unconstrained one. A local search heuristic is used to improve the overall performance of the algorithm. We compare the performance of SwarmCAGTVVC with a greedy algorithm GreedyCAG and a genetic algorithm GenNAG for minimization analysis of several largescale costsensitive attack graphs. On average, the weight of a critical set of countermeasures found by SwarmCAGTVVC is 6.15 percent less than the weight of a critical set of countermeasures found by GreedyCAG. Also, SwarmCAGTVVC performs better than GenNAG in terms of convergence speed and accuracy. The results of the experiments show that SwarmCAGTVVC can be successfully used for minimization analysis of largescale costsensitive attack graphs. c ○ 2010 ISC. All rights reserved. 1