Predicting an adversary's capabilities, intentions, and probable vectors of attack is in general a complex arduous task. Cyber space particularly vulnerable to unforeseen attacks, as most computer networks have large, complex, opaque surface area are therefore extremely difficult analyze. Abstract adversarial models which capture the pertinent features needed for analysis, can reduce complexity sufficiently make analysis feasible. Game theory allows mathematical models; however, its...

Efficient and accurate malware detection is increasingly becoming a necessity for society to operate. Existing systems have excellent performance in identifying known which signatures are available, but poor anomaly zero day exploits not yet been made available or targeted attacks against specific entity. The primary goal of this paper provide evidence the potential learning classifier improve accuracy detection. A proof concept presented adaptive rule-based employing systems, combines...

Restricting the class of problems we want to perform well on allows Black Box Search Algorithms (BBSAs) specifically tailored that significantly outperform more general purpose problem solvers. However, fields encompass BBSAs, including Evolutionary Computing, are mostly focused improving algorithm performance over increasingly diversified classes. By definition, payoff for designing a high quality solver is far larger in terms number it can address, than specialized BBSA. This paper...

This paper presents the Coevolutionary Automated Software Correction system, which addresses in an integral and fully automated manner complete cycle of software artifact testing, error location, correction phases. It employs a coevolutionary approach where artifacts test cases are evolved tandem. The evolve to better find flaws behave specification when exposed cases, thus causing evolutionary arms race. Experimental results presented on same problem employed published previous...

For a given program, testing, locating the errors identified, and correcting those is critical, yet expensive process. The field of Search Based Software Engineering (SBSE) addresses these phases by formulating them as search problems. Coevolutionary Automated Correction (CASC) system targets correction phase coevolving test cases programs at source code level. This paper presents latest version CASC featuring multi-objective optimization an enhanced representation language. Results are...

Crossover is a core genetic operator in many evolutionary algorithms (EAs). The performance of such EAs on given problem dependent properly configuring crossover. A small set common crossover operators used the vast majority EAs, typically fixed for entire run. Selecting which to use and tuning its associated parameters obtain acceptable specific often time consuming manual process. Even then custom may be required achieve optimal performance. Finally, best configuration state This paper...

General-purpose optimization algorithms are often not well suited for real-world scenarios where many instances of the same problem class need to be repeatedly and efficiently solved. Hyper-heuristics automate design a particular scenario, making them good match solving. For instance, hardware model checking induced Boolean Satisfiability Problem (SAT) have very specific distribution which general SAT solvers necessarily targeted to. can solver customized instances.

One of the obstacles to Evolutionary Algorithms (EAs) fulfilling their promise as easy use general-purpose problem solvers, is difficulty correctly configuring them for specific problems such obtain satisfactory performance. Having a mechanism automatically parameters and operators every stage evolutionary life-cycle would give EAs more widely spread popularity in non-expert community. This paper investigates automatic configuration one stages life-cycle, parent selection, via new concept...

The No-Free-Lunch theorem is a fundamental result in the field of black-box function optimization. Recent work has shown that coevolution can exhibit free lunches. question as to which classes lunches still open. In this paper we present novel framework for analyzing like results coevolutionary algorithms. Our advantage inquiries terms solution concepts and isomorphisms on weak preference relation configurations. This allows algorithms be naturally classified by type they seek. Using also...

Fuzzy Adaptive Resonance Theory (ART) is a classic unsupervised learning algorithm. Its performance on particular clustering problem sensitive to the suitability of category function for said problem. However, ART employs fixed and thus unable benefit from potential adjust its function. This paper presents an exploration into employing evolutionary computation automated design functions obtain significantly enhanced through tailoring specific classes. We employ genetic programming powered...

Black-Box Search Algorithms (BBSAs) tailored to a specific problem class may be expected significantly outperform more general purpose solvers, including canonical evolutionary algorithms. Recent work has introduced novel approach evolving BBSAs through genetic programming hyper-heuristic. However, that first generation of hyper-heuristics suffered from over-specialization. This paper presents study on the second hyper-heuristic which employs multi-sample training alleviate...

Practitioners often need to solve real world problems for which no custom search algorithms exist. In these cases they tend use general-purpose solvers that have guarantee perform well on their specific problem. The relatively new field of hyper-heuristics provides an alternative the potential pit-falls solvers, by allowing practitioners generate a algorithm optimized problem interest. Hyper-heuristics are meta-heuristics operating space employing targeted primitives compose algorithms. This...

Abstract The genetic blueprint for the essential functions of life is encoded in DNA, which translated into proteins—the engines driving most our metabolic processes. Recent advancements genome sequencing have unveiled a vast diversity protein families, but compared with massive search space all possible amino acid sequences, set known functional families minimal. One could say nature has limited ”vocabulary.” A major question computational biologists, therefore, whether this vocabulary can...

Discovering and exploiting the linkage between genes during evolutionary search allows Linkage Tree Genetic Algorithm (LTGA) to maximize crossover effectiveness, greatly reducing both population size total number of evaluations required reach success on decomposable problems. This paper presents a comparative analysis most prominent LTGA variants newly introduced variant. While deceptive trap problem (Trap-k) is one canonical benchmarks for testing LTGA, when combined with applying steepest...

Creating an Evolutionary Algorithm (EA) which is capable of automatically configuring itself and dynamically controlling its parameters a challenging problem. However, solving this problem can reduce the amount manual configuration required to implement EA, allow EA be more adaptable, produce better results on range problems without requiring specific tuning. Using Supportive Coevolution (SuCo) evolve Self-Configuring Crossover (SCX) combines automatic technique multiple populations from...

The number of parameters that need to be manually tuned achieve good performance evolutionary algorithms and the dependency on each other make this potentially robust efficient computational method very time consuming difficult use. This paper introduces a greedy population sizing for (GPS-EA), an automated size tuning does not require any related specified or priori. Theoretical analysis function evaluations needed by GPS-EA produce solutions is provided. We also perform empirical...

Traditional evolutionary algorithms are powerful problem solvers that have several fixed parameters which require prior specification. Determining good values for any of these can be difficult, as generally very sensitive, requiring expert knowledge to set optimally without extensive use trial and error. Parameter control is a promising approach achieving this automation has the added potential increasing EA performance based on both theoretical empirical evidence optimal strategy change...

Evolutionary Algorithms (EAs) are inherently parallel due to their ability simultaneously evaluate the fitness of individuals. Synchronous Parallel EAs (SPEAs) leverage this with intent gain significant speed-ups when executed on multiple processors. However, many important problem classes lead large variations in evaluation times, such as is often case hyper-heuristics where time complexity executing one individual may differ greatly from that another. Asynchronous (APEAs) omit generational...

Random graph generation techniques provide an invaluable tool for studying related concepts. Unfortunately, traditional random models tend to produce artificial representations of real-world phenomenon. Manually developing customized every application would require unreasonable amount time and effort. In this work, a platform is developed automate the production generators that are tailored specific applications. Elements existing used create set graph-based primitive operations. A...