Downloads & Free Reading Options - Results

Evolutionary Computation is a branch of computer science with which, traditionally, High Energy Physics has fewer connections. Its methods were investigated in this field, mainly for data analysis tasks. These methods and studies are, however, less known in the high energy physics community and this motivated us to prepare this lecture. The lecture presents a general overview of the main types of algorithms based on Evolutionary Computation, as well as a review of their applications in High Energy Physics.

Risto Miikkulainen is a computer scientist at UT Austin. Please support this podcast by checking out our sponsors: - The Jordan Harbinger Show: https://jordanharbinger.com/lex/ - Grammarly: https://grammarly.com/lex to get 20% off premium - Belcampo: https://belcampo.com/lex and use code LEX to get 20% off first order - Indeed: https://indeed.com/lex to get $75 credit EPISODE LINKS: Risto's Website: https://www.cs.utexas.edu/users/risto/ PODCAST INFO: Podcast website: https://lexfridman.com/podcast Apple Podcasts: https://apple.co/2lwqZIr Spotify: https://spoti.fi/2nEwCF8 RSS: https://lexfridman.com/feed/podcast/ YouTube Full Episodes: https://youtube.com/lexfridman YouTube Clips: https://youtube.com/lexclips SUPPORT & CONNECT: - Check out the sponsors above, it's the best way to support this podcast - Support on Patreon: https://www.patreon.com/lexfridman - Twitter:

Co-authorship graphs (that is, the graph of authors linked by co-authorship of papers) are complex networks, which expresses the dynamics of a complex system. Only recently its study has started to draw interest from the EC community, the first paper dealing with it having been published two years ago. In this paper we will study the co-authorship network of EC at a microscopic level. Our objective is ascertaining which are the most relevant nodes (i.e. authors) in it. For this purpose, we examine several metrics defined in the complex-network literature, and analyze them both in isolation and combined within a Pareto-dominance approach. The result of our analysis indicates that there are some well-known researchers that appear systematically in top rankings. This also provides some hints on the social behavior of our community.

This dissertation presents principles, techniques, and performance of evolutionary computation optimization methods. Concentration is on concepts, design formulation, and prescription for multiobjective problem solving and explicit building block (BB) multiobjective evolutionary algorithms (MOEAs). Current state-of-the-art explicit BB MOEAs are addressed in the innovative design, execution, and testing of a new multiobjective explicit BB MOEA. Evolutionary computation concepts examined are algorithm convergence, population diversity and sizing, genotype and phenotype partitioning, archiving, BB concepts, parallel evolutionary algorithm (EA) models, robustness, visualization of evolutionary process, and performance in terms of effectiveness and efficiency. The main result of this research is the development of a more robust algorithm where MOEA concepts are implicitly employed. Testing shows that the new MOEA can be more effective and efficient than previous state-of-the-art explicit BB MOEAs for selected test suite multiobjective optimization problems (MOPs) and U.S. Air Force applications. Other contributions include the extension of explicit BB definitions to clarify the meanings for good single and multiobjective BBs. A new visualization technique is developed for viewing genotype, phenotype, and the evolutionary process in finding Pareto front vectors while tracking the size of the BBs. The visualization technique is the result of a BB tracing mechanism integrated into the new MOEA that enables one to determine the required BB sizes and assign an approximation epistasis level for solving a particular problem. The culmination of this research is explicit BB state-of-the-art MOEA technology based on the MOEA design, BB classifier type assessment, solution evolution visualization, and insight into MOEA test metric validation and usage as applied to test suite, deception, bioinformatics, unmanned vehicle flight pattern, and digital symbol set design MOPs.

This project contributed new principles for the development of intelligent, mobile robots performing complex tasks in unpredictable environments. In the behavior-based approach to robot design, the overall performance of the robot arises through the interaction of multiple, relatively simple, behaviors. The manual design of multiple interacting behaviors is difficult, labor-intensive and error-prone. One way to reduce the effort in the design of behavior-based robots is to develop an evolutionary approach in which the various behaviors, as well as their modes of interaction, evolve over time. Evolution may also provide a basis for the development of strategies for multiple-robot environments, for example, environments in which a robot is expected to adapt its behavior based on the current behavior of other agents or environmental conditions which themselves are changing over time. This project addressed in four complementary areas concerning the effectiveness of evolutionary algorithms for the design of autonomous robots: (1) learning multiple behaviors by asynchronous co-evolution; (2) continuous and embedded learning; (3) comparison with other reinforcement learning methods, and (4) the ability to evolve responses to changing environments. Results in each of these tasks are reported.

We investigate the ability of a genetic algorithm to design cellular automata that perform computations. The computational strategies of the resulting cellular automata can be understood using a framework in which ``particles'' embedded in space-time configurations carry information and interactions between particles effect information processing. This structural analysis can also be used to explain the evolutionary process by which the strategies were designed by the genetic algorithm. More generally, our goals are to understand how machine-learning processes can design complex decentralized systems with sophisticated collective computational abilities and to develop rigorous frameworks for understanding how the resulting dynamical systems perform computation.

We investigate the ability of a genetic algorithm to design cellular automata that perform computations. The computational strategies of the resulting cellular automata can be understood using a framework in which ``particles'' embedded in space-time configurations carry information and interactions between particles effect information processing. This structural analysis can also be used to explain the evolutionary process by which the strategies were designed by the genetic algorithm. More generally, our goals are to understand how machine-learning processes can design complex decentralized systems with sophisticated collective computational abilities and to develop rigorous frameworks for understanding how the resulting dynamical systems perform computation.

We investigate the ability of a genetic algorithm to design cellular automata that perform computations. The computational strategies of the resulting cellular automata can be understood using a framework in which ``particles'' embedded in space-time configurations carry information and interactions between particles effect information processing. This structural analysis can also be used to explain the evolutionary process by which the strategies were designed by the genetic algorithm. More generally, our goals are to understand how machine-learning processes can design complex decentralized systems with sophisticated collective computational abilities and to develop rigorous frameworks for understanding how the resulting dynamical systems perform computation.

We investigate the ability of a genetic algorithm to design cellular automata that perform computations. The computational strategies of the resulting cellular automata can be understood using a framework in which ``particles'' embedded in space-time configurations carry information and interactions between particles effect information processing. This structural analysis can also be used to explain the evolutionary process by which the strategies were designed by the genetic algorithm. More generally, our goals are to understand how machine-learning processes can design complex decentralized systems with sophisticated collective computational abilities and to develop rigorous frameworks for understanding how the resulting dynamical systems perform computation.

EO contains classes for any kind of evolutionary computation, especially genetic algorithms, that you might encounter. It is component-based so that if you don't find the class you need in it, it is very easy to subclass the existing abstract or concrete classes.

EO contains classes for any kind of evolutionary computation, especially genetic algorithms, that you might encounter. It is component-based so that if you don't find the class you need in it, it is very easy to subclass the existing abstract or concrete classes.

EO contains classes for any kind of evolutionary computation, especially genetic algorithms, that you might encounter. It is component-based so that if you don't find the class you need in it, it is very easy to subclass the existing abstract or concrete classes.

The discovery and study of a vast number of regulatory RNAs in all kingdoms of life over the past decades has allowed the design of new synthetic RNAs that can regulate gene expression in vivo. Riboregulators, in particular, have been used to activate or repress gene expression. However, to accelerate and scale up the design process, synthetic biologists require computer-assisted design tools, without which riboregulator engineering will remain a case-by-case design process requiring expert attention. Recently, the design of RNA circuits by evolutionary computation and adapting strand displacement techniques from nanotechnology has proven to be suited to the automated generation of DNA sequences implementing regulatory RNA systems in bacteria. Herein, we present our method to carry out such evolutionary design and how to use it to create various types of riboregulators, allowing the systematic de novo design of genetic control systems in synthetic biology.

The discovery and study of a vast number of regulatory RNAs in all kingdoms of life over the past decades has allowed the design of new synthetic RNAs that can regulate gene expression in vivo. Riboregulators, in particular, have been used to activate or repress gene expression. However, to accelerate and scale up the design process, synthetic biologists require computer-assisted design tools, without which riboregulator engineering will remain a case-by-case design process requiring expert attention. Recently, the design of RNA circuits by evolutionary computation and adapting strand displacement techniques from nanotechnology has proven to be suited to the automated generation of DNA sequences implementing regulatory RNA systems in bacteria. Herein, we present our method to carry out such evolutionary design and how to use it to create various types of riboregulators, allowing the systematic de novo design of genetic control systems in synthetic biology.

This work presents a new concept for finding the optimal values for the entire three fundamental design vectors namely: scheduling, projection and processor so that not only architecture design could be feasible along with that maximum hardware utilizing efficiency could be achieved. This Approach also having the focus to minimize the total delay involved with systolic architecture design. Evolutionary programming has applied to find the optimal solution. Presented work and result will provide facility to designer without any involvement to find out best suited architecture for a particular application. The Proposed method having capability to find the large number of optimal vectors for any algorithm which can be implemented in systolic architecture. The increasing demands of speed and performance in modern signal and image processing applications necessitate a revolutionary super -computing technology. The proposed method is coded in MATLAB editor and simulation environment. Shilpa V | Suma V Shetty"Design of Systolic Architecture Using Evolutionary Computation" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-2 | Issue-4 , June 2018, URL: http://www.ijtsrd.com/papers/ijtsrd15776.pdf http://www.ijtsrd.com/engineering/electronics-and-communication-engineering/15776/design-of-systolic-architecture-using-evolutionary-computation/shilpa-v

Hybrid optimization algorithms have gained popularity as it has become apparent there cannot be a universal optimization strategy which is globally more beneficial than any other. Despite their popularity, hybridization frameworks require more detailed categorization regarding: the nature of the problem domain, the constituent algorithms, the coupling schema and the intended area of application. This report proposes a hybrid algorithm for solving small to large-scale continuous global optimization problems. It comprises evolutionary computation (EC) algorithms and a sequential quadratic programming (SQP) algorithm; combined in a collaborative portfolio. The SQP is a gradient based local search method. To optimize the individual contributions of the EC and SQP algorithms for the overall success of the proposed hybrid system, improvements were made in key features of these algorithms. The report proposes enhancements in: i) the evolutionary algorithm, ii) a new convergence detection mechanism was proposed; and iii) in the methods for evaluating the search directions and step sizes for the SQP local search algorithm. The proposed hybrid design aim was to ensure that the two algorithms complement each other by exploring and exploiting the problem search space. Preliminary results justify that an adept hybridization of evolutionary algorithms with a suitable local search method, could yield a robust and efficient means of solving wide range of global optimization problems. Finally, a discussion of the outcomes of the initial investigation and a review of the associated challenges and inherent limitations of the proposed method is presented to complete the investigation. The report highlights extensive research, particularly, some potential case studies and application areas.

Hybrid optimization algorithms have gained popularity as it has become apparent there cannot be a universal optimization strategy which is globally more beneficial than any other. Despite their popularity, hybridization frameworks require more detailed categorization regarding: the nature of the problem domain, the constituent algorithms, the coupling schema and the intended area of application. This report proposes a hybrid algorithm for solving small to large-scale continuous global optimization problems. It comprises evolutionary computation (EC) algorithms and a sequential quadratic programming (SQP) algorithm; combined in a collaborative portfolio. The SQP is a gradient based local search method. To optimize the individual contributions of the EC and SQP algorithms for the overall success of the proposed hybrid system, improvements were made in key features of these algorithms. The report proposes enhancements in: i) the evolutionary algorithm, ii) a new convergence detection mechanism was proposed; and iii) in the methods for evaluating the search directions and step sizes for the SQP local search algorithm. The proposed hybrid design aim was to ensure that the two algorithms complement each other by exploring and exploiting the problem search space. Preliminary results justify that an adept hybridization of evolutionary algorithms with a suitable local search method, could yield a robust and efficient means of solving wide range of global optimization problems. Finally, a discussion of the outcomes of the initial investigation and a review of the associated challenges and inherent limitations of the proposed method is presented to complete the investigation. The report highlights extensive research, particularly, some potential case studies and application areas.

Hybrid optimization algorithms have gained popularity as it has become apparent there cannot be a universal optimization strategy which is globally more beneficial than any other. Despite their popularity, hybridization frameworks require more detailed categorization regarding: the nature of the problem domain, the constituent algorithms, the coupling schema and the intended area of application. This report proposes a hybrid algorithm for solving small to large-scale continuous global optimization problems. It comprises evolutionary computation (EC) algorithms and a sequential quadratic programming (SQP) algorithm; combined in a collaborative portfolio. The SQP is a gradient based local search method. To optimize the individual contributions of the EC and SQP algorithms for the overall success of the proposed hybrid system, improvements were made in key features of these algorithms. The report proposes enhancements in: i) the evolutionary algorithm, ii) a new convergence detection mechanism was proposed; and iii) in the methods for evaluating the search directions and step sizes for the SQP local search algorithm. The proposed hybrid design aim was to ensure that the two algorithms complement each other by exploring and exploiting the problem search space. Preliminary results justify that an adept hybridization of evolutionary algorithms with a suitable local search method, could yield a robust and efficient means of solving wide range of global optimization problems. Finally, a discussion of the outcomes of the initial investigation and a review of the associated challenges and inherent limitations of the proposed method is presented to complete the investigation. The report highlights extensive research, particularly, some potential case studies and application areas.

Hybrid optimization algorithms have gained popularity as it has become apparent there cannot be a universal optimization strategy which is globally more beneficial than any other. Despite their popularity, hybridization frameworks require more detailed categorization regarding: the nature of the problem domain, the constituent algorithms, the coupling schema and the intended area of application. This report proposes a hybrid algorithm for solving small to large-scale continuous global optimization problems. It comprises evolutionary computation (EC) algorithms and a sequential quadratic programming (SQP) algorithm; combined in a collaborative portfolio. The SQP is a gradient based local search method. To optimize the individual contributions of the EC and SQP algorithms for the overall success of the proposed hybrid system, improvements were made in key features of these algorithms. The report proposes enhancements in: i) the evolutionary algorithm, ii) a new convergence detection mechanism was proposed; and iii) in the methods for evaluating the search directions and step sizes for the SQP local search algorithm. The proposed hybrid design aim was to ensure that the two algorithms complement each other by exploring and exploiting the problem search space. Preliminary results justify that an adept hybridization of evolutionary algorithms with a suitable local search method, could yield a robust and efficient means of solving wide range of global optimization problems. Finally, a discussion of the outcomes of the initial investigation and a review of the associated challenges and inherent limitations of the proposed method is presented to complete the investigation. The report highlights extensive research, particularly, some potential case studies and application areas.

Hybrid optimization algorithms have gained popularity as it has become apparent there cannot be a universal optimization strategy which is globally more beneficial than any other. Despite their popularity, hybridization frameworks require more detailed categorization regarding: the nature of the problem domain, the constituent algorithms, the coupling schema and the intended area of application. This report proposes a hybrid algorithm for solving small to large-scale continuous global optimization problems. It comprises evolutionary computation (EC) algorithms and a sequential quadratic programming (SQP) algorithm; combined in a collaborative portfolio. The SQP is a gradient based local search method. To optimize the individual contributions of the EC and SQP algorithms for the overall success of the proposed hybrid system, improvements were made in key features of these algorithms. The report proposes enhancements in: i) the evolutionary algorithm, ii) a new convergence detection mechanism was proposed; and iii) in the methods for evaluating the search directions and step sizes for the SQP local search algorithm. The proposed hybrid design aim was to ensure that the two algorithms complement each other by exploring and exploiting the problem search space. Preliminary results justify that an adept hybridization of evolutionary algorithms with a suitable local search method, could yield a robust and efficient means of solving wide range of global optimization problems. Finally, a discussion of the outcomes of the initial investigation and a review of the associated challenges and inherent limitations of the proposed method is presented to complete the investigation. The report highlights extensive research, particularly, some potential case studies and application areas.

In the manuscript titled "Computation environment (1)", we introduced a notion called computation environment as an interactive model for computation and complexity theory. In this model, Turing machines are not autonomous entities and find their meanings through the interaction between a computist and a universal processor, and thus due to evolution of the universal processor, the meanings of Turing machines could change. In this manuscript, we discuss persistently evolutionary intensions. We introduce a new semantics, called persistently evolutionary semantics, for predicate logic that the meaning of function and predicate symbols are not already predetermined, and predicate and function symbols find their meaning through the interaction of the subject with the language. In (classic) model theory, the mathematician who studies a structure is assumed as a god who lives out of the structure, and the study of the mathematician does not effect the structure. The meaning of predicate and function symbols are assumed to be independent of the mathematician who does math. The persistently evolutionary semantics could be regarded as a start of "Interactive Model Theory" as a new paradigm in model theory (similar to the paradigm of interactive computation). In interactive model theory, we suppose that a mathematical structure should consist of two parts: 1) an intelligent agent (a subject), and 2) an environment (language), and every things should find its meaning through the interaction of these two parts. We introduce persistently evolutionary Kripke structure for propositional and predicate logic. Also, we propose a persistently evolutionary Kripke semantics for the notion of computation, where the intension of a code of a Turing machine persistently evolve. We show that in this Kripke model the subject can never know P = NP.

Mutation has traditionally been regarded as an important operator in evolutionary algorithms. In particular, there have been many experimental studies which showed the effectiveness of adapting mutation rates for various static optimization problems. Given the perceived effectiveness of adaptive and self-adaptive mutation for static optimization problems, there have been speculations that adaptive and self-adaptive mutation can benefit dynamic optimization problems even more since adaptation and self-adaptation are capable of following a dynamic environment. However, few theoretical results are available in analyzing rigorously evolutionary algorithms for dynamic optimization problems. It is unclear when adaptive and self-adaptive mutation rates are likely to be useful for evolutionary algorithms in solving dynamic optimization problems. This paper provides the first rigorous analysis of adaptive mutation and its impact on the computation times of evolutionary algorithms in solving certain dynamic optimization problems. More specifically, for both individual-based and population-based EAs, we have shown that any time-variable mutation rate scheme will not significantly outperform a fixed mutation rate on some dynamic optimization problem instances. The proofs also offer some insights into conditions under which any time-variable mutation scheme is unlikely to be useful and into the relationships between the problem characteristics and algorithmic features (e.g., different mutation schemes).

Mutation has traditionally been regarded as an important operator in evolutionary algorithms. In particular, there have been many experimental studies which showed the effectiveness of adapting mutation rates for various static optimization problems. Given the perceived effectiveness of adaptive and self-adaptive mutation for static optimization problems, there have been speculations that adaptive and self-adaptive mutation can benefit dynamic optimization problems even more since adaptation and self-adaptation are capable of following a dynamic environment. However, few theoretical results are available in analyzing rigorously evolutionary algorithms for dynamic optimization problems. It is unclear when adaptive and self-adaptive mutation rates are likely to be useful for evolutionary algorithms in solving dynamic optimization problems. This paper provides the first rigorous analysis of adaptive mutation and its impact on the computation times of evolutionary algorithms in solving certain dynamic optimization problems. More specifically, for both individual-based and population-based EAs, we have shown that any time-variable mutation rate scheme will not significantly outperform a fixed mutation rate on some dynamic optimization problem instances. The proofs also offer some insights into conditions under which any time-variable mutation scheme is unlikely to be useful and into the relationships between the problem characteristics and algorithmic features (e.g., different mutation schemes).

pages

pages

This dissertation investigates several innovative approaches to evolutionary optimization that are relevant to numerous applications in astronautical engineering. The challenges and shortfalls associated with evolutionary algorithms are translated into three overarching goals that directly motivate the research and innovations of this dissertation. The first goal is to investigate and employ techniques that enable evolutionary algorithms to effectively handle constraints in a way that allows for feasible solutions to constrained optimization problems. The second goal is to improve computation times and efficiencies associated with evolutionary algorithms. The last goal is to enhance the evolutionary algorithm's robustness and ability to consistently find accurate solutions within a finite number of iterations. Novel techniques involving the application of unscented sampling, parallel computation, and various forms of exact penalty functions are developed and applied to both genetic algorithms and evolution strategies to achieve these goals. The results of this research offer a promising new set of modified evolutionary algorithms that outperform state-of-the-art techniques on a number of challenging multimodal optimization problems. In addition, these new methods are shown to be very effective in solving a minimum-propellant lunar lander optimal control problem, representing a class of problems that are historically difficult to solve using evolutionary algorithms.

EC paper authors form a complex network of co-authorship which is, by itself, a example of an evolving system with its own rules, concept of fitness, and patterns of attachment. In this paper we explore the network of authors of evolutionary computation papers found in a major bibliographic database. We examine its macroscopic properties, and compare it with other co-authorship networks; the EC co-authorship network yields results in the same ballpark as other networks, but exhibits some distinctive patterns in terms of internal cohesion. We also try to find some hints on what makes an author a sociometric star. Finally, the role of proceeding editorship as the origin of long-range links in the co-authorship network is studied as well.

The challenge of ad-hoc computing is to find the way of taking advantage of spare cycles in an efficient way that takes into account all capabilities of the devices and interconnections available to them. In this paper we explore distributed evolutionary computation based on the Ruby on Rails framework, which overlays a Model-View-Controller on evolutionary computation. It allows anybody with a web browser (that is, mostly everybody connected to the Internet) to participate in an evolutionary computation experiment. Using a straightforward farming model, we consider different factors, such as the size of the population used. We are mostly interested in how they impact on performance, but also the scaling behavior when a non-trivial number of computers is applied to the problem. Experiments show the impact of different packet sizes on performance, as well as a quite limited scaling behavior, due to the characteristics of the server. Several solutions for that problem are proposed.

The challenge of ad-hoc computing is to find the way of taking advantage of spare cycles in an efficient way that takes into account all capabilities of the devices and interconnections available to them. In this paper we explore distributed evolutionary computation based on the Ruby on Rails framework, which overlays a Model-View-Controller on evolutionary computation. It allows anybody with a web browser (that is, mostly everybody connected to the Internet) to participate in an evolutionary computation experiment. Using a straightforward farming model, we consider different factors, such as the size of the population used. We are mostly interested in how they impact on performance, but also the scaling behavior when a non-trivial number of computers is applied to the problem. Experiments show the impact of different packet sizes on performance, as well as a quite limited scaling behavior, due to the characteristics of the server. Several solutions for that problem are proposed.

The challenge of ad-hoc computing is to find the way of taking advantage of spare cycles in an efficient way that takes into account all capabilities of the devices and interconnections available to them. In this paper we explore distributed evolutionary computation based on the Ruby on Rails framework, which overlays a Model-View-Controller on evolutionary computation. It allows anybody with a web browser (that is, mostly everybody connected to the Internet) to participate in an evolutionary computation experiment. Using a straightforward farming model, we consider different factors, such as the size of the population used. We are mostly interested in how they impact on performance, but also the scaling behavior when a non-trivial number of computers is applied to the problem. Experiments show the impact of different packet sizes on performance, as well as a quite limited scaling behavior, due to the characteristics of the server. Several solutions for that problem are proposed.

Analogy plays an important role in creativity, and is extensively used in science as well as art. In this paper we introduce a technique for the automated generation of cross-domain analogies based on a novel evolutionary algorithm (EA). Unlike existing work in computational analogy-making restricted to creating analogies between two given cases, our approach, for a given case, is capable of creating an analogy along with the novel analogous case itself. Our algorithm is based on the concept of "memes", which are units of culture, or knowledge, undergoing variation and selection under a fitness measure, and represents evolving pieces of knowledge as semantic networks. Using a fitness function based on Gentner's structure mapping theory of analogies, we demonstrate the feasibility of spontaneously generating semantic networks that are analogous to a given base network.

Analogy plays an important role in creativity, and is extensively used in science as well as art. In this paper we introduce a technique for the automated generation of cross-domain analogies based on a novel evolutionary algorithm (EA). Unlike existing work in computational analogy-making restricted to creating analogies between two given cases, our approach, for a given case, is capable of creating an analogy along with the novel analogous case itself. Our algorithm is based on the concept of "memes", which are units of culture, or knowledge, undergoing variation and selection under a fitness measure, and represents evolving pieces of knowledge as semantic networks. Using a fitness function based on Gentner's structure mapping theory of analogies, we demonstrate the feasibility of spontaneously generating semantic networks that are analogous to a given base network.

Analogy plays an important role in creativity, and is extensively used in science as well as art. In this paper we introduce a technique for the automated generation of cross-domain analogies based on a novel evolutionary algorithm (EA). Unlike existing work in computational analogy-making restricted to creating analogies between two given cases, our approach, for a given case, is capable of creating an analogy along with the novel analogous case itself. Our algorithm is based on the concept of "memes", which are units of culture, or knowledge, undergoing variation and selection under a fitness measure, and represents evolving pieces of knowledge as semantic networks. Using a fitness function based on Gentner's structure mapping theory of analogies, we demonstrate the feasibility of spontaneously generating semantic networks that are analogous to a given base network.

XV, 724 Seiten :

XV, 724 Seiten :

This erratum points out an error in the simplified drift theorem (SDT) [Algorithmica 59(3), 369-386, 2011]. It is also shown that a minor modification of one of its conditions is sufficient to establish a valid result. In many respects, the new theorem is more general than before. We no longer assume a Markov process nor a finite search space. Furthermore, the proof of the theorem is more compact than the previous ones. Finally, previous applications of the SDT are revisited. It turns out that all of these either meet the modified condition directly or by means of few additional arguments.

This paper analises distributed evolutionary computation based on the Representational State Transfer (REST) protocol, which overlays a farming model on evolutionary computation. An approach to evolutionary distributed optimisation of multilayer perceptrons (MLP) using REST and language Perl has been done. In these experiments, a master-slave based evolutionary algorithm (EA) has been implemented, where slave processes evaluate the costly fitness function (training a MLP to solve a classification problem). Obtained results show that the parallel version of the developed programs obtains similar or better results using much less time than the sequential version, obtaining a good speedup.

This paper analises distributed evolutionary computation based on the Representational State Transfer (REST) protocol, which overlays a farming model on evolutionary computation. An approach to evolutionary distributed optimisation of multilayer perceptrons (MLP) using REST and language Perl has been done. In these experiments, a master-slave based evolutionary algorithm (EA) has been implemented, where slave processes evaluate the costly fitness function (training a MLP to solve a classification problem). Obtained results show that the parallel version of the developed programs obtains similar or better results using much less time than the sequential version, obtaining a good speedup.

This paper analises distributed evolutionary computation based on the Representational State Transfer (REST) protocol, which overlays a farming model on evolutionary computation. An approach to evolutionary distributed optimisation of multilayer perceptrons (MLP) using REST and language Perl has been done. In these experiments, a master-slave based evolutionary algorithm (EA) has been implemented, where slave processes evaluate the costly fitness function (training a MLP to solve a classification problem). Obtained results show that the parallel version of the developed programs obtains similar or better results using much less time than the sequential version, obtaining a good speedup.

This paper analises distributed evolutionary computation based on the Representational State Transfer (REST) protocol, which overlays a farming model on evolutionary computation. An approach to evolutionary distributed optimisation of multilayer perceptrons (MLP) using REST and language Perl has been done. In these experiments, a master-slave based evolutionary algorithm (EA) has been implemented, where slave processes evaluate the costly fitness function (training a MLP to solve a classification problem). Obtained results show that the parallel version of the developed programs obtains similar or better results using much less time than the sequential version, obtaining a good speedup.

This paper analises distributed evolutionary computation based on the Representational State Transfer (REST) protocol, which overlays a farming model on evolutionary computation. An approach to evolutionary distributed optimisation of multilayer perceptrons (MLP) using REST and language Perl has been done. In these experiments, a master-slave based evolutionary algorithm (EA) has been implemented, where slave processes evaluate the costly fitness function (training a MLP to solve a classification problem). Obtained results show that the parallel version of the developed programs obtains similar or better results using much less time than the sequential version, obtaining a good speedup.

The rising sophistication of cyber threats as well as the improvement of physical computer network properties present increasing challenges to contemporary Intrusion Detection (ID) techniques. To respond to these challenges, a multi agent system (MAS) coupled with flow-based ID techniques may effectively complement traditional ID systems. This paper develops: 1) a scalable software architecture for a new, self-organized, multi agent, flow-based ID system; and 2) a network simulation environment suitable for evaluating implementations of this MAS architecture and for other research purposes. Self-organization is achieved via 1) a reputation system that influences agent mobility in the search for effective vantage points in the network; and 2) multi objective evolutionary algorithms that seek effective operational parameter values. This paper illustrates, through quantitative and qualitative evaluation, 1) the conditions for which the reputation system provides a significant benefit; and 2) essential functionality of a complex network simulation environment supporting a broad range of malicious activity scenarios. These results establish an optimistic outlook for further research in flow-based multi agent systems for ID in computer networks.

Volume 2-Computational Fluid dynamics and Wave Propagation; Parallel Computing; Concurrent and Supercomputing; Computational Physics/Computational Chemistry and Evolutionary Systems.

In evolutionary graph theory biologists study the problem of determining the probability that a small number of mutants overtake a population that is structured on a weighted, possibly directed graph. Currently Monte Carlo simulations are used for estimating such fixation probabilities on directed graphs, since no good analytical methods exist. In this paper, we introduce a novel deterministic algorithm for computing fixation probabilities for strongly connected directed, weighted evolutionary graphs under the case of neutral drift, which we show to be a lower bound for the case where the mutant is more fit than the rest of the population (previously, this was only observed from simulation). We also show that, in neutral drift, fixation probability is additive under the weighted, directed case. We implement our algorithm and show experimentally that it consistently outperforms Monte Carlo simulations by several orders of magnitude, which can allow researchers to study fixation probability on much larger graphs.

Leading polymorphous computing architecture (PCA) efforts include the Raw Architecture Workstation (RAW) and the Tera-op Reliable and Intelligently Adaptive Processing System (TRIPS), both of which are tile-based. The Raw toolchain places responsibility for program decomposition on the programmer but the TRIPS toolchain automatically generates hyperblocks and allocates them to processing elements. This report identifies evolutionary computation (EC) techniques that enable and that are enabled by PCA technology, focusing on application of EC in enhancing the effectiveness of the TRIPS toolchain, including the Scalable Compiler for Analytic Experiments (SCALE) compiler. In particular, computational experiments are described that investigate the application of genetic programming to the meta-optimization of the priority function used to increase the number of instructions per hyperblock in the in-lining optimization phase of SCALE.

This dissertation investigates several innovative approaches to evolutionary optimization that are relevant to numerous applications in astronautical engineering. The challenges and shortfalls associated with evolutionary algorithms are translated into three overarching goals that directly motivate the research and innovations of this dissertation. The first goal is to investigate and employ techniques that enable evolutionary algorithms to effectively handle constraints in a way that allows for feasible solutions to constrained optimization problems. The second goal is to improve computation times and efficiencies associated with evolutionary algorithms. The last goal is to enhance the evolutionary algorithms robustness and ability to consistently find accurate solutions within a finite number of iterations. Novel techniques involving the application of unscented sampling, parallel computation, and various forms of exact penalty functions are developed and applied to both genetic algorithms and evolution strategies to achieve these goals. The results of this research offer a promising new set of modified evolutionary algorithms that outperform state-of-the-art techniques on a number of challenging multimodal optimization problems. In addition, these new methods are shown to be very effective in solving a minimum-propellant lunar lander optimal control problem, representing a class of problems that are historically difficult to solve using evolutionary algorithms.

The standardization of very-large-scale integration (VLSI) physical architecture for VLSI chips and multichip platforms is now in its early stages of development. The purpose of VLSI partitioning is to divide the circuit into numerous smaller circuits with few connections in between. Partitioning is the fundamental problem in circuit design and division. The efficient method of evolutionary computation may be used to tackle the partitioning problem in VLSI circuit design. It provides a heuristic approach to solve this problem by exploring the solution space and incrementally improving the quality of the solutions. In order to obtain the shortest wire length (WL), area, and connections, an evolutionary optimized simulated annealing memetic algorithm (OSAMA) that incorporates one or more local search phases inside its evolutionary cycle was developed.

The standardization of very-large-scale integration (VLSI) physical architecture for VLSI chips and multichip platforms is now in its early stages of development. The purpose of VLSI partitioning is to divide the circuit into numerous smaller circuits with few connections in between. Partitioning is the fundamental problem in circuit design and division. The efficient method of evolutionary computation may be used to tackle the partitioning problem in VLSI circuit design. It provides a heuristic approach to solve this problem by exploring the solution space and incrementally improving the quality of the solutions. In order to obtain the shortest wire length (WL), area, and connections, an evolutionary optimized simulated annealing memetic algorithm (OSAMA) that incorporates one or more local search phases inside its evolutionary cycle was developed.

The two most important criteria for vectorial Boolean functions used as S-boxes in block ciphers are differential uniformity and nonlinearity. Previous work in this field has focused only on nonlinearity and a different criterion, autocorrelation. In this paper, we describe the results of experiments in using simulated annealing, memetic algorithms, and ant colony optimisation to create vectorial Boolean functions with low differential uniformity. Keywords: Metaheuristics, simulated annealing, memetic algorithms, ant colony optimization, cryptography, Boolean functions, vectorial Boolean functions.