On coevolutionary genetic algorithms

 The use of evolutionary computing techniques in coevolutionary/multi-agent systems is becoming increasingly popular. This paper presents simple models of the genetic algorithm in such systems, with the aim of examining the effects of different types of interdependence between individuals. Using the model it is shown that, for a fixed amount of interdependence between coevolving individuals, the existence of partner gene variance and the level at which fitness is applied can have significant effects, as does the evaluation partnering strategy used.     

Nonlinear System Identification Using Coevolution of Models and Tests

We present a coevolutionary algorithm for inferring the topology and parameters of a wide range of hidden nonlinear systems with a minimum of experimentation on the target system. The algorithm synthesizes an explicit model directly from the observed data produced by intelligently generated tests. The algorithm is composed of two coevolving populations. One population evolves candidate models that estimate the structure of the hidden system. The second population evolves informative tests that either extract new information from the hidden system or elicit desirable behavior from it. The fitness of candidate models is their ability to explain behavior of the target system observed in response to all tests carried out so far; the fitness of candidate tests is their ability to make the models disagree in their predictions. We demonstrate the generality of this estimation-exploration algorithm by applying it to four different problems—grammar induction, gene network inference, evolutionary robotics, and robot damage recovery—and discuss how it overcomes several of the pathologies commonly found in other coevolutionary algorithms. We show that the algorithm is able to successfully infer and/or manipulate highly nonlinear hidden systems using very few tests, and that the benefit of this approach increases as the hidden systems possess more degrees of freedom, or become more biased or unobservable. The algorithm provides a systematic method for posing synthesis or analysis tasks to a coevolutionary system.     

Coevolution and the Red Queen effect shape virtual plants

According to the Red Queen hypothesis a population of individuals may be improving some trait even though fitness remains constant. We have tested this hypothesis using a population of virtual plants. The plants have to compete with each other for virtual sunlight. Plants are modeled using Lindenmayer systems and rendered with OpenGL. Reproductive success of a plant depends on the amount of virtual light received as well as on the structural complexity of the plant. We experiment with two different modes of evaluation. In one experiment, plants are evaluated in isolation, while in other experiments plants are evaluated using coevolution. When using coevolution plants have to compete with each other for sunlight inside the same environment. Coevolution produces much thinner and taller plants in comparison to bush-like plants which are obtained when plants are evaluated in isolation. The presence of other individuals leads to an evolutionary arms race. Because plants are evaluated inside the same environment, the leaves of one plant may be shadowed by other plants. In an attempt to gain more sunlight, plants grow higher and higher. The Red Queen effect was observed when individuals of a single population were coevolving. Communicated by: Una-May O'Reilly     

Evolutionary consequences of coevolving targets

Most evolutionary optimization models incorporate a fitness evaluation that is based on a predefined static set of test cases or problems. In the natural evolutionary process, selection is of course not based on a static fitness evaluation. Organisms do not have to combat every existing disease during their lifespan; organisms of one species may live in different or changing environments; different species coevolve. This leads to the question of how information is integrated over many generations. This study focuses on the effects of different fitness evaluation schemes on the types of genotypes and phenotypes that evolve. The evolutionary target is a simple numerical function. The genetic representation is in the form of a program (i.e., a functional representation, as in genetic programming). Many different programs can code for the same numerical function. In other words, there is a many-to-one mapping between "genotypes" (the programs) and "phenotypes". We compare fitness evaluation based on a large static set of problems and fitness evaluation based on small coevolving sets of problems. In the latter model very little information is presented to the evolving programs regarding the evolutionary target per evolutionary time step. In other words, the fitness evaluation is very sparse. Nevertheless the model produces correct solutions to the complete evolutionary target in about half of the simulations. The complete evaluation model, on the other hand, does not find correct solutions to the target in any of the simulations. More important, we find that sparse evaluated programs are better generalizable compared to the complete evaluated programs when they are evaluated on a much denser set of problems. In addition, the two evaluation schemes lead to programs that differ with respect to mutational stability; sparse evaluated programs are less stable than complete evaluated programs.     

Fitness approximation for bot evolution in genetic programming

Estimating the fitness value of individuals in an evolutionary algorithm in order to reduce the computational expense of actually calculating the fitness has been a classical pursuit of practitioners. One area which could benefit from progress in this endeavour is bot evolution, i.e. the evolution of non-playing characters in computer games. Because assigning a fitness value to a bot (or rather, the decision tree that controls its behaviour) requires playing the game, the process is very costly. In this work, we introduce two major contributions to speed up this process in the computer game Unreal Tournament 2004?. Firstly, a method for fitness value approximation in genetic programming which is based on the idea that individuals that behave in a similar fashion will have a similar fitness. Thus, similarity of individuals is taken at the performance level, in contrast to commonly employed approaches which are either based on similarity of genotypes or, less frequently, phenotypes. The approximation performs a weighted average of the fitness values of a number of individuals, attaching a confidence level which is based on similarity estimation. The latter is the second contribution of this work, namely a method for estimating the similarity between individuals. This involves carrying out a number of tests consisting of playing a ‘static’ version of the game (with fixed inputs) with the individuals whose similarity is under evaluation and comparing the results. Because the tests involve a limited version of the game, the computational expense of the similarity estimation plus that of the fitness approximation is much lower than that of directly calculating the fitness. The success of the fitness approximation by similarity estimation method for bot evolution in UT2K4 allows us to expect similar results in environments that share the same characteristics.     

Models for evaluating information updating intervals

Management scientists studying the decision-making process are in agreement that information and decision making are closely tied together. Unfortunately, improvements in information technology have not included methodologies which capture this interdependence. This paper is concerned with treating this interdependence in an aspect of systems design termed the “information updating interval.” Three evaluation models are presented, along with a special-case model for estimating related probability distributions.     

Assessing moderating effect in meta-analysis: a re-analysis of top management support studies and suggestions for researchers

Meta-analysis has been increasingly used as a knowledge cumulation tool by IS researchers. In recent years many meta-analysts have conducted moderator analyses in an attempt to develop and test theories. These studies suffer from several methodological problems and, as a result, may have contributed to rather than resolved inconsistent research findings. For example, a previous meta-analysis reports that task interdependence moderates the effect of top management support to render it a non-critical component in systems implementation projects when task interdependence is low. We show that this conclusion is the result of uncorrected measurement error and an erroneous application of a fixed effects regression analysis. We discuss other pitfalls in the detection and confirmation of moderators including the use of the Q statistic and significance tests. Our recommended approach is to break the sample into subgroups and compare their credibility and confidence intervals. This approach is illustrated in a re-analysis of the top management support literature. Our results indicate that top management support is important in both high and low task interdependence groups and in fact may be equally important in both groups. Guidelines are developed to help IS researchers properly conduct moderator analyses in future meta-analytic studies.     

Automating the analysis of variance of orthogonal designs

A new algorithm is presented which for the wide class of orthogonal designs is capable of deducing the appropriate analysis of variance from the design only. As a consequence the use of a model equation for specifying the analysis becomes dispensable. The proposed approach can simplify the analysis of complex models with iterative crossing and nesting of factors, where treatment factors have fixed and plot factors have random effects. An implementation is described and its use is illustrated with several examples.     

Generalized linear models with clustered data: Fixed and random effects models

The statistical analysis of mixed effects models for binary and count data is investigated. In the statistical computing environment R, there are a few packages that estimate models of this kind. The package lme4 is a de facto standard for mixed effects models. The package glmmML allows non-normal distributions in the specification of random intercepts. It also allows for the estimation of a fixed effects model, assuming that all cluster intercepts are distinct fixed parameters; moreover, a bootstrapping technique is implemented to replace asymptotic analysis. The random intercepts model is fitted using a maximum likelihood estimator with adaptive Gauss-Hermite and Laplace quadrature approximations of the likelihood function. The fixed effects model is fitted through a profiling approach, which is necessary when the number of clusters is large. In a simulation study, the two approaches are compared. The fixed effects model has severe bias when the mixed effects variance is positive and the number of clusters is large.     

Estimation in the probit normal model for binary outcomes using the SAEM algorithm

Generalized linear mixed models (GLMM) form a very general class of random effects models for discrete and continuous responses in the exponential family. They are useful in a variety of applications. The traditional likelihood approach for GLMM usually involves high dimensional integrations which are computationally intensive. In this work, we investigate the case of binary outcomes analyzed under a two stage probit normal model with random effects. First, it is shown how ML estimates of the fixed effects and variance components can be computed using a stochastic approximation of the EM algorithm (SAEM). The SAEM algorithm can be applied directly, or in conjunction with a parameter expansion version of EM to speed up the convergence. A procedure is also proposed to obtain REML estimates of variance components and REML-based estimates of fixed effects. Finally an application to a real data set involving a clinical trial is presented, in which these techniques are compared to other procedures (penalized quasi-likelihood, maximum likelihood, Bayesian inference) already available in classical softwares (SAS Glimmix, SAS Nlmixed, WinBUGS), as well as to a Monte Carlo EM (MCEM) algorithm.     

An approach to parameters estimation of a chromatography model using a clustering genetic algorithm based inverse model

Genetic algorithms are tools for searching in complex spaces and they have been used successfully in the system identification solution that is an inverse problem. Chromatography models are represented by systems of partial differential equations with non-linear parameters which are, in general, difficult to estimate many times. In this work a genetic algorithm is used to solve the inverse problem of parameters estimation in a model of protein adsorption by batch chromatography process. Each population individual represents a supposed condition to the direct solution of the partial differential equation system, so the computation of the fitness can be time consuming if the population is large. To avoid this difficulty, the implemented genetic algorithm divides the population into clusters, whose representatives are evaluated, while the fitness of the remaining individuals is calculated in function of their distances from the representatives. Simulation and practical studies illustrate the computational time saving of the proposed genetic algorithm and show that it is an effective solution method for this type of application.     

Putting more genetics into genetic algorithms

The majority of current genetic algorithms (GAs), while inspired by natural evolutionary systems, are seldom viewed as biologically plausible models. This is not a criticism of GAs, but rather a reflection of choices made regarding the level of abstraction at which biological mechanisms are modeled, and a reflection of the more engineering-oriented goals of the evolutionary computation community. Understanding better and reducing this gap between GAs and genetics has been a central issue in an interdisciplinary project whose goal is to build GA-based computational models of viral evolution. The result is a system called Virtual Virus (VIV). VIV incorporates a number of more biologically plausible mechanisms, including a more flexible genotype-to-phenotype mapping. In VIV the genes are independent of position, and genomes can vary in length and may contain noncoding regions, as well as duplicative or competing genes. Initial computational studies with VIV have already revealed several emergent phenomena of both biological and computational interest. In the absence of any penalty based on genome length, VIV develops individuals with long genomes and also performs more poorly (from a problem-solving viewpoint) than when a length penalty is used. With a fixed linear length penalty, genome length tends to increase dramatically in the early phases of evolution and then decrease to a level based on the mutation rate. The plateau genome length (i.e., the average length of individuals in the final population) generally increases in response to an increase in the base mutation rate. When VIV converges, there tend to be many copies of good alternative genes within the individuals. We observed many instances of switching between active and inactive genes during the entire evolutionary process. These observations support the conclusion that noncoding regions serve as scratch space in which VIV can explore alternative gene values. These results represent a positive step in understanding how GAs might exploit more of the power and flexibility of biological evolution while simultaneously providing better tools for understanding evolving biological systems.     

Designing Polymer Blends Using Neural Networks, Genetic Algorithms, and Markov Chains

In this paper we present a new technique to simulate polymer blends that overcomes the shortcomings in polymer system modeling. This method has an inherent advantage in that the vast existing information on polymer systems forms a critical part in the design process. The stages in the design begin with selecting potential candidates for blending using Neural Networks. Generally the parent polymers of the blend need to have certain properties and if the blend is miscible then it will reflect the properties of the parents. Once this step is finished the entire problem is encoded into a genetic algorithm using various models as fitness functions. We select the lattice fluid model of Sanchez and Lacombe (J. Polym. Sci. Polym. Lett. Ed., vol. 15, p. 71, 1977), which allows for a compressible lattice. After reaching a steady-state with the genetic algorithm we transform the now stochastic problem that satisfies detailed balance and the condition of ergodicity to a Markov Chain of states. This is done by first creating a transition matrix, and then using it on the incidence vector obtained from the final populations of the genetic algorithm. The resulting vector is converted back into a population of individuals that can be searched to find the individuals with the best fitness values. A high degree of convergence not seen using the genetic algorithm alone is obtained. We check this method with known systems that are miscible and then use it to predict miscibility on several unknown systems.     

Novel adaptive genetic algorithm sample consensus

Random Sample Consensus (RANSAC) is a successful algorithm in model fitting applications when there are numerous outliers within the dataset. Achieving a proper model is guaranteed through the pure exploration strategy of RANSAC. However, finding the optimum result requires exploitation. Genetic Algorithm Sample Consensus (GASAC) is an evolutionary paradigm which adds the exploitation capability to RANSAC. Although GASAC improves the results of RANSAC, it has a fixed strategy for balancing between exploration and exploitation. In this paper, a new paradigm is proposed based on genetic algorithms using an adaptive strategy. We propose an adaptive genetic operator to select the proper number of high fitness individuals as parents and mutate the rest. This operator can adjust the ratio of exploration vs. exploitation phases according to the amount of outliers. Also, a learning method is proposed for the mutation operator to gradually learn which gene is the best replacement for the mutated gene. This operator guides the exploration phase towards good solution areas and therefore produces better individuals for further exploitation. The proposed method is extensively evaluated in two sets of experiments. In all tests, our method outperformed the other methods in terms of both the number of inliers found and the speed of the algorithm.     

自适应二次变异差分进化算法

提出一种基于群体适应度方差自适应二次变异的差分进化算法．该算法在运行过程中根据群体适应度方差的大小,增加一种新的变异算子对最优个体和部分其他个体同时进行变异操作,以提高种群多样性,增强差分进化算法跳出局部最优解的能力．对几种典型Benchmarks函数进行了测试,实验结果表明,该方法能有效避免早熟收敛,显著提高算法的全局搜索能力。     

INTERDEPENDENCE AND INFORMATION SHARING IN SOCIAL SYSTEMS

Our world is becoming increasingly interdependent; events in one part of the globe combine with events in other parts to strike a host of social systems. As a result of this combination of distant events, systems find that the time interval between one disturbance and the next is significantly shortened. On the other hand, interdependence among social systems results in a longer response time, which is, the period of time between a system being disturbed and its regaining equilibrium. Moreover, response time expands exponentially as interdependence increases, leaving our modern social systems in a very tight predicament. The sharing of decision-making information among systems can significantly reduce the response time and dissipate the exponential pattern that emerges as a result of interdependence.     

Efficient search for robust solutions by means of evolutionary algorithms and fitness approximation

For many real-world optimization problems, the robustness of a solution is of great importance in addition to the solution's quality. By robustness, we mean that small deviations from the original design, e.g., due to manufacturing tolerances, should be tolerated without a severe loss of quality. One way to achieve that goal is to evaluate each solution under a number of different scenarios and use the average solution quality as fitness. However, this approach is often impractical, because the cost for evaluating each individual several times is unacceptable. In this paper, we present a new and efficient approach to estimating a solution's expected quality and variance. We propose to construct local approximate models of the fitness function and then use these approximate models to estimate expected fitness and variance. Based on a variety of test functions, we demonstrate empirically that our approach significantly outperforms the implicit averaging approach, as well as the explicit averaging approaches using existing estimation techniques reported in the literature.     

The impact of partnership attributes on EDI implementation success

The growing interest in the use of electronic data interchange (EDI) to revolutionize the way in which business conducts trading activities leads to research questions concerning the characteristics of electronic partnerships that promote implementation success. The research model in this study shows that partnership attributes affect EDI implementation, which has three dimensions: integration, utilization, and diversity, and performance. Partnership attributes encompass partner trust, interdependence, and commitment. Results show that partner trust, interdependence, and commitment affect the extent to which companies undertake EDI integration and increase the percentage of EDI exchange and performance. This suggests that companies contemplating EDI should set up compatible business processes and a high-quality partner relationship for implementation and performance.     

An application of a GA with Markov network surrogate to feature selection

Surrogate models of fitness have been presented as a way of reducing the number of fitness evaluations required by evolutionary algorithms. This is of particular interest with expensive fitness functions where the time taken for building the model is outweighed by the savings of using fewer function evaluations. In this article, we show how a Markov network model can be used as a surrogate fitness function for a genetic algorithm in a new algorithm called Markov Fitness Model Genetic Algorithm (MFM-GA). We thoroughly investigate its application to a fitness function for feature selection in Case-Based Reasoning (CBR), using a range of standard benchmarks from the CBR community. This fitness function requires considerable computation time to evaluate and we show that using the surrogate offers a significant decrease in total run-time compared to a GA using the true fitness function. This comes at the cost of a reduction in the global best fitness found. We demonstrate that the quality of the solutions obtained by MFM-GA improves significantly with model rebuilding. Comparisons with a classic GA, a GA using fitness inheritance and a selection of filter selection methods for CBR shows that MFM-GA provides a good trade-off between fitness quality and run-time.     

Dynamics of two-element populations in the space of population states

An analysis of population dynamics in the space of population states is presented. The simplest case of the phenotypic evolution-a population consisting of two individuals with one real-valued trait, evolving under proportional selection and mutation with an underlying normal distribution-is considered. The focus is on the trajectories of the expected population state values generating a discrete dynamical system. The system models the expected asymptotic behavior of the evolutionary process. The analysis and the simulation results shed light on the dynamics of approaching evolutionary equilibria. The effect of two-speed convergence is observed: 1) initially fast convergence toward an approximately homogenous population and then 2) a slow drift of the population toward optima. The system's fixed points and their stability are determined. Periodic and chaotic behaviors are observed for some fitness functions.