An efficient multi-objective evolutionary fuzzy system for regression problems

During the last years, multi-objective evolutionary algorithms (MOEAs) have been extensively employed as optimization tools for generating fuzzy rule-based systems (FRBSs) with different trade-offs between accuracy and interpretability from data. Since the size of the search space and the computational cost of the fitness evaluation depend on the number of input variables and instances, respectively, managing high-dimensional and large datasets is a critical issue.In this paper, we focus on MOEAs applied to learn concurrently the rule base and the data base of Mamdani FRBSs and propose to tackle the issue by exploiting the synergy between two different techniques. The first technique is based on a novel method which reduces the search space by learning rules not from scratch, but rather from a heuristically generated rule base. The second technique performs an instance selection by exploiting a co-evolutionary approach where cyclically a genetic algorithm evolves a reduced training set which is used in the evolution of the MOEA.The effectiveness of the synergy has been tested on twelve datasets. Using non-parametric statistical tests we show that, although achieving statistically equivalent solutions, the adoption of this synergy allows saving up to 97.38% of the execution time with respect to a state-of-the-art multi-objective evolutionary approach which learns rules from scratch.     

Integrating memetic search into the BioHEL evolutionary learning system for large-scale datasets

Local search methods are widely used to improve the performance of evolutionary computation algorithms in all kinds of domains. Employing advanced and efficient exploration mechanisms becomes crucial in complex and very large (in terms of search space) problems, such as when employing evolutionary algorithms to large-scale data mining tasks. Recently, the GAssist Pittsburgh evolutionary learning system was extended with memetic operators for discrete representations that use information from the supervised learning process to heuristically edit classification rules and rule sets. In this paper we first adapt some of these operators to BioHEL, a different evolutionary learning system applying the iterative learning approach, and afterwards propose versions of these operators designed for continuous attributes and for dealing with noise. The performance of all these operators and their combination is extensively evaluated on a broad range of synthetic large-scale datasets to identify the settings that present the best balance between efficiency and accuracy. Finally, the identified best configurations are compared with other classes of machine learning methods on both synthetic and real-world large-scale datasets and show very competent performance.     

Handling equality constraints with agent-based memetic algorithms

In addition to inequality constraints, many mathematical models require equality constraints to represent the practical problems appropriately. The existence of equality constraints reduces the size of the feasible space significantly, which makes it difficult to locate feasible and optimal solutions. This paper presents a new equality constraint handling technique which enhances the performance of an agent-based evolutionary algorithm in solving constrained optimization problems with equality constraints. The technique is basically used as an agent learning process in the agent-based evolutionary algorithm. The performance of the proposed algorithm is tested on a set of well-known benchmark problems including seven new problems. The experimental results confirm the improved performance of the proposed technique.     

Hybrid evolutionary algorithm with Hermite radial basis function interpolants for computationally expensive adjoint solvers

In this paper, we present an evolutionary algorithm hybridized with a gradient-based optimization technique in the spirit of Lamarckian learning for efficient design optimization. In order to expedite gradient search, we employ local surrogate models that approximate the outputs of a computationally expensive Euler solver. Our focus is on the case when an adjoint Euler solver is available for efficiently computing the sensitivities of the outputs with respect to the design variables. We propose the idea of using Hermite interpolation to construct gradient-enhanced radial basis function networks that incorporate sensitivity data provided by the adjoint Euler solver. Further, we conduct local search using a trust-region framework that interleaves gradient-enhanced surrogate models with the computationally expensive adjoint Euler solver. This ensures that the present hybrid evolutionary algorithm inherits the convergence properties of the classical trust-region approach. We present numerical results for airfoil aerodynamic design optimization problems to show that the proposed algorithm converges to good designs on a limited computational budget.     

Learning genetic algorithm parameters using hidden Markov models

Genetic algorithms (GAs) are routinely used to search problem spaces of interest. A lesser known but growing group of applications of GAs is the modeling of so-called “evolutionary processes”, for example, organizational learning and group decision-making. Given such an application, we show it is possible to compute the likely GA parameter settings given observed populations of such an evolutionary process. We examine the parameter estimation process using estimation procedures for learning hidden Markov models, with mathematical models that exactly capture expected GA behavior. We then explore the sampling distributions relevant to this estimation problem using an experimental approach.     

Combining boosting and evolutionary algorithms for learning of fuzzy classification rules

This paper presents a novel boosting algorithm for genetic learning of fuzzy classification rules. The method is based on the iterative rule learning approach to fuzzy rule base system design. The fuzzy rule base is generated in an incremental fashion, in that the evolutionary algorithm optimizes one fuzzy classifier rule at a time. The boosting mechanism reduces the weight of those training instances that are classified correctly by the new rule. Therefore, the next rule generation cycle focuses on fuzzy rules that account for the currently uncovered or misclassified instances. The weight of a fuzzy rule reflects the relative strength the boosting algorithm assigns to the rule class when it aggregates the casted votes. The approach is compared with other classification algorithms for a number problem sets from the UCI repository.     

An estimation of distribution algorithm with intelligent local search for rule-based nurse rostering

This paper proposes a new memetic evolutionary algorithm to achieve explicit learning in rule-based nurse rostering, which involves applying a set of heuristic rules for each nurse's assignment. The main framework of the algorithm is an estimation of distribution algorithm, in which an ant-miner methodology improves the individual solutions produced in each generation. Unlike our previous work (where learning is implicit), the learning in the memetic estimation of distribution algorithm is explicit, that is, we are able to identify building blocks directly. The overall approach learns by building a probabilistic model, that is, an estimation of the probability distribution of individual nurse–rule pairs that are used to construct schedules. The local search processor (ie the ant-miner) reinforces nurse–rule pairs that receive higher rewards. A challenging real-world nurse rostering problem is used as the test problem. Computational results show that the proposed approach outperforms most existing approaches. It is suggested that the learning methodologies suggested in this paper may be applied to other scheduling problems where schedules are built systematically according to specific rules.     

Autonomous vehicle navigation using evolutionary reinforcement learning

Reinforcement learning schemes perform direct on-line search in control space. This makes them appropriate for modifying control rules to obtain improvements in the performance of a system. The effectiveness of a reinforcement learning strategy is studied here through the training of a learning classifier system (LCS) that controls the movement of an autonomous vehicle in simulated paths including left and right turns. The LCS comprises a set of condition-action rules (classifiers) that compete to control the system and evolve by means of a genetic algorithm (GA). Evolution and operation of classifiers depend upon an appropriate credit assignment mechanism based on reinforcement learning. Different design options and the role of various parameters have been investigated experimentally. The performance of vehicle movement under the proposed evolutionary approach is superior compared with that of other (neural) approaches based on reinforcement learning that have been applied previously to the same benchmark problem.     

Neuroevolution strategies for episodic reinforcement learning

Because of their convincing performance, there is a growing interest in using evolutionary algorithms for reinforcement learning. We propose learning of neural network policies by the covariance matrix adaptation evolution strategy (CMA-ES), a randomized variable-metric search algorithm for continuous optimization. We argue that this approach, which we refer to as CMA Neuroevolution Strategy (CMA-NeuroES), is ideally suited for reinforcement learning, in particular because it is based on ranking policies (and therefore robust against noise), efficiently detects correlations between parameters, and infers a search direction from scalar reinforcement signals. We evaluate the CMA-NeuroES on five different (Markovian and non-Markovian) variants of the common pole balancing problem. The results are compared to those described in a recent study covering several RL algorithms, and the CMA-NeuroES shows the overall best performance.     

An interactive evolutionary multi-objective optimization algorithm with a limited number of decision maker calls

This paper presents a preference-based method to handle optimization problems with multiple objectives. With an increase in the number of objectives the computational cost in solving a multi-objective optimization problem rises exponentially, and it becomes increasingly difficult for evolutionary multi-objective techniques to produce the entire Pareto-optimal front. In this paper, an evolutionary multi-objective procedure is combined with preference information from the decision maker during the intermediate stages of the algorithm leading to the most preferred point. The proposed approach is different from the existing approaches, as it tries to find the most preferred point with a limited budget of decision maker calls. In this paper, we incorporate the idea into a progressively interactive technique based on polyhedral cones. The idea is also tested on another progressively interactive approach based on value functions. Results are provided on two to five-objective unconstrained as well as constrained test problems.     

A multivariate evolutionary generalised linear model framework with adaptive estimation for claims reserving

In this paper, we develop a multivariate evolutionary generalised linear model (GLM) framework for claims reserving, which allows for dynamic features of claims activity in conjunction with dependency across business lines to accurately assess claims reserves. We extend the traditional GLM reserving framework on two fronts: GLM fixed factors are allowed to evolve in a recursive manner, and dependence is incorporated in the specification of these factors using a common shock approach.We consider factors that evolve across accident years in conjunction with factors that evolve across calendar years. This two-dimensional evolution of factors is unconventional as a traditional evolutionary model typically considers the evolution in one single time dimension. This creates challenges for the estimation process, which we tackle in this paper. We develop the formulation of a particle filtering algorithm with parameter learning procedure. This is an adaptive estimation approach which updates evolving factors of the framework recursively over time.We implement and illustrate our model with a simulated data set, as well as a set of real data from a Canadian insurer.     

Evolutionary learning of small networks

Results are presented of a simulation that mimics an evolutionary learning process for small networks. Special features of these networks include a high recurrency, a transition function which decreases for large input activities, and the absence of tunable weights attached to the lines—the line is either there (weight 1) or absent (weight 0). It is remarkable that these simple systems already exhibit a complex learning behavior and a phenomenon of punctuated equilibrium in the evolutionary process. These findings should be of interest for both, the general understanding of evolutionary dynamics and, more specifically, the understanding of the role of recurrence in combination with nonmonotic response patterns for learning processes. © 2007 Wiley Periodicals, Inc. Complexity, 2008     

Genetic learning of accurate and compact fuzzy rule based systems based on the 2-tuples linguistic representation

One of the problems that focus the research in the linguistic fuzzy modeling area is the trade-off between interpretability and accuracy. To deal with this problem, different approaches can be found in the literature. Recently, a new linguistic rule representation model was presented to perform a genetic lateral tuning of membership functions. It is based on the linguistic 2-tuples representation that allows the lateral displacement of a label considering an unique parameter. This way to work involves a reduction of the search space that eases the derivation of optimal models and therefore, improves the mentioned trade-off.Based on the 2-tuples rule representation, this work proposes a new method to obtain linguistic fuzzy systems by means of an evolutionary learning of the data base a priori (number of labels and lateral displacements) and a simple rule generation method to quickly learn the associated rule base. Since this rule generation method is run from each data base definition generated by the evolutionary algorithm, its selection is an important aspect. In this work, we also propose two new ad hoc data-driven rule generation methods, analyzing the influence of them and other rule generation methods in the proposed learning approach. The developed algorithms will be tested considering two different real-world problems.     

A predictive gradient strategy for multiobjective evolutionary algorithms in a fast changing environment

An essential feature of a dynamic multiobjective evolutionary algorithm (MOEA) is to converge quickly to the Pareto-optimal Set before it changes. In cases where the behavior of the dynamic problem follows a certain trend, convergence can be accelerated by anticipating the characteristics of future changes in the problem. A prediction model is usually used to exploit past information and estimate the location of the new Pareto-optimal Set. In this work, we propose the novel approach of tracking and predicting the changes in the location of the Pareto Set in order to minimize the effects of a landscape change. The predicted direction and magnitude of the next change, known as the predictive gradient, is estimated based on the history of previously discovered solutions using a weighted average approach. Solutions updated with the predictive gradient will remain in the vicinity of the new Pareto-optimal Set and help the rest of the population to converge. The prediction strategy is simple to implement, making it suitable for fast-changing problems. In addition, a new memory technique is introduced to exploit any periodicity in the dynamic problem. The memory technique selects only the more promising stored solutions for retrieval in order to reduce the number of evaluations used. Both techniques are incorporated into a variant of the multi-objective evolutionary gradient search (MO-EGS) and two other MOEAs for dynamic optimization and results indicate that they are effective at improving performance on several dynamic multiobjective test problems.     

单纯形调优法的收敛性质

在非线性规划中,单纯形调优法是一种可信用的算法,然而却缺乏理论分析,本文对单纯形调优法的理论进行了一些研究,所考虑的方法类似于Spendley,Hext,Himsworth的正规单纯形调优法,但采用了不同的反映条件,其中带有某种下山门槛或具有三点下降形式,这里的基本思想是证明:上述单纯形调优法是定步长下山法的特殊情形,所以,此研究紧密联系着作者关于定步长下山法收敛定理的工作。     

Integrated exact,hybrid and metaheuristic learning methods for confidentiality protection

A vital task facing government agencies and commercial organizations that report data is to represent the data in a meaningful way and simultaneously to protect the confidentiality of critical components of this data. The challenge is to organize and disseminate data in a form that prevents such critical components from being inferred by groups bent on corporate espionage, to gain competitive advantages, or having a desire to penetrate the security of the information underlying the data. Controlled tabular adjustment is a recently developed approach for protecting sensitive information by imposing a special form of statistical disclosure limitation on tabular data. The underlying model gives rise to a mixed integer linear programming problem involving both continuous and discrete (zero-one) variables. We develop stratified ordered (s-ordered) heuristics and a new meta-heuristic learning approach for solving this model, and compare their performance to previous heuristics and to an exact algorithm embodied in the state-of-the-art ILOG- CPLEX software. Our new approaches are based on partitioning the problem into its discrete and continuous components, first creating an s-ordered heuristic that reduces the number of binary variables through a grouping procedure that combines an exact mathematical programming model with constructive heuristics. To gain further advantages we then replace the mathematical programming model with an evolutionary scatter search approach that makes it possible to extend the method to large problems with over 9000 entries. Finally, we introduce a new metaheuristic learning method that significantly improves the quality of solutions obtained.     

Hybrid immune algorithm with Lamarckian local search for multi-objective optimization

Lamarckian learning has been introduced into evolutionary computation as local search mechanism. The relevant research topic, memetic computation, has received significant amount of interests. In this study, a novel Lamarckian learning strategy is designed for improving the Nondominated Neighbor Immune Algorithm, a novel hybrid multi-objective optimization algorithm, Multi-objective Lamarckian Immune Algorithm (MLIA), is proposed. The Lamarckian learning performs a greedy search which proceeds towards the goal along the direction obtained by Tchebycheff approach and generates the improved progenies or improved decision vectors, so single individual will be optimized locally and the newcomers yield an enhanced exploitation around the nondominated individuals in less-crowded regions of the current trade-off front. Simulation results based on twelve benchmark problems show that MLIA outperforms the original immune algorithm and NSGA-II in approximating Pareto-optimal front in most of the test problems. When compared with the state of the art algorithm MOEA/D, MLIA shows better performance in terms of the coverage of two sets metric, although it is laggard in the hypervolume metric.     

Memetic informed evolutionary optimization via data mining

This paper proposes a novel Informed Evolutionary algorithm (InEA) which implements the idea of learning with a generation. An association rule miner is used to identify the norm of a population. Subsequently, a knowledge based mutation operator is used to help guide the search of the evolutionary optimizer. The approach breaks away from the current practice of treating the optimization and analysis process as two independent processes. It shows how a rule mining module can be used to mine knowledge and hybridized into EA to improve the performance of the optimizer. The proposed memetic algorithm is examined via various benchmarks problems, and the simulation results show that InEA is competitive as compared to existing approaches in literature.     

A population global optimization algorithm to solve the image alignment problem in electron crystallography

Knowledge of the structure of biological specimens is critical to understanding their function. Electron crystallography is an electron microscopy (EM) approach that derives the 3D structure of specimens at high-resolution, even at atomic detail. Prior to the tomographic reconstruction, the images taken from the microscope have to be properly aligned. Traditional alignment methods in electron crystallography are based on a phase residual function to be minimized by inefficient exhaustive search procedures. This work addresses this minimization problem from an evolutionary perspective. Universal Evolutionary Global Optimizer (UEGO), an evolutionary multimodal optimization algorithm, has been applied and evaluated for the task of image alignment in this field. UEGO has turned out to be a promising technique alternative to the standard methodology. The alignments found out by UEGO show high levels of accuracy, while reducing the number of function evaluations by a significant factor with respect to the standard method.     

Numerical solution of the time-dependent incompressible Navier–Stokes equations by piecewise linear finite elements

In this paper we present a new method to solve the 2D generalized Stokes problem in terms of the stream function and the vorticity. Such problem results, for instance, from the discretization of the evolutionary Stokes system. The difficulty arising from the lack of the boundary conditions for the vorticity is overcome by means of a suitable technique for uncoupling both variables. In order to apply the above technique to the Navier–Stokes equations we linearize the advective term in the vorticity transport equation as described in the development of the paper. We illustrate the good performance of our approach by means of numerical results, obtained for benchmark driven cavity problem solved with classical piecewise linear finite element.