Oliver Kramer

Evolutionary self-adaptation: a survey of operators and strategy parameters

The success of evolutionary search depends on adequate parameter settings. Ill conditioned strategy parameters decrease the success probabilities of genetic operators. Proper settings may change during the optimization process. The question arises if adequate settings can be found automatically during the optimization process. Evolution strategies gave an answer to the online parameter control problem decades ago: self-adaptation. Self-adaptation is the implicit search in the space of strategy parameters. The self-adaptive control of mutation strengths in evolution strategies turned out to be exceptionally successful. Nevertheless, for years self-adaptation has not achieved the attention it deserves. This paper is a survey of self-adaptive parameter control in evolutionary computation. It classifies self-adaptation in the taxonomy of parameter setting techniques, gives an overview of automatic online-controllable evolutionary operators and provides a coherent view on search techniques in the space of strategy parameters. Beyer and Sendhoff’s covariance matrix self-adaptation evolution strategy is reviewed as a successful example for self-adaptation and exemplarily tested for various concepts that are discussed.

Iterated local search with Powell's method: a memetic algorithm for continuous global optimization

In combinatorial solution spaces Iterated Local Search (ILS) turns out to be exceptionally successful. The question arises: is ILS also capable of improving the optimization process in continuous solution spaces? To demonstrate that hybridization leads to powerful techniques in continuous domains, we introduce a hybrid meta-heuristic that integrates Powell’s direct search method. It combines direct search with elements from population based evolutionary optimization. The approach is analyzed experimentally on a set of well known test problems and compared to a state-of-the-art technique, i.e., a restart variant of the Covariance Matrix Adaptation Evolution Strategy with increasing population sizes (G-CMA-ES). It turns out that the population-based Powell-ILS is competitive to the CMA-ES, in some cases even significantly faster and behaves more robust than the pure strategy of Powell in multimodal fitness landscapes. Further experiments on the perturbation mechanism, population sizes, and problems with noise complete the analysis of the hybrid methodology and lead to parameter recommendations.

Wind energy prediction and monitoring with neural computation

Wind energy has an important part to play as renewable energy resource in a sustainable world. For a reliable integration of wind energy high-dimensional wind time-series have to be analyzed. Fault analysis and prediction are an important aspect in this context. The objective of this work is to show how methods from neural computation can serve as forecasting and monitoring techniques, contributing to a successful integration of wind into sustainable and smart energy grids. We will employ support vector regression as prediction method for wind energy time-series. Furthermore, we will use dimension reduction techniques like self-organizing maps for monitoring of high-dimensional wind time-series. The methods are briefly introduced, related work is presented, and experimental case studies are exemplarily described. The experimental parts are based on real wind energy time-series data from the National Renewable Energy Laboratory (NREL) western wind resource data set.

Statistical Learning for Short-Term Photovoltaic Power Predictions

A reliable prediction of photovoltaic (PV) power plays an important part as basis for operation and management strategies for a efficient and economical integration into the power grid. Due to changing weather conditions, e.g., clouds and fog, a precise forecast in a few hour range can be a difficult task. The growing IT infrastructure allows a fine screening of PV power. On the basis of big data sets of PV measurements, we apply methods from statistical learning for one- to six-hour ahead predictions based on data with hourly resolution. In this work, we employ nearest neighbor regression and support vector regression for PV power predictions based on measurements and numerical weather predictions. We put an emphasis on the analysis of feature combinations based on these two data sources. After optimizing the settings and comparing the employed statistical learning models, we build a hybrid predictor that uses forecasts of both employed models.

Surrogate constraint functions for CMA evolution strategies

Many practical optimization problems are constrained black boxes. Covariance Matrix Adaptation Evolution Strategies (CMA-ES) belong to the most successful black box optimization methods. Up to now no sophisticated constraint handling method for Covariance Matrix Adaptation optimizers has been proposed. In our novel approach we learn a meta-model of the constraint function and use this surrogate model to adapt the covariance matrix during the search at the vicinity of the constraint boundary. The meta-model can be used for various purposes, i.e. rotation of the mutation ellipsoid, checking the feasibility of candidate solutions or repairing infeasible mutations by projecting them onto the constraint surrogate function. Experimental results show the potentials of the proposed approach.

Precise Wind Power Prediction with SVM Ensemble Regression

In this work, we propose the use of support vector regression ensembles for wind power prediction. Ensemble methods often yield better classification and regression accuracy than classical machine learning algorithms and reduce the computational cost. In the field of wind power generation, the integration into the smart grid is only possible with a precise forecast computed in a reasonable time. Our support vector regression ensemble approach uses bootstrap aggregating (bagging), which can easily be parallelized. A set of weak predictors is trained and then combined to an ensemble by aggregating the predictions. We investigate how to choose and train the individual predictors and how to weight them best in the prediction ensemble. In a comprehensive experimental analysis, we show that our SVR ensemble approach renders significantly better forecast results than state-of-the-art predictors.

On the hybridization of SMS-EMOA and local search for continuous multiobjective optimization

In the recent past, hybrid metaheuristics became famous as successful optimization methods. The motivation for the hybridization is a notion of combining the best of two worlds: evolutionary black box optimization and local search. Successful hybridizations in large combinatorial solution spaces motivate to transfer the idea of combining the two worlds to continuous domains as well. The question arises: Can local search also improve the convergence to the Pareto front in continuous multiobjective solutions spaces? We introduce a relay and a concurrent hybridization of the successful multiobjective optimizer SMS-EMOA and local optimization methods like Hooke & Jeeves and the Newton method. The concurrent approach is based on a parameterized probability function to control the local search. Experimental analyses on academic test functions show increased convergence speed as well as improved accuracy of the solution set of the new hybridizations.

Classification Cascades of Overlapping Feature Ensembles for Energy Time Series Data

The classification of high-dimensional time series data can be a challenging task due to the curse-of-dimensionality problem. The classification of time series is relevant in various applications, e.g., in the task of learning meta-models of feasible schedules for flexible components in the energy domain. In this paper, we introduce a classification approach that employs a cascade of classifiers based on features of overlapping time series steps. To evaluate the feasibility of the whole time series, each overlapping pattern is evaluated and the results are aggregated. We apply the approach to the problem of combined heat and power plant operation schedules and an artificial similarly structured data set. We identify conditions under which the cascade approach shows better results than a classic One-Class-SVM.

Premature Convergence in Constrained Continuous Search Spaces

The optimum of numerical problems quite often lies on the constraint boundary or even in a vertex of the feasible search space. In such cases the evolutionary algorithm (EA) frequently suffers from premature convergence because of a low success probability near the constraint boundaries. We analyze premature fitness stagnation and the success rates experimentally for an EA using self-adaptive step size control. For a (1+1)-EA with a Rechenberg-like step control mechanism we prove premature step size reduction at the constraint boundary. The proof is based on a success rate analysis considering a simplified mutation distribution model. From the success rates and the possible state transitions, the expected step size change can be derived at each step. We validate the theoretical model with an experimental analysis.

On Evolutionary Approaches to Wind Turbine Placement with Geo-Constraints

Wind turbine placement, i.e., the geographical planning of wind turbine locations, is an important first step to an efficient integration of wind energy. The turbine placement problem becomes a difficult optimization problem due to varying wind distributions at different locations and due to the mutual interference in the wind field known as wake effect. Artificial and environmental geological constraints make the optimization problem even more difficult to solve. In our paper, we focus on the evolutionary turbine placement based on an enhanced wake effect model fed with real-world wind distributions. We model geo-constraints with real-world data from OpenStreetMap. Besides the realistic modeling of wakes and geo-constraints, the focus of the paper is on the comparison of various evolutionary optimization approaches. We propose four variants of evolution strategies with turbine-oriented mutation operators and compare to state-of-the-art optimizers like the CMA-ES in a detailed experimental analysis on three benchmark scenarios.