Pavel Krömer

Genetically Evolved Fuzzy Predictor for Photovoltaic Power Output Estimation

Fuzzy sets and fuzzy logic can be used for efficient data mining, classification, and value prediction. We propose a genetically evolved fuzzy predictor to estimate the output of a Photovoltaic Power Plant. Photovoltaic Power Plants (PVPPs) are classified as power energy sources with unstable supply of electrical energy. It is necessary to back up power energy from PVPPs for stable electric network operation. An optimal value of back up power can be set with reliable prediction models and significantly contribute to the robustness of the electric network and therefore help in the building of intelligent power grids.

Implementing GP on Optimizing both Boolean and Extended Boolean Queries in IR and Fuzzy IR systems with Respect to the Users Profiles

Rapidly growing amount of the data available on World Wide Web (WWW) as the ultimate collection of public accessible text documents complicates the task of acquiring usable and relevant information from raw data. Several methods of improvement and optimization of information retrieval (IR) systems providing user interaction with WWW documents have been introduced and discussed. In this paper, theory and application of genetic programming (GP) as an optimization method of modern user oriented IR and fuzzy IR (FIR) systems optimization of search queries is presented.

Matrix Factorization Approach for Feature Deduction and Design of Intrusion Detection Systems

Current Intrusion Detection Systems (IDS) examine all data features to detect intrusion or misuse patterns. Some of the features may be redundant or contribute little (if anything) to the detection process. The purpose of this research is to identify important input features in building an IDS that is computationally efficient and effective. This paper propose a novel matrix factorization approach for feature deduction and design of intrusion detection systems. Experiment results indicate that the proposed method is efficient.

Learning the Classification of Traffic Accident Types

This paper presents an application of evolutionary fuzzy classifier design to a road accident data analysis. A fuzzy classifier evolved by the genetic programming was used to learn the labeling of data in a real world road accident data set. The symbolic classifier was inspected in order to select important features and the relations among them. Selected features provide a feedback for traffic management authorities that can exploit the knowledge to improve road safety and mitigate the severity of traffic accidents.

Evolutionary Approaches to Linear Ordering Problem

Linear ordering problem (LOP) is a well know optimization problem attractive for its complexity (it is a NP hard problem), rich collection of testing data and variety of real world applications. In this paper, we investigate the usage and performance of two variants of genetic algorithms - mutation only genetic algorithms and higher level chromosome genetic algorithms - on the linear ordering problem. Both methods are tested and evaluated on a collection of real world and artificial LOP instances.

Review of nature-inspired methods for wake-up scheduling in wireless sensor networks

Abstract Over the last few decades, algorithms inspired by nature have matured into a widely used class of computing methods. They have shown the ability to adjust to variety of conditions, and have been frequently employed for solving complex, real-world optimization problems. They are especially suitable for problems that require adaptation, and that involve optimization of complex, distributed systems, operating in dynamic environments. Among other application domains, nature-inspired methods have been extensively used in the areas of networking in general, and wireless sensor networks in particular. Energy management and network lifetime optimization are two great research and implementation challenges for wireless sensor networks. Duty cycle management, synchronization, and wake-up scheduling are complementary approaches that facilitate this complex optimization process. This review focuses on the intersection of nature-inspired computing and wake-up scheduling algorithms for wireless sensor networks. It describes the state-of-the-art in these fields and provides an up-to-date review of the most recent developments in this interdisciplinary domain. It discusses the motivation for using nature-inspired methods for wake-up scheduling, and presents related open issues and research challenges.

On the Implementation of Boolean Matrix Factorization

Matrix factorization or factor analysis is an important task helpful in the analysis of high dimensional real world data. There are several well known methods and algorithms for factorization of real data but many application areas including information retrieval, pattern recognition and data mining require processing of binary rather than real data. Unfortunately, the methods used for real matrix factorization fail in the latter case. In this paper we introduce the background of the task as well as genetic algorithm based solver and present results obtained from computer experiments.

On Genetic Algorithms for Boolean Matrix Factorization

Matrix factorization or factor analysis is an important task in the analysis of high dimensional real world data. There are several well known methods and algorithms for factorization of real data but they are rather inefficient when dealing with binary information. In this paper we introduce background and initial version of genetic algorithm for binary matrix factorization.

Harvesting-aware control of wireless sensor nodes using fuzzy logic and differential evolution

Environmental monitoring systems are expected to provide detailed data about their surroundings in a timely fashion. Frequent data sampling is desired to obtain data with high temporal resolution, and frequent wireless transmissions are needed in order to get the data to end users without unnecessary delays. Energy harvesting systems depend, to a large extent, on the energy available in the environment. They need sophisticated control approaches in order to consolidate these requirements with the need for autonomous operation and longevity of the system. This study proposes a novel approach for dynamic control of energy harvesting nodes. It uses fuzzy logic controller evolved by differential evolution. The evolved control system is evaluated in a series of computational experiments that show very promising results.

A brief survey of advances in Particle Swarm Optimization on Graphic Processing Units

In the last few years, the Graphic Processing Units (GPUs) emerged as an exciting new hardware environment available for a truly parallel implementation and execution of Nature and Bio-inspired Algorithms. In contrast to common multicore CPUs that contain up to tens of independent cores, the GPUs represent a massively parallel single-instruction multiple-data (SIMD) devices that can nowadays reach peak performance of hundreds and thousands of giga FLOPS (floating-point operations per second). Nature and Bio-inspired Algorithms often adopt populational problem solving approaches and implement parallel optimization strategies in which group or groups of candidate solutions search for optimal solutions. Swarm Intelligence and Particle Swarm Optimization (PSO) in particular can be seen as multiagent methods in which the interaction of simple independent agents yields intelligent collective behavior. Such algorithms especially fit to the architecture of the GPUs. This survey provides a brief overview of the latest state-of-the-art research on the design, implementation, and applications of PSO-based methods on the GPUs.