Daniel Kostrzewa

The Comparison of an Adapted Evolutionary Algorithm with the Invasive Weed Optimization Algorithm Based on the Problem of Predetermining the Progress of Distributed Data Merging Process

The goal of the project was to adapt the idea of the Invasive Weed Optimization (IWO) algorithm to the problem of predetermining the progress of distributed data merging process and to compare the results of the conducted experiments with analogical outcomes produced by the evolutionary algorithm. The main differences between both compared algorithms constituted by operators used for transformation of individuals and for creation of a new population were taken into consideration during the implementation of the IWO algorithm. The construction of an environment for experimental research made it possible to carry out a set of tests to explore the characteristics of the tested algorithms. The results of the conducted experiments formed the main topic of analysis.

The Expanded Invasive Weed Optimization Metaheuristic for Solving Continuous and Discrete Optimization Problems

This paper introduces an expanded version of the Invasive Weed Optimization algorithm (exIWO) distinguished by the hybrid strategy of the search space exploration proposed by the authors. The algorithm is evaluated by solving three well-known optimization problems: minimization of numerical functions, feature selection, and the Mona Lisa TSP Challenge as one of the instances of the traveling salesman problem. The achieved results are compared with analogous outcomes produced by other optimization methods reported in the literature.

The modified IWO algorithm for optimization of numerical functions

The Invasive Weed Optimization algorithm (IWO) is an optimization method inspired by dynamic growth of weeds colony. The authors of the present paper have modified the IWO algorithm introducing a hybrid strategy of the search space exploration. The goal of the project was to evaluate the modified version by testing its usefulness for numerical functions minimization. The optimized multidimensional functions: Griewank, Rastrigin, and Rosenbrock are frequently used as benchmarks which allows to compare the experimental results with outcomes reported in the literature. Both the results produced by the original version of the IWO algorithm and the Adaptive Particle Swarm Optimization (APSO) method served as the reference points.

Verification of the Search Space Exploration Strategy Based on the Solutions of the Join Ordering Problem

The paper addresses the problem of quality estimation of the search space exploration strategy. The strategy is used to find a satisfying solution to the join ordering problem, which constitutes a crucial part of the database query optimization task. The method of strategy verification is based on the comparison of the execution time for the solution produced by the Invasive Weed Optimization (IWO) algorithm with the analogous value for the solution determined by the SQL Server 2008 optimizer. Solutions were generated for star queries that are common in data warehousing applications. The authors discuss representations of the single solution and describe the modified version of the IWO algorithm emphasizing features of the proposed hybrid method of the search space exploration. The results of the conducted experiments form the main topic of analysis.

The exIWO Metaheuristic – A Recapitulation of the Research on the Join Ordering Problem

The authors summarize the several years research on the join ordering problem presenting a method based on the exIWO metaheuristic which is characterized by both the hybrid strategy of the search space exploration and three variants of selection of individuals as candidates for next population. The nub of the problem was recalled along with details of adaptation of the exIWO including representation of a single solution and transformation of an individual. Results of the experiments show that the exIWO algorithm can successfully compete with the SQL Server 2008 DBMS in optimization of join order in database queries.

Towards Evolutionary Super-Resolution

Super-resolution reconstruction (SRR) allows for producing a high-resolution (HR) image from a set of low-resolution (LR) observations. The majority of existing methods require tuning a number of hyper-parameters which control the reconstruction process and configure the imaging model that is supposed to reflect the relation between high and low resolution. In this paper, we demonstrate that the reconstruction process is very sensitive to the actual relation between LR and HR images, and we argue that this is a substantial obstacle in deploying SRR in practice. We propose to search the hyper-parameter space using a genetic algorithm (GA), thus adapting to the actual relation between LR and HR, which has not been reported in the literature so far. The results of our extensive experimental study clearly indicate that our GA improves the capacities of SRR. Importantly, the GA converges to different values of the hyper-parameters depending on the applied degradation procedure, which is confirmed using statistical tests.

Parametric Optimization of the Selected Classifiers in Binary Classification

The conception of classification is one of the major aspects in data processing. Conducted research present comparison of chosen classifiers’ results of classification for a few data sets. All data were chosen from these available on UCI Machine Learning Repository web site. During realization of research, the optimization process of the results of classification was made on the modifiable parameters for particular classifiers. In this work, gathered result of classification was presented as well as conclusion and possibility of future work.

Feature Extraction and HMM-Based Classification of Gait Video Sequences for the Purpose of Human Identification

The authors present results of the research on human recognition based on the video gait sequences from the CASIA Gait Database. Both linear (principal component analysis; PCA) and non-linear (isometric features mapping; Isomap and locally linear embedding; LLE) methods were applied in order to reduce data dimensionality, whereas a concept of hidden Markov model (HMM) was used for the purpose of data classification. The results of the conducted experiments formed the main subject of analysis of classification accuracy expressed by means of the Correct Classification Rate (CCR).

Towards Robust Evaluation of Super-Resolution Satellite Image Reconstruction

Super-resolution reconstruction (SRR) consists in processing an image or a bunch of images to generate a new image of higher spatial resolution. This problem has been intensively studied, but seldom is SRR applied in practice for satellite data. In this paper, we briefly review the state of the art on SRR algorithms and we argue that commonly adopted strategies for their evaluation do not reflect the operational conditions. We report our study on assessing the SRR outcome, relying on new quantitative measures. The obtained results allow us to outline the most important research pathways to improve the performance of SRR.

Adjusting Parameters of the Classifiers in Multiclass Classification

The article presents the results of the optimization process of classification for five selected data sets. These data sets contain the data for the realization of the multiclass classification. The article presents the results of initial classification, carried out by dozens of classifiers, as well as the results after the process of adjusting parameters, this time obtained for a set of selected classifiers. At the end of article, a summary and the possibility of further work are provided.