Search

Optical character recognition (OCR) systems help to digitize paper-based historical achieves. However, poor quality of scanned documents and limitations of text recognition techniques result in different kinds of errors in OCR outputs. Post-processing is an essential step in improving the output quality of OCR systems by detecting and cleaning the errors. In this paper, we present an automatic model consisting of both error detection and error correction phases for OCR post-processing. We propose a novel approach of OCR post-processing error correction using correction pattern edits and evolutionary algorithm which has been mainly used for solving optimization problems. Our model adopts a variant of the self-organizing migrating algorithm along with a fitness function based on modifications of important linguistic features. We illustrate how to construct the table of correction pattern edits involving all types of edit operations and being directly learned from the training dataset. Through efficient settings of the algorithm parameters, our model can be performed with high-quality candidate generation and error correction. The experimental results show that our proposed approach outperforms various baseline approaches as evaluated on the benchmark dataset of ICDAR 2017 Post-OCR text correction competition.

Natural systems often exhibit chaotic behavior in their space-time evolution. Systems transiting between chaos and order manifest a potential to compute, as shown with cellular automata and artificial neural networks. We demonstrate that swarm optimization algorithms also exhibit transitions from chaos, analogous to a motion of gas molecules, when particles explore solution space disorderly, to order, when particles follow a leader, similar to molecules propagating along diffusion gradients in liquid solutions of reagents. We analyze these ‘phase-like’ transitions in swarm optimization algorithms using recurrence quantification analysis and Lempel-Ziv complexity estimation. We demonstrate that converging iterations of the optimization algorithms are statistically different from non-converging ones in a view of applied chaos, complexity and predictability estimating indicators. An identification of a key factor responsible for the intensity of their phase transition is the main contribution of this paper. We examined an optimization as a process with three variable factors—an algorithm, number generator and optimization function. More than 9000 executions of the optimization algorithm revealed that the nature of an applied algorithm itself is the main source of the phase transitions. Some of the algorithms exhibit larger transition-shifting behavior while others perform rather transition-steady computing. These findings might be important for future extensions of these algorithms.

Differential evolution (DE) is one popular meta-heuristic, which is used to solve difficult optimization problems. In the last years, a huge number of new variants of the differential evolution has been introduced to outperform previously presented algorithms. To provide solutions of higher quality or to speed-up the convergence principles as control parameters adaptation, novel mutation strategies, or combination of different mutation strategies are often used. In this work, five different variants of the differential evolution have been chosen with the goal to investigate their inner dynamics, especially spread of positive genomes within the population. To capture relationships between individuals, temporal networks, more precisely contact sequences, are used. Based on the empirical results, we have concluded that temporal networks generated on the basis of the DE algorithms dynamics are non-Markovian temporal networks. For this reason, to analyze the causality-driven changes of diffusion speed in these networks, analytical methods described by Scholtes et al. have been used.

This paper presents an obstacle avoidance algorithm for swarm robot in unknown environment based on self-organizing migrating algorithm in which the trajectory is divided into a set of points created by the SOMA that robot must pass through. During the move, obstacles will be detected by sensors on the robot, it only knows the target position without knowing the obstacles location until detecting them. The fitness function is based on the principle of attraction from the target and the repulsion of the obstacles, which helps robot find the trajectory for moving, and the ways leaving it to move safety away from the trapped area. Simulation results are also presented to show the feasibility of the algorithm.

In this article, we describe the design and implementation of a variant version of SOMA named SOMA Pareto to solve ten hard problems of the 100-Digit Challenge. The algorithm consists of the following operations: Organization, Migration, and Update. In which, we focus on improving the Organization operation with the adaptive parameters of PRT and Step. When applying the SOMA Pareto to solve ten hard problems to 10 digits of accuracy, we achieved a competitive result: 85.04 points.

The paper proposes a method for the drone to catch the given target and avoid detected obstacles in its path based on the self-organizing migrating algorithm. In particular, a two-component fitness function is proposed based on the principle that the closer the target, the lower the fitness value, and the closer the obstacle, the higher the fitness value. Self-organizing migrating algorithm, a swarm intelligence algorithm, is used to predict the next positions that the drone will move to. These positions both satisfy the requirement to avoid obstacles and shorten the distance to the target. A map of two drones, two corresponding targets and four static obstacles was modeled on Matlab. The simulation results verify the correctness and effectiveness of the proposed method.

This paper discusses the mutual combination of the unconventional algorithm (in this case evolutionary algorithms), deterministic chaos and modeling on real data from astrophysics. Analytical programming with selected evolutionary algorithm is used to synthesize suitable models. The main attention in this paper is focused on various chaotic generators, which are used instead of classical pseudo-random number generators. Chaotic generators are used in conjunction with a special case - a generator based on a strange non-chaotic attractor. The performance of all chaotic and non-chaotic based generators is then mutually compared at the end.

The examination of the Elliott Wave theory is the main motivation of this contribution. All of the fundamental features of an proper Elliott Wave pattern (EW pattern) are reviewed and explained. Based on this knowledge, an algorithm for detection of these patterns is designed, developed and tested. Under several different algorithm settings, several EW pattern sets are obtained. They differ in amount of found EW patterns, quality and size. The following application of the developed detection algorithm was based on recognition of an incomplete EW patterns with aim of the prediction of the following progress of the time set. The Random Decision Forest and the Support Vector Machine are the machine learning algorithms employed for this task. The accuracy of trend prediction above 70 proves the relevancy of EW patterns on stock market data as well as the validity of the algorithm as a tool for detection of such patterns. © 2018 - IOS Press and the authors. All rights reserved.

In this paper, we apply the SOMA T3A algorithm to solve 10 hard problems of the 100-Digit Challenge of the GECCO 2019 Competition. With effective improvements in choosing Migrants and Leader in the organization process, as well as the Step and PRT adaptive parameters in the migration process, the algorithm has achieved promising results. The total score that the algorithm achieved is 92.04 points.

In this paper, we propose a new method named Pareto-based self-organizing migrating algorithm (SOMA Pareto), in which the algorithm is divided into the Organization, Migration, and Update processes. The important key in the Organization process is the application of the Pareto Principle to select the Migrant and the Leader, increasing the performance of the algorithm. The adaptive PRT, Step, and PRTVector parameters are applied to enhance the ability to search for promising subspaces and then to focus on exploiting that subspaces. Based on the testing results on the well-known benchmark suites such as CEC'13, CEC'15, and CEC'17, the superior performance of the proposed algorithm compared to the SOMA family and the state-of-the-art algorithms such as variant DE and PSO are confirmed. These results demonstrate that SOMA Pareto is an effective, promising algorithm.

In this article, a novel leader selection strategy for the self-organizing migrating algorithm is introduced. This strategy replaces original AllToOne and AllToRand strategies. It is shown and statistically tested, that the new strategy outperforms the original ones. All the experiments were conducted on well known CEC 2014 benchmark functions according to the CEC competition rules and reported here.

In this article we continue in our research which combines three different areas - swarm and evolutionary algorithms, networks and coupled map lattices control. Main aim of this article is to compare networks obtained from best and worst self-organizing migrating algorithm runs. All experiments were done on well known CEC 2014 benchmark functions. For each selected function we picked 30 best and 30 worst runs, converted each run into a network, counted selected properties and compared the results. All obtained results are reported in this article.

This paper presents a method for swarm robot to catch the moving target and to avoid multiple dynamic obstacles in the unknown environment. An imaginary map is built, including the highest mountain, some small hills, and a lowest lying land, respectively corresponding to the starting position of the robot, the detected obstacles, and the target. The robot is considered as a flow of water flowing from high to low. The flow of water is the robot trajectory that is divided into a set of points created by an algorithm called Self-organizing migrating algorithm. Simulation results are also presented to show that the obstacle avoidance and catching target task can be reached using this method.

In this study, we propose a repulsive mechanism for the Particle Swarm Optimization algorithm that improves its performance on multi-modal problems. The repulsive mechanism is further extended with a distance-based modification. The results are presented and tested for statistical significance. We discuss the observations and propose further directions for the research.

This research represents a detailed insight into the modern and popular hybridization of deterministic chaotic dynamics and evolutionary computation. It is aimed at the influence of chaotic sequences on the performance of four selected Differential Evolution (DE) variants. The variants of interest were: original DE/Rand/1/ and DE/Best/1/ mutation schemes, simple parameter adaptive jDE, and the recent state of the art version SHADE. Experiments are focused on the extensive investigation of the different randomization schemes for the selection of individuals in DE algorithm driven by the nine different two-dimensional discrete deterministic chaotic systems, as the chaotic pseudorandom number generators. The performances of DE variants and their chaotic/non-chaotic versions are recorded in the one-dimensional settings of 10D and 15 test functions from the CEC 2015 benchmark, further statistically analyzed. © 2018, Brno University of Technology. All rights reserved.

This article describes the initial experiments using a new approach to automatic identification of Be and B[e] stars spectra in large archives. With enormous amount of these data it is no longer feasible to analyze it using classical approaches. We introduce an evolutionary synthesis of the classification by means of analytic programming, one of methods of symbolic regression. By this method, we synthesize the most suitable mathematical formulas that approximate chosen samples of the stellar spectra. As a result is then selected the category whose formula has the lowest difference compared to the particular spectrum. The results show us that classification of stellar spectra by means of analytic programming is able to identify different shapes of the spectra.

Automated comparison of faces in the photographs is a well established discipline. The main aim of this paper is to describe an approach whereby face recognition can be used in suggestion of a new contacts. The new contact suggestion is a common technique used across all main social networks. Our approach uses a freely available face comparison called "Betaface" together with our automated processig of the user´s Facebook profile. The research´s main point of interest is the comparison of friend´s facial images in a social network itself, how to process such a great amount of photos and what additional sources of data should be used. In this approach we used our automated processing algorithm Betaface in the social network Facebook and for the additional data, the Flickr social network was used. The results and their quality are discussed at the end.

In this article, we review how to capture the swarm and evolutionary algorithm dynamic into a network, and we show how we can cover the dynamical network into a coupled map lattices. The main part of this article focuses on the control of the algorithm via the coupled map lattices. Mainly, we concentrate on the statical control and the differential evolution algorithm. In the experimental part, we show that we can successfully lower or raise the control property. All the experiments are done on well-known CEC 2015 benchmark functions.

The self-organizing migrating algorithm is a population-based algorithm belonging to swarm intelligence, which has been successfully applied in several areas for solving non-trivial optimization problems. However, based on our experiments, the original formulation of this algorithm suffers with some shortcomings as loss of population diversity, premature convergence, and the necessity of the control parameters hand-tuning. The main contribution of this paper is the development of the novel algorithm mitigating the mentioned issues of the original self-organizing migrating algorithm. We have applied the ideas of the self-adaptation of the control parameters, the different principle of the leader creation, and the external archive of the successful particles. For some special cases, we are able to utilize the differential grouping to detect the interacting variables effectively removing the need for the perturbation parameter. To prove the efficiency of the novel algorithm, we have performed experiments on fifteen unconstrained problems from the CEC 2015 benchmark. The algorithm is compared with seven well-known evolutionary and swarm algorithms. The results of the experiments indicate that the mechanisms used in the novel algorithm had significantly improved the performance of the original self-organizing migrating algorithm, and the new algorithm can now compete with the selected algorithms.

In this chapter we describe possibility how to create complex network from SOMA algorithm and we describe what this network represents according to algorithm. Also we look at the visualization of such networks and show basic global complex network properties and how they look during the time.

In this chapter, that is a continuation of the previous one; we show some basic complex network analysis of SOMA algorithm on selected cost functions. The interesting facts, which can be used to improve the algorithm, are discussed here. Also, comparison of different runs of SOMA algorithm on the same test functions is presented here.

In this chapter, we focus on improvement of SOMA algorithm based on the previous analysis. We show some possibilities how to improve SOMA algorithm according to complex network analysis. At the end of the chapter, we show the best possible option how to improve standard SOMA algorithm together with results of a statistical test. Proposed improvements can be made (in principle) on arbitrary algorithm, SOMA here is used only for demonstrative purposes.

In this work, we present an overview of the various real-world application of Particle Swarm Optimization Algorithm. We argue that the PSO is showing superior performance on different optimization problems such as temperature prediction, battery storage optimization or leukemia diagnosis. The diversity of real-world applications covers the fields of electronic, informatics, energetics, medicine and many other areas of industry and research. This study should encourage new researchers for applying this method and take advantage of its unique inner dynamic and performance.

In this article, there is reported our progress of our research that merges evolutionary (swarm) algorithms, (complex) networks, and coupled map lattices. The article is mainly focused on the statical control of the self-organizing migrating algorithm. There is described how to convert networks into a coupled map lattices, how to create a feedback loop into the algorithm, and how to control the run of the algorithm through the coupled map lattice. Also, you can find here a small revision of our previous research. All experiments are done on a well known CEC 2015 benchmarks.

This research deals with the hybridization of several approaches for evolutionary algorithms, which are the adaptive control parameter adjusting strategy and multi-chaotic dynamics driving the selection of indices in Differential Evolution (DE). The novelty of the paper is given by the experiments with the multi-chaos-driven adaptive DE concept inside adaptive parameter adjusting DE strategies. These experiments are representing the investigations on the mutual influences of several different randomizations types together with adaptive DE strategies. The multi-chaotic concept is representing the adaptive switching between two different chaotic systems based on the progress of individuals within population. This paper is aimed at the embedding of discrete dissipative chaotic systems in the form of multi-chaotic pseudo random number generators for the jDE, which is the state of the art representative of simple adaptive control parameter adjusting strategy for DE. Repeated simulations for two different combinations of driving chaotic systems were performed on the IEEE CEC 13 benchmark set. Finally, the obtained results are compared with the canonical not-chaotic jDE.

In this paper, the utilization of different chaotic systems as pseudo-random number generators (PRNGs) for velocity calculation in the PSO algorithm are proposed. Two chaos-based PRNGs are used alternately within one run of the PSO algorithm and dynamically switched over when a certain criterion is met. By using this unique technique, it is possible to improve the performance of PSO algorithm as it is demonstrated on different benchmark functions.

In this research, evolutionary based algorithm Differential Evolution (DE) is compared with swarm based optimization technique PSO in the task of optimal evolutionary tuning of controller parameters for the stabilization of nonlinear chaotic oscillations of the two-dimensional Lozi map. The novelty of the approach is that the most utilized examples of evolutionary/swarm based algorithms are compared directly on the highly nonlinear and complex multimodal optimization and simulation task. The simulations were performed for two different required final behavior of the chaotic system.

This paper deals with mutual comparison of different pseudorandom number generators and its impact on the performance of selected algorithms of the symbolic regression. In this paper we discuss the use of genetic programming (GP) with use of chaotic systems as well as its main attributes and universal features. It is also explained why deterministic chaos is used and compared here with standard pseudorandom number generators. Based on the characterization of deterministic chaos, universal features of that kind of behavior are explained. In the second part of the paper we discuss selected examples of GP powered by deterministic chaos and classical pseudorandom number generators and its results.

In this paper, the utilization of chaos pseudorandom number generators based on three different chaotic maps to alter the behavior and overall performance of PSO algorithm is proposed. This paper presents results of testing the performance and behavior of the proposed algorithm on typical benchmark functions that represent unimodal and multimodal problems. The promising results are analyzed and discussed.

This paper introduces a novel Discrete Self-Organising Migrating Algorithm for the task of flow-shop scheduling with no-wait makespan. The new algorithm is tested with the small and medium Taillard benchmark problems and the obtained results are competitive with the best performing heuristics in the literature.

This paper discusses the possibility of using evolutionary algorithms for the reconstruction of chaotic systems. The main aim of this work is to show that evolutionary algorithms are capable of the reconstruction of chaotic systems without any partial knowledge of internal structure, i.e. based only on measured data and a predefined set of basic mathematical “objects”. Algorithm SOMA and differential evolution were used in reported experiments here. Systems selected for numerical experiments here is the well-known Lorenz system, Simplest Quadratic Flow, Double Sroll, Damped Driven Pendulum and Nose—Hoover oscillator. For each algorithm repeated simulations were done, totaling 20 simulations. According to obtained results it can be stated that evolutionary reconstruction is an alternative and promising way as to how to identify chaotic systems.

Presence of partial discharges implies the fault behavior on insulation system of medium voltage overhead lines, especially with covered conductors (CC). This paper covers the machine learning model based on features, which are derived from complex networks. These features are applied to predict whether the measured signal contains phenomenon indicating CC fault behavior or not. The comparison of different threshold levels of similarity values brings more information about complex network modeling. The final performance of the Random Forest classification algorithm shows valuable results for future research.

This article describes using an evolutionary algorithms to finding an original rule of generated 1D cellular automaton. In this article will be verified possibility and precision of searching rule from generated 1D cellular automaton using different evolutionary algorithms, random generators and rule definition. This extractor can be used for finding cellular automaton rule generating animals patterns like pigment patterns on the shells of mollusks.

In this paper we are continuing in our research to show mutual intersection of two different areas of research: complex network and evolutionary computation. This research parer is focused on possibility to convert run of evolution algorithm to a complex network inspired by ants. Such network can then be analyzed to get useful information about algorithm dynamics. In this paper we focused on one global network property, average network strength. This property is described and used for different types of cost functions and different representations of network conversion.

In this participation, we are continuing in our research on swarm based algorithm SOMA — Self Organized Migrating Algorithm and its use on problems defined in Competition track at WCCI 2016. In this paper we described SOMA algorithm, its conversion into complex network and we present modification called SOMARemove. Also we compare classical SOMA algoritm with SOMARemove on the first 8 functions of CEC15 benchmark in order to test the increase of the performance of new modification.

In this participation, we are continuing in our research on swarm based algorithm SOMA — Self Organized Migrating Algorithm and its use on problems defined in Competition track at WCCI 2016. SOMA algorithm has been used exactly according to defined propositions and obtained results are reported here. The most problematics step has been setting of the SOMA algorithm. In this paper we describe not only SOMA principles, but also how and why the SOMA parameters has been set to its values. The used approach has been used on learning — based real-parameter single objective optimization in order to test an increase the performance of selected SOMA strategy.

A discrete Self Organising Migrating Algorithm (DSOAM) is described in this chapter. This variant is specifically designed for the permutative based combinatorial optimisation problem, where the problem domain in generally NP-Hard. Specific sampling between individuals in the search space is introduced as a means of constructing new feasible individuals. These feasible solutions are improved using 2-Opt routines. DSOMA has proven successful in solving manufacturing scheduling and assignment problems.

In this paper the idea of capturing inner dynamics of evolutionary process in complex network is discussed. Complex network, in itself, can contain large number of information about running evolutionary process. Consequently this information can be used in subsequent adaptive control of algorithm. We present initial study of methodology for capturing the information about algorithm and present possible uses for created networks. For this study the Firefly algorithm was selected.

In this chapter a new approach for Particle Swarm Optimization (PSO) algorithm driven by chaotic pseudorandom number generator based on chaotic Lozi map is presented. This research represents the continuation of the satisfactory results obtained by means of chaos embedded (driven) swarm based algorithms, which utilize the chaotic dynamics in the place of pseudorandom number generators. The perturbation vector, which is introduced here, was inspired by the swarm based Self-organizing Migrating Algorithm (SOMA). It was embedded into the PSO algorithm to help overcome the issue of premature convergence.

In this paper a previously proposed method is extended with pseudo-random number generator based on chaotic sequences. Several recent approaches for designing the evolutionary computational techniques are merged in the proposed method. The proposed method represents a hybridization of heterogeneous swarm based PSO and differential evolution extended with the chaotic sequences implementation. The performance of the proposed method is tested on IEEE CEC 2013 benchmark set.

This paper deals with novel approach for hybridization of two scientific techniques: the evolutionary computational techniques and deterministic chaos. The Particle Swarm Optimization algorithm is enhanced with two pseudo-random number generators based on chaotic systems. The chaotic pseudo-random number generators (CPRNGs) are used to guide the particles movement through multiplying the accelerating constants. Different CPRNGs are used simultaneously in order to improve the performance of the algorithm. The IEEE CEC’13 benchmark suite is used to test the performance of the proposed method.

This research deals with the comparison of the two modern approaches for evolutionary algorithms, which are the adaptivity and complex chaotic dynamics. This paper aims on the investigations on the chaos-driven Differential Evolution (DE) concept. This paper is aimed at the embedding of discrete dissipative chaotic systems in the form of chaotic pseudo random number generators for the DE and comparing the influence to the performance with the state of the art adaptive representative jDE. This research is focused mainly on the possible disadvantages and advantages of both compared approaches. Repeated simulations for Lozi map driving chaotic systems were performed on the simple benchmark functions set, which are more close to the real optimization problems. Obtained results are compared with the canonical not-chaotic and not adaptive DE. Results show that with used simple test functions, the performance of ChaosDE is better in the most cases than jDE and Canonical DE, furthermore due to the unique sequencing in CPRNG given by the hidden chaotic dynamics, thus better and faster selection of unique individuals from population, ChaosDE is faster. © 2016 Author(s).

Evolutionary algorithms (EAs) have proven to be a powerful and robust optimizing technique even for complex optimization problems. The main aim of this work is to show that such a powerful simulating and optimizing of a non-linear dynamic process. In this paper, the complex reaction sequence used to study various reaction kinetics by optimization the rate constants. Two algorithms from the field of artificial intelligent—Differential evolution (DE), Self-organizing migrating algorithm (SOMA) are used in this investigation. Two optimization techniques were developed using Mathematica for accurately determining the rate constants of the reaction at certain temperature from the experimental data. The results show that EAs are used successfully in the process optimization.

A novel clustered population paradigm is presented in this paper which is based on Chaos principles of edges and attractors. Convergence in evolutionary algorithms is viewed as a manifestation through cyclic dynamics and thus a new population is developed which is clustered and separated through new segregation bias rules. This population is embedded on the Enhanced Differential Evolution and the flow shop scheduling problem with blocking is solved. The two flow shop benchmark problems of Rec/Car/Hel and Taillard are solved with this new approach and the results favorably compared with published results in literature. A total of 49 new upper bounds for the Taillard problems was obtained during experimentation.

The closeness centrality can be considered as the natural distance metric between pairs of nodes in connected graphs. This paper is the initial study of the influence of the closeness centrality of the graph built on the basis of the differential evolution dynamics to the differential evolution convergence rate. Our algorithm is based on the principle that the differential evolution creates graph for each generation, where nodes represent the individuals and edges the relationships between them. For each individual the closeness centrality is computed and on the basis of its value the individuals are selected in the mutation step of the algorithm. The higher value of the closeness centrality means the higher probability to become the parent in the mutation step. This enhancement has been incorporated in the classical differential evolution and a set of 21 well-known benchmark functions has been used to test and evaluate the performance of the proposed enhancement of the differential evolution. The experimental results and statistical analysis indicate that the enhanced algorithm performs better or at least comparable to its original version.

In this paper, we are continuing to show mutual intersection of two different areas of research: complex networks and evolutionary computation. We demonstrate that dynamics of evolutionary algorithms, that are based on Darwin theory of evolution and Mendel theory of genetic heritage, can be also visualized as complex networks. Such network can be then analyzed by means of classical tools of complex networks science. Results presented in our previous papers were currently numerical demonstration rather than theoretical mathematical proofs. We opened question whether evolutionary algorithms really create complex network structures and whether this knowledge can be successfully used like feedback for control of evolutionary dynamics and its improvement in order to increase the performance of evolutionary algorithms. This research paper is focused on the dynamics of complex networks from windows time point of view with proposition of a new windows time algorithm to evaluated evolution dynamics. There are described by temporal centralities and change centrality. These centralities are implemented as Gephi plugin and an own tool. At the end are examples of analysis of some networks using implemented algorithms.

This participation is focused on artificial intelligence techniques and their practical use in computer game. The aim is to show how program (based on evolutionary algorithms) can replace a man in the strategy game StarCraft: Brood War. Implementation used in our experiments use classic techniques of artificial intelligence environments, as well as unconventional techniques, such as evolutionary computation. An artificial player, proposed in this paper, is the combination of the decision tree and evolutionary algorithm SOMA. Whole code for experiments was written in the Java programming language. The proposed code provides a simple implementation of the artificial computer player in combination with slightly modified algorithm SOMA. This provides an opportunity for effective, coordinated movement of combat units around the combat landscape. Research reported here has shown potential benefit of evolutionary computation in the field of strategy games.

In this paper it is investigated the effect of tuning of control parameters of the Lozi Chaotic Map employed as a chaotic pseudo-random number generator for the particle swarm optimization algorithm. Three different benchmark functions are selected from the IEEE CEC 2013 competition benchmark set. The Lozi map is extensively tuned and the performance of PSO is evaluated.