This paper presents an experimental study that compares the performance of four selected metaheuristic algorithms for optimizing a time delay system model. Time delay system models are complex and challenging to optimize due to their inherent characteristics, such as non-linearity, multi-modality, and constraints. The study includes an explanation of the choice and core functionality of the selected algorithms, which are both baseline and state-of-the-art variants of self-organizing migrating algorithm (SOMA), state-of-the-art variant from the Success-History-based Adaptive Differential Evolution family of algorithms, with emphasis on diverse search (DISH algorithm), and Covariance Matrix Adaptation Evolution Strategy (CMA-ES) algorithm. The hyperparameters of the metaheuristic algorithms were set using the iRace automatic algorithm configuration framework. The paper emphasizes the importance of metaheuristic algorithms in control engineering for time-delay systems to develop more effective and efficient control strategies and precise model identifications. The experimental results highlight the effectiveness of the state-of-the-art algorithms with specific adaptive mechanisms like population organization process, diverse search and adaptation mechanisms ensuring a gradual transition from exploration to exploitation. Overall, this study contributes to understanding the challenges and advantages of using metaheuristic algorithms in control engineering for time delay systems. The results provide valuable insights into the performance of modern metaheuristic algorithms and can help guide the selection of appropriate adaptive mechanisms of metaheuristics.

Sequential pattern mining in general and one particular form, clickstream pattern mining, are data mining topics that have recently attracted attention due to their potential applications of discovering useful patterns. However, in order to provide them as real-world service applications, one issue that needs to be addressed is that traditional algorithms often view databases as static. In reality, databases often grow over time and invalidate parts of the previous results after updates, forcing the algorithms to rerun from scratch on the updated databases to obtain updated frequent patterns. This can be inefficient as a service application due to the cost in terms of resources, and the returning of results to users can take longer when the databases get bigger. The response time can be shortened if the algorithms update the results based on incremental changes in databases. Thus, we propose PF-CUP (pre-frequent clickstream mining using pseudo-IDList), an approach towards incremental clickstream pattern mining as a service. The algorithm is based on the pre-large concept to maintain and update results and a data structure called a pre-frequent hash table to maintain the information about patterns. The experiments completed on different databases show that the proposed algorithm is efficient in incremental clickstream pattern mining.

This paper describes the approach for the Parasitic Egg Detection and Classification in Microscopic Images challenge. Our solution relies on a robust deep learning pipeline implementing a five-fold training schema to pursue the challenge goal. The final methodology utilizes the TOOD model, further enhanced by slicing-aided fine-tuning and inference. The slicing helps to overcome the image size invariability of the dataset and allows the model to access all images in high resolution, and consequently helps it learn detailed features needed to distinguish different classes and find a precise object position. Our results demonstrate the importance of proper data analysis and consequent pre and post-processing to improve prediction performance.

Heating-cooling processes are widely used in industry and are often subject to latencies and delays. Although the parameters of such processes can be identified using mathematical and physical modeling, this technique is highly demanding both when it comes to the amount of a priori knowledge required and the time it takes to find a solution. In this work, we provide a robust formulation of a time-delay control problem and propose a numerical approach to finding its parameters in a form of an optimization task. Consequently, we apply various techniques and metaheuristics to obtain a high quality solution and discuss key observations made in the process.

Metaheuristics are universal optimization algorithms which should be used for solving difficult problems, unsolvable by classic approaches. In this paper we aim at constructing novel socio-cognitive metaheuristic based on castes, and apply several versions of this algorithm to optimization of time-delay system model. Besides giving the background and the details of the proposed algorithms we apply them to optimization of selected variants of the problem and discuss the results.

Following the introduction of the socio-cognitive caste-based algorithms into the classic evolutionary metaheuristics, in this paper we focus on similar task regarding agent-based universal optimization methods. We tackle EMAS and DE algorithms and enrich them also with TOPSIS-inspired mechanism. Besides giving the details of the methods and the background, we present preliminary results after applying those techniques to solving the problem of optimization of time-delay system model.

Closed sequential pattern (CSP) mining is an optimization technique in sequential pattern mining because they produce more compact representations. Additionally, the runtime and memory usage required for mining CSPs is much lower than the sequential pattern mining. This task has fascinated numerous researchers. In this study, we propose a novel approach for closed clickstream pattern mining using C-List (CCPC) data structure. Closed clickstream pattern mining is a more specific task of CSP mining that has been lacking in research investment; nevertheless, it has promising applications in various fields. CCPC consists of two key steps: It initially builds the SPPC-tree and the C-List for each frequent 1-pattern and then determines all frequently closed clickstream 1-patterns; next, it constructs the C-List for each frequent k-pattern and mines the remaining frequently closed k-patterns. The proposed method is optimized by modifying the SPPC-tree structure and a new property is added into each node element in both the SPPC-tree and C-Lists to quickly prune nonclosed clickstream. Experimental results conducted on several datasets show that the proposed method is better than the previous techniques and improves the runtime and memory usage in most cases, especially when using low minimum support thresholds on the huge databases.

The Bison Algorithm is a recent swarm optimization algorithm based on bison behavior. The algorithm divides the population into two groups, simulating exploitation and exploration patterns separately. The exploration is particularly linked with crossing the search space boundaries. This paper compares several boundary violation protocols: the hypersphere, reflection, random positioning, and clipping strategies on IEEE CEC 2017 benchmark and seeks the most fitting method for the Bison Algorithm. © ECMS Mauro Iacono, Francesco Palmieri, Marco Gribaudo, Massimo Ficco (Editors).

To date, there remains a lack of works that focus on the problem of mining clickstream patterns. Although it is an alternative to use the general algorithms for sequential pattern mining to mine clickstreams, their performance may suffer and the resources needed are more than necessary. In this paper, we present a novel data structure, called index-IDList, that is suitable for clickstream pattern mining. Based on this data structure, we present a vertical format algorithm named CUI (Clickstream pattern mining Using Index-IDList). The experiments are carried out on four real-life clickstream databases and the results show that our proposed method is effective and efficient in terms of runtimes and memory consumption.

This paper represents a comprehensive review of selected methods for visualization of the population dynamics of the swarm and evolutionary algorithms in the form of networks. The whole idea is based on the obvious similarity between interactions between individuals in a swarm and evolutionary algorithms and for example, users of social networks, society, etc. The analogy between individuals from the population used in an arbitrary evolutionary or swarm-based algorithm and vertices (nodes) of a network is discussed here, as well as between edges in a network and communication between individuals in a population. Simple experiments with four well-known heuristic algorithms are described here, giving an insight into different approaches to the building of the network during metaheuristic run. © ECMS Lars Nolle, Alexandra Burger, Christoph Tholen, Jens Werner, Jens Wellhausen

This paper discuss the utilization of the complex chaotic dynamics given by the selected time-continuous chaotic systems as well as by the discrete chaotic maps, as the chaotic pseudo random number generators and driving maps for the chaos based optimization. Such an optimization concept is utilizing direct output iterations of chaotic system transferred into the required numerical range or it uses the chaotic dynamics for mapping the search space mostly within the local search techniques. This paper shows totally three groups of complex chaotic dynamics given by chaotic flows, oscillators and discrete maps. Simulations of examples of chaotic dynamics mapped to the search space were performed and related issues like parametric plots, distributions of such a systems, periodicity, and dependency on internal accessible parameters are briefly discussed in this paper. © ECMS Zita Zoltay Paprika, Péter Horák, Kata Váradi,Péter Tamás Zwierczyk, Ágnes Vidovics-Dancs, János Péter Rádics (Editors).

This paper investigates the utilization of the complex chaotic dynamics given by the several selected time-continuous chaotic systems, as the chaotic pseudo random number generators and driving maps for the chaotic optimization. Chaos based optimization concept is utilizing direct output iterations of chaotic system transferred into the required numerical range or it uses the chaotic dynamics for mapping the search space mostly within the local search techniques. This paper shows two groups of complex chaotic dynamics given by either chaotic flows or oscillators. Simulations of examples of chaotic dynamics transferred to the pseudo random number generators were performed and related issues like distributions of such a systems, periodicity, and dependency on sampling times are briefly discussed in this paper. © ECMS Thorsten Claus, Frank Herrmann, Michael Manitz, Oliver Rose (Editors).

This paper proposes a new technique for the estimation of values of constants in programs synthesized by Analytical Programming (AP). Proposed technique divides the features of an individual in Evolutionary Algorithm (EA) into two parts - program part and constant part. Features in program part are mapped to synthesized program and features in constant part are used as constant values. AP with implementation of this technique is tested on four benchmark functions - Quintic, Sextic, 3Sine and 4Sine. © ECMS Thorsten Claus, Frank Herrmann, Michael Manitz, Oliver Rose (Editors).

This research deals with the hybridization of the two softcomputing fields, which are chaos theory and evolutionary computation. This paper aims on the experimental investigations on the chaos-driven evolutionary algorithm Differential Evolution (DE) concept. This research represents the continuation of the preliminary satisfactory results obtained by means of chaos embedded (driven) DE, which utilizes the chaotic dynamics in the place of pseudorandom number generators The novelty of this work represents the experimental analysis of the chaotic dynamics directly injected into the DE. To be more precise, this research investigates the influence of parameter settings to the performance of chaos driven DE. Both settings for mutation and crossover DE control parameters and adjustable chaotic system parameters are experimentally investigated here. Repeated simulations were performed on the selected set of well-known benchmark functions in higher dimensions. Finally, the obtained results are compared with canonical DE and state-of-the art representative of basic adaptive variant jDE.

This article is focused on presenting our research about direct punch force. Professional defence is very important part of our life and punches are basic technique in majority of martial arts. Our aim was to measure dependence of force on time. Then we found differences between genders and among groups of participants with different level of training. The analysis started with measuring of the force with help of strain gauge L6E-C3-300kg and then we continued with finding of the dependence of this force on time and on input parameters. For data analysis two pieces of software were used-Office Excel and MINITAB. Our goal was to prepare data for an artificial neural network. © ECMS Valen M. Mladenov, Petia Georgieva, Grisha Spasov, Galidiya Petrova (Editors).

This research deals with the analysis of the behaviour of artificial neural nets for prediction of raw material batches for the production of clinker. During the production several oxides that are present in raw materials in quarries have to be extracted for homogenization of the mixture suitable for clinker production. There is some delay between the measurement of the mixture and the material which is send from quarry. It is necessary to “send” precise chemical composition to ensure a good quality of clinker and resulting product cement. Artificial neural networks (ANN) are suitable for such kind of timeindependent prediction. The results show that not all oxides are necessary to use for the prediction of one oxide. The ANN were designed into several nets with one input similarly as pseudo neural networks are able to work. The results will be used for the purpose of further research of pseudo neural nets which currently serve only as classifiers. INTRODUCTION The paper deals with the analysis of behaviour of artificial neural networks during the data training for prediction of raw materials batches for the production of clinker. This analysis should help for further research of pseudo neural networks. Pseudo neural networks were developed in 2013 and further tested (Kominkova Oplatkova et al. 2013a, 2013b, 2014). Pseudo neural networks (PNN) were inspired in artificial neural nets (ANN). There is synthesis of the model structure relation between inputs and outputs in both pseudo neural networks and artificial neural nets. ANN uses structures of nodes with transfer functions like sigmoid and training algorithms like back propagation or Levenberg-Marquardt. On the contrary, pseudo neural net synthesizes the structure via evolutionary process of symbolic regression there are no nodes and weights from the respective of ANN but there is a function dependent on inputs. The first simulations with PNN were connected with the area of classification. The principle of pseudo neural network allows usage of only one output but it can be used multiple inputs. Therefore classification into several classes was done in the way either of continuous intervals in one output node or connection of several independent nets (symbolically designed models) gives the combinations of values in more output nodes. As the process of pseudo neural networks is evolutionary driven some inputs were omitted in the final model. In some cases, it might serve as an advantage which reduces the complexity of the structure. And at the same time, the most important inputs connected with the problem are evolutionary selected. Tasks like prediction need values from the history and also historical data of other features and most probably all of them are necessary. As these values are given to the input layer, it might happen that PNN would omit some inputs. For the purpose of the further research with pseudo neural nets, the analysis of behaviour of standard multilayered perceptron (feedforward artificial neural net) was studied in this paper. As the dataset, data from the production of clinker was used. Researchers have explored some stages of the cement and clinker production with usage of ANN (Svinning et al. 2010, Stanisic et al. 2014, Topalov 2004) but they do not solve the prediction of the chemical composition of raw material. Firstly, introduction into the clinker production is given, than a description of artificial neural network (ANN) follows. Afterwards, the proposed experiment with results is explained. Conclusion and future work overview finish the paper. INTRODUCTION INTO CEMENT PRODUCTION Cement is the most common structural binder. It is a powdered inorganic hydraulic binder, which contains mainly a finely ground clinker. It is possible to affect its properties by means of selection, composition of raw materials and technological process of cement production and thus it can be predetermined to the widest possible usage or the specific usage. Nowadays, factories check workflow production and associated operations regularly and everything is subject to the maintenance and development of a functioning quality management system. Proceedings 29th European Conference on Modelling and Simulation ©ECMS Valeri M. Mladenov, Petia Georgieva, Grisha Spasov, Galidiya Petrova (Editors) ISBN: 978-0-9932440-0-1 / ISBN: 978-0-9932440-1-8 (CD) Cement production is very complicated, technically, time and energy consuming process, which comprises several stages: 1. Resource extraction. The crushed and milled material progresses through homogenization silos into stock silos. In this phase, regular analysis of chemical composition of mined rock are carried out. Based on analysis results, eg. the composition of raw materials or ground fineness can be adjusted. 2. Firing of clinker. The most important part of the cement production process. The raw powder passes through the heat exchanger, where the raw material is preheated and partial decarbonized, into the rotary kiln. During burning process artificial minerals are produced and by quenching they are stabilized and formed into Portland cement clinker. The most of the firing process is automatically controlled focusing on stability and little variability of resulting clinker parameters. 3. Grinding of clinker and storage. Clinker, together with other ingredients is ground to fine meal and then stored in silos. Relevant parameters are continuously checked. Testing is carried out on the modern equipment, such as a laser granulometer, automatic X-ray analyser and others. 4. Expedition. Cement is supplied either in bulk or packaged in paper bags. Prior to the shipment of cement is testing in accordance to technical standards. Modern control systems and the computer technology used in the cement production should guarantee its stable properties. An important contribution to the optimization, uniformity and control of the entire cement production process is automatic service laboratory. It is used for automated sampling at regular intervals, analyses and results archiving with minimizing of the human factor influence on inspection and test procedures during the cement production. The paper deals with the part of cement production which is aimed on optimal regulation (prediction) of the dosage of each mineral resource for the correct composition of clinker. PRODUCTION PROCESS The main raw materials for the production of cement are limestone and sodalite. These materials are mined in quarries with the help of blasting or only dragline using heavy equipment (Fig. 1). Wheel loaders and trucks transport the raw material into a crushing. Large stones are crushed to the size of the road gravel. Figure 1: Mining of the raw material in quarry and its crushing The crushed material with the help of a belt conveyor is transported to a materials landfill. The raw material is stored in the prehomogenized silo (Fig. 2) and then homogenized. Figure 2: Homogenization of the raw material Required mixture of crushed basic raw materials and correction of raw materials (silica sand, iron oxide correction) is prepared with the help of precision measuring equipment. Cylindrical or ball mills grind the raw material mixture to fine meal dry it at the same time. Then the meal is stored in silos and further homogenized (Fig. 3). Figure 3: Milling and drying of the raw material Firing of the raw materials needs temperatures about 1450 °C in an oven or in a rotary kiln with superior heat exchanger. These types of furnaces work in different techniques, while the main difference is the method of preparation and preheating the raw meal. The chemical reactions in a process known as sintering and subsequent quenching arises new product Portland cement clinker (Fig. 4). Clinker is the basic raw material for cement production. The production begins by grinding of input materials limestone, marl, clay shale, silica sand, fluorspar and iron ore. Raw materials are batched onto the conveyor belt, which is continuously analysed and the chemical composition of aggregates monitored continuously throughout volume and time. Dosing of the individual UTB ve Zlině, Fakulta aplikovane informatiky, 2010 13

This paper investigates the utilization of the time-continuous chaotic system, which is UEDA oscillator, as the chaotic pseudo random number generator. (CPRNG). Repeated simulations were performed investigating the influence of the oscillator sampling time to the selected heuristic, which is differential evolution algorithm (DE). Initial experiments were performed on the selected test function in higher dimensions. © ECMS Valen M. Mladenov, Petia Georgieva, Grisha Spasov, Galidiya Petrova (Editors).

Public Venture Capital financing often fail rigorous scrutiny in their selection of high-potential start-ups as compared to Private Venture capital. In some developing countries, decision making on final selection for financial support of early stage but high potential Small and Medium sized Enterprises (SMEs) are often ‘clouded’ by several factors including consideration of political party affiliations. This results in low capital recovery rate and a mischance in choosing deserving start-ups. This paper applies Fuzzy Preference Ranking Organization METHod for Enrichment Evaluation (Fuzzy PROMETHEE) method to evaluate and select early-stage but high potential start-up businesses in a government high priority area such as in Information and Communications Technology. A numerical example with pre-defined linguistic terms parameterized by triangular fuzzy numbers is provided. The framework could serve as a useful tool for decision makers in scrutinizing selection of start-ups in other government priority areas.

The utilization of Social media tools in business enterprises has tremendously increased with an increased number of users and a corresponding upsurge in time spent online. Online social media services such as Facebook and Twitter are used by companies to introduce new products and services, provide various supports and interact with customers on daily basis. This regular interaction of businesses and consumers results in huge amount of customer-generated content which is becoming a source of insight in analysing the often erratic consumer behaviour. For companies to harness the business potential of social media to increase competitive advantage, sentiments behind textual data of both their customers and that of their competitors must be keenly monitored and analysed. This paper demonstrates how companies especially those in the Telecommunication industry can seize the opportunity presented by social media to mine textual data to gain advantage over competitors by cumulatively understanding consumer opinions, frustrations and satisfaction. Using Facebook and Twitter sites of the top three telecommunication companies in Ghana: MTN, Vodafone and Tigo the paper reveals insights from unstructured texts of customers of these three companies. The results show (1) the exponential growth of social media users in Ghana (2) impact and numbers behind active social media participation in the telecommunication industry (3) the power of social media opinion mining for competitive analysis (4) how business value could be extracted from the huge unstructured textual data available on social media and (5) the company that is more responsive to customer concerns.

Cluster analysis or clustering is a task of grouping a set of objects in such a way that objects in the same group (called a cluster) are more similar (in some sense or another) to each other than to those in other groups (clusters). It is the main task of exploratory data mining and a common technique for statistical data analysis used in many fields, including machine learning, pattern recognition, image analysis, information retrieval, and bioinformatics.The topic of this paper is modern methods of clustering. The paper describes the theory needed to understand the principle of clustering and descriptions of algorithms used with clustering, followed by a comparison of the chosen methods. Proceedings 28th European Conference on Modelling and Simulation © ECMS Flaminio Squazzoni, Fabio Baronio, Claudia Archetti, Marco Castellani (Editors).

In this initial study it is investigated the possibility of using simple artificial neural network for classification of kick techniques based on their specific force course profile. The aim is to investigate whether the neural networks could be a suitable tool for such task and can be possibly used in following research that will deal with classification of punch techniques and also the striker's gender and level of training. Proceedings 28th European Conference on Modelling and Simulation © ECMS Flaminio Squazzoni, Fabio Baronio, Claudia Archetti, Marco Castellani (Editors).

This research deals with the deeper analysis of the novel concept of a multi-chaos-driven evolutionary algorithm Differential Evolution (DE). This paper is aimed at the embedding and alternating of set of two discrete dissipative chaotic systems in the form of chaos pseudo random number generator for DE. Repeated simulations were performed on the selected test function in higher dimensions. Finally, the obtained results are compared with canonical DE.

Proliferation of health related information on the internet is both welcoming and a concern. For instance when solicited information goes wrong, it tends to have dire consequences on the general public. Assessing the quality of internet health information is often difficult but a rational and systematic approach can be useful in evaluating the quality of the services they render to the public. The paper proposes a fuzzy VIKOR framework for evaluating and ranking internet health information providers under a fuzzy environment where uncertainties and subjectivities are catered for with linguistic variables. Linguistic variables with triangular fuzzy numbers (TFN) are used to evaluate weights of the evaluation criteria and the rankings of each internet health information provider. A numerical example is demonstrated using HIV/AIDS online information providers in the most adult prevalent country in the world. The proposed method is compared with TOPSIS and can be applied in evaluating the quality of other specific internet health information providers.

In this paper, Differential Evolution (DE) is used in the task of optimization of batch reactor geometry. The novality of the approach is that the six selected discrete dissipative chaotic maps are used as the chaotic pseudo random number generator to drive the mutation and crossover process in the DE. The optimized results obtained are compared with original reactor geometry and process parameters adjustment. The statistical analysis of the results given by six versions of chaos driven DE is compared with canonical DE strategy.

This research deals with a utilization of a tool for symbolic regression, which is analytic programming, for the purpose of the synthesis of a new robust control law. This universal synthesized robust chaotic controller secures the fully stabilization of selected set of discrete chaotic systems. The paper consists of the descriptions of analytic programming as well as selected chaotic systems, used heuristic and cost function. For experimentation, Self-Organizing Migrating Algorithm (SOMA) and Differential evolution (DE) were used.

This research deals with a utilization of tool for symbolic regression, which is analytic programming, for the purpose of the synthesis of a new control law. This synthesized chaotic controller secures the stabilization of high periodic orbit - oscillations between several values of discrete chaotic system, which is Henon map. The paper consists of the descriptions of analytic programming as well as chaotic system, used heuristic and cost function. For experimentation, Self-Organizing Migrating Algorithm (SOMA) and Differential evolution (DE) were used.

The paper deals with a steganalysis of PQ algorithm by means of neural networks. The paper continues with the research of steganalysis by means of neural networks and brings results for the other steganography tool and also simulation results for different settings of neural network.

The main idea of this paper is to compare feature selection methods for dimension reduction of the original dataset to reach optimization of steganalysis process by artificial neural networks (ANN). Feature selection methods are tools based on statistic exploited in pre-processing step of data mining workflow. These methods are very useful in a dimension reduction, removing of insignificant data, increasing comprehensibility and learning accuracy. Dimension reduction leads to reduced computational resource consumptions, which is validated by ANN simulations. Steganalysis is a field of the computer security, which deals with a discovering of hidden information in images which is normally unrecognizable. All dataming processes, which reduce the dimension of ANN input layer, should keep accuracy of steganalysis on the original level. © ECMS.

This research deals with a synthesis of control law for selected discrete chaotic system by means of analytic programming. The novelty of the approach is that a tool for symbolic regression - analytic programming - is used for the purpose of stabilization of higher periodic orbits - oscillations between several values of chaotic system. The paper consists of the descriptions of analytic programming as well as chaotic system and used blackbox type cost function. For experimentation, Self-Organizing Migrating Algorithm (SOMA) with analytic programming and Differential evolution (DE) as second algorithm for meta-evolution were used. © ECMS.

This paper focuses on the application of chaos driven Differential Evolution to the chaos control optimization problem. The focus of this paper is the embedding of chaotic systems in the form of the chaos number generator for Differential Evolution. The aim of this paper is also to show extreme sensitivity of quality of results on the selection of evolutionary algorithm, setting-up of evolutionary algorithm, construction of cost function and any small change in its design. As a model of deterministic chaotic system, the two dimensional Henon map was used. Two complex targeting cost functions were tested in this work. The optimization was realized in several ways, each one for another desired behavior of system. Repeated simulations demonstrated the robustness of the used method and constructed cost function. Finally, the obtained results are compared with evolutionary algorithms, Self-Organizing Migrating Algorithm (SOMA) and canonical Differential Evolution.

The paper deals with a synthesis of control law for a discrete chaotic Henon system by means of analytic programming. This is a preliminary study in which the aim is to show that tool for symbolic regression – analytic programming - is possible to use for such kind of problems. The paper consists of description of analytic programming as well as chaotic Henon system. This article contents only 21 successful simulations in the result section and will be extended within future tests in this field. SOMA (Self-Organizing Migrating Algorithm) with analytic programming was used for experiments in this case.

Evolutionary Programming (EP) seems a promising methodology to automatically find programs to solve new computing challenges. The Evolutionary Programming techniques use classical genetic operators (selection, crossover and mutation) to automatically generate programs targeted to solve computing problems or specifications. Among the methodologies related with Evolutionary Programming we can find Genetic Programming, Analytic Programming and Grammatical Evolution. In this paper we present the Evolutionary Programming Multiagent Systems (EPMAS) framework based on Grammatical Evolution (GE) to evolutionary generate Multi-agent systems (MAS) ad-hoc. We also present two case studies in MAS scenarios for applying our EPMAS framework: the predator-prey problem and the Iterative Prisoner's Dilemma. © ECMS.

This work deals with the comparison of performance of two selected evolutionary algorithms (EA) in the task of optimization of the control of chaos. The main aim of this work is to show that evolutionary algorithms are capable of optimization of chaos control, leading to satisfactory results and to show extreme sensitivity of quality of results on the selection of EA, setting-up of EA, construction of cost function (CF) and any small change in the CF design. As a model of deterministic chaotic system, the two dimensional Henon map was used. Two complex targeting cost functions were tested in this work. The optimization was realized in several ways, each one for another evolutionary algorithm or another desired periodic orbit and behavior of system. The evolutionary algorithms, SOMA (self-organizing migrating algorithm) and DE (differential evolution) were used in several versions. For each version, repeated simulations demonstrated the robustness of the used method and constructed CF. Finally, the obtained results are compared.

This research deals with the optimization of the control of chaos by means of evolutionary algorithms. The main aim of this work is to show that powerful optimizing tools like evolutionary algorithms can in reality be used for the optimization of deterministic chaos control. This work is aimed on an explanation of how to use evolutionary algorithms (EAs) and how to properly define the advanced targeting cost function (CF) securing very fast and precise stabilization of desired state for any initial conditions. As a model of deterministic chaotic system, the two dimensional Henon map was used. The evolutionary algorithm Self-Organizing Migrating Algorithm (SOMA) was used in four versions. For each version, repeated simulations were conducted to outline the effectiveness and robustness of used method and targeting CF.

The paper deals with a alternative tool for symbolic regression – Analytic Programming which is able to solve various problems from the symbolic regression domain. In this contribution main principles of Analytic Programming are described and explained. Then follows how Analytic Programming was used for setting an optimal trajectory for an artificial ant according to Koza. An ability to create so called programs, as well as Genetic Programming or Grammatical Evolution do, is shown in that part. Analytic Programming is a superstructure of evolutionary algorithms which are necessary to run Analytic Programming. In this contribution SelfOrganizing Migrating Algorithm and Differential Evolution as two evolutionary algorithms were used to carry simulations out.