Mario Köppen

Substitute distance assignments in NSGA-II for handling many-objective optimization problems

Many-objective optimization refers to optimization problems with a number of objectives considerably larger than two or three. In this paper, a study on the performance of the Fast Elitist Nondominated Sorting Genetic Algorithm (NSGA-II) for handling such many-objective optimization problems is presented. In its basic form, the algorithm is not well suited for the handling of a larger number of objectives. The main reason for this is the decreasing probability of having Pareto-dominated solutions in the initial external population. To overcome this problem, substitute distance assignment schemes are proposed that can replace the crowding distance assignment, which is normally used in NSGA-II. These distances are based on measurement procedures for the highest degree, to which a solution is nearly Pareto-dominated by any other solution: like the number of smaller objectives, the magnitude of all smaller or larger objectives, or a multi-criterion derived from the former ones. For a number of many-objective test problems, all proposed substitute distance assignments resulted into a strongly improved performance of the NSGA-II.

Hybrid information systems

Neural Networks and Applications.- A Full Explanation Facility for a MLP Network that Classifies Low-Back-Pain Patients and for Predicting its Reliability.- Use of Multi-category Proximal SVM for Data Set Reduction.- Neural Techniques in Logo Recognition.- Motion Detection Using Cellular Neural Network.- Speech Separation Based on Higher Order Statistics Using Recurrent Neural Networks.- Speaker Recognition Using Radial Basis Function Neural Networks.- A Multifaceted Investigation into Speech Reading.- Global Optimisation of Neural Networks Using a Deterministic Hybrid Approach.- AppART: An ART Hybrid Stable Learning Neural Network for Universal Function Approximation.- Monitoring System Security Using Neural Networks and Support Vector Machines.- A Hybrid Detection and Classification System for Human Motion Analysis.- Integrated Technique with Neurocomputing for Temporal Video Segmentation.- Matching Data Mining Algorithm Suitability to Data Characteristics Using a Self-Organizing Map.- Perceptual Grouping of Contours via Gated Diffusion of Boundary Signals.- Fuzzy Logic and Applications.- Fusion of Fuzzy System and Conventional Technique to Evaluate Weather and Terrain Effects on the Vehicle Operations.- Soft Computing for Developing Short Term Load Forecasting Models in Czech Republic.- An Induction Algorithm with Selection Significance Based on a Fuzzy Derivative.- Adaptive Database Learning in Decision Support Systems Using Evolutionary Fuzzy Systems: A Generic Framework.- Histogram-Based Fuzzy Clustering and its Comparison to Fuzzy C-Means Clustering in One-Dimensional Data.- Optimizing Linear Programming Technique Using Fuzzy Logic.- Semantics for Fuzzy Disjunctive Programs with Weak Similarity.- An Integration of Fuzzy and Two-Valued Logics on Natural Language Semantics.- Fuzzy Hyperplanes in the Hypothesis Space.- Evolutionary Computation and Other Heuristics.- A Genetic Algorithm for Optimizing Throughput in Non-broadcast WDM Optical Networks.- Solving Trigonometric Identities with Tree Adjunct Grammar Guided Genetic Programming.- Integrated Evolutionary Algorithms.- Evolving Natural Language Parser with Genetic Programming.- A Linear Genetic Programming Approach for Modeling Electricity Demand Prediction in Victoria.- Flexible Generator Maintenance Scheduling in a Practical System Using Fuzzy Logic and Genetic Algorithm.- Information Space Optimization with Real-Coded Genetic Algorithm for Inductive Learning.- A Comparison of GRASP and an Exact Method for Solving a Production and Delivery Scheduling Problem.- Intelligent Agents and Applications.- MEBRL: Memory Evolution Based Reinforcement Learning Algorithm of MAS.- Agent Representation and Communication in CBR-Tutor.- Agent-based Software Engineering and Agent Mediations.- Virtual AI Classroom: A Proposal.- An Argumentation-Based Multi-Agent System for eTourism Dialogue.- Modeling a Distributed Knowledge Management for Cooperative Agents.- Bayesian Methods / Rough Sets and Applications.- Linear Discriminant Text Classification in High Dimension.- A Bayesian Track-before-Detect Algorithm for IR Point Target Detection.- Application of Bayesian Controllers to Dynamic Systems.- An Algorithm for Automatic Generation of a Case Base from a Database Using Similarity-Based Rough Approximation.- A Family of Algorithms for Implementing the Main Concepts of the Rough Set Theory.- Intelligent Data Mining and Information Analysis.- Value of Information Analysis in Dynamic Influence Diagrams.- An Automated Report Generation Tool for the Data Understanding Phase.- Determining the Validity of Clustering for Data Fusion.- The Performance of Small Support Spatial and Temporal Filters for Dim Point Target Detection in Infrared Image Sequences.- Using Petri Nets for Modeling Branch Control of Pipeline Processors.- Extended Vector Annotated Logic Program and its Applications to Robot Action Control and Automated Safety Verification.- Hybrid Intelligent Systems Applications: Reviews and Frameworks.- Overview of Markov Chain Monte Carlo for Statistical Inference and its Application.- Insurance Applications of Soft Computing Technologies.- Teaming Human and Machine: A Conceptual Framework.- Dynamics and Thinking of Social Systems.- Author Index.

A fuzzy scheme for the ranking of multivariate data and its application

This paper presents a generic fuzzy scheme for the ranking of multivariate data. The scheme is based on a comparison function of two numbers. The comparison function values are fused by a T-norm for all components of two vectors, giving the comparison values. For each vector in a set of vectors, the smallest value of its comparison values with all other elements of the set is assigned to this vector as its ranking value. Then, all further processing is based on the ranking values alone. As a suitable comparison function, bounded division is identified. The application of the scheme to define a color morphology and an evolutionary multiobjective optimization algorithm is demonstrated.

Data Swarm Clustering

Data clustering is concerned with the division of a set of objects into groups of similar objects. In social insects there are many examples of clustering processes. Brood sorting observed in ant colonies can be considered as clustering according to the developmental state of the larvae. Also nest cleaning by forming piles of corpse or items is another example. These observed sorting and cluster capabilities of ant colonies have already been the inspiration of an ant-based clustering algorithm. Another kind of clustering mechanism can be observed in flocks of birds. In some rainforests mixed-species flocks of birds can be observed. From time to time different species of birds are merging to become a multi-species swarm. The separation of this multi-species swarm into its single species can be considered as a kind of species clustering. This chapter introduces a data clustering algorithm based on species clustering. It combines methods of Particle Swarm Optimization and Flock Algorithms. A given set of data is interpreted as a multi-species swarm which wants to separate into single-species swarms, i.e., clusters. The data to be clustered are assigned to datoids which form a swarm on a twodimensional plane. A datoid can be imagined as a bird carrying a piece of data on its back. While swarming, this swarm divides into sub swarms moving over the plane and consisting of datoids carrying similar data. After swarming, these sub swarms of datoids can be grouped together as clusters.

Visualization of Pareto-Sets in Evolutionary Multi-Objective Optimization

In this paper, a method for the visualization of the population of an evolutionary multi-objective optimization (EMO) algorithm is presented. The main characteristic of this approach is the preservation of Pareto-dominance relations among the individuals as good as possible. It will be shown that in general, a Pareto- dominance preserving mapping from higher- to lower- dimensional spaces does not exist. Thus, the demand is to find a mapping with as few wrongly indicated dominance relations as possible, which gives one more objective in addition to other mapping objectives like preserving nearest neighbor relations. Therefore, such a mapping poses a multi-objective optimization problem by itself, which is also handled by an EMO algorithm (NSGA-II in this case). The resulting mappings are shown for the run of a NSGA-II version on the 15 objective DTLZ2 problem as an example. From such plots, some insights into evolutionary dynamics can be obtained.

Many-Objective Particle Swarm Optimization by Gradual Leader Selection

Many-objective optimization refers to multi-objective optimization problems with a number of objectives considerably larger than two or three. This papers contributes to the use of Particle Swarm Optimization (PSO) for the handling of such many-objective optimization problems. Multi-objective PSO approaches typically rely on the employment of a so-called set of leaders that generalizes the global best particle used in the standard PSO algorithm. The exponentially decreasing probability of finding non-dominated points in search spaces with increasing number of objectives poses a problem for the selection from this set of leaders, and renders multi-objective PSOs easily unusable. Gradual Pareto dominance relation can be used to overcome this problem. The approach will be studied by means of the problem to minimize the Euclidian distances to a number of points, where each distance to the points is considered an independent objective. The Pareto set of this problem is the convex closure of the set of points. The conducted experiments demonstrate the usefulness of the proposed approach and also show the higher resemblance of the proposed PSO variation with the standard PSO.

No-Free-Lunch theorems and the diversity of algorithms

In this paper, the no-free-lunch theorem is extended to subsets of functions. It is shown that for algorithm a performing better on a set of functions than algorithm b, three has to be another subset of functions on which b performs better in average than a. to achieve a performance evaluation for an algorithm, it is not sufficient to demonstrate its better performance on a given set of functions. Instead of this, the diversity of an algorithm is considered in this paper in more detail. The total number of possible algorithms is computed and compared with the number of algorithms instances that a random search or a population-based algorithm can have. It comes out that the number of different random searches is very small in comparison to the total number of algorithms. On the other hand, population-based algorithms are principally able to cover the set of all possible algorithms. The smaller variance of algorithm performance, measured by the repeated application of the algorithm under different settings on different random sets of functions, comes out to be a value reflecting the higher count of instances.In this paper, the no-free-lunch theorem is extended to subsets of functions. It is shown that for algorithm a performing better on a set of functions than algorithm b, three has to be another subset of functions on which b performs better in average than a. to achieve a performance evaluation for an algorithm, it is not sufficient to demonstrate its better performance on a given set of functions. Instead of this, the diversity of an algorithm is considered in this paper in more detail. The total number of possible algorithms is computed and compared with the number of algorithms instances that a random search or a population-based algorithm can have. It comes out that the number of different random searches is very small in comparison to the total number of algorithms. On the other hand, population-based algorithms are principally able to cover the set of all possible algorithms. The smaller variance of algorithm performance, measured by the repeated application of the algorithm under different settings on different random sets of functions, comes out to be a value reflecting the higher count of instances.

Multiobjective optimization using adaptive Pareto archived evolution strategy

This paper proposes a novel adaptive representation for evolutionary multiobjective optimization for solving a stock modeling problem. The standard Pareto achieved evolution strategy (PAES) uses real or binary representation for encoding solutions. Adaptive Pareto archived evolution strategy (APAES) uses dynamic alphabets for encoding solutions. APAES is applied for modeling two popular stock indices involving 4 objective functions. Further, two bench mark test functions for multiobjective optimization are also used to illustrate the performance of the algorithm. Empirical results demonstrate APAES performs well when compared to the standard PAES,.

Multi-Jain Fairness Index of Per-Entity Allocation Features for Fair and Efficient Allocation of Network Resources

Due to its simplicity and its easy comprehension, Jains fairness index is still among the most popular measures to compare justness of allocations. However, it was already argued in the original paper that while the way of computing the index is well established, it is not immediately clear to which metric to apply the computation. Thereby, metric stands for a specific choice of a system observable. Here we study the extension of Jains index to multiple metrics at once. We propose a set of per-entity allocation features to represent justness of an allocation, and to derive corresponding vectors of feature-wise taken Jains fairness indices. The features give a numerical representation of fulfilling common fairness properties like proportionality, envy-freeness and equity of an allocation. Then, maximizing the smallest index gives an efficient procedure for allocation of goods. We study this procedure for the problem of allocating wireless channels in a multi-user setup and compare the influence of the various feature choices on the efficiency of the solution.

Color Effect on Subjective Perception of Progress Bar Speed

This present study attempts to find relationship between progress bar colors and subjective speed by subjective evaluation experiment. We prepared the six combinations of colors, blue/red for progress bar foreground color and cyan/orange/gray for background color. The test progress bars are designed under the same condition, for example progress bar size or animation speed, except colors. The test progress bars were displayed one after the other of all pair combinations, and made the subjects mark the test progress bar which they felt faster. As a result, there were no obvious significant color effects related to subjective speed impression in this subjective evaluation experiments. We consider the reason why not found significant color effects as follows, (i) its not enough data due to the small-scale experiments with only 10 subjects, (ii) color effect might not strong on progress bar.