A5101435853- Evolutionary Algorithms and Applications
- Metaheuristic Optimization Algorithms Research
- Advanced Multi-Objective Optimization Algorithms
- Reinforcement Learning in Robotics
- Evolutionary Game Theory and Cooperation
- Evolution and Genetic Dynamics
- Artificial Intelligence in Games
- Modular Robots and Swarm Intelligence
- Game Theory and Applications
- Surface Modification and Superhydrophobicity
- Multi-Agent Systems and Negotiation
- Neural Networks and Applications
- Sports Analytics and Performance
- Logic, Reasoning, and Knowledge
- Micro and Nano Robotics
- Electrowetting and Microfluidic Technologies
- Flood Risk Assessment and Management
- Model Reduction and Neural Networks
- Disaster Management and Resilience
- Advanced Software Engineering Methodologies
- Direction-of-Arrival Estimation Techniques
- Sparse and Compressive Sensing Techniques
- Power Transformer Diagnostics and Insulation
- Logic, programming, and type systems
- Semantic Web and Ontologies
Utrecht University
2003-2024
Radboud University Nijmegen
2023-2024
Kema International (Netherlands)
2021-2023
University of Groningen
2014-2023
Eindhoven University of Technology
1981-2022
DNV (Netherlands)
2017
Saarland University
2007
Carnegie Mellon University
2007
Indiana University Bloomington
2007
University of Sussex
2007
We study the question of how a local learning algorithm, executed by multiple distributed agents, can lead to global system communication. First, notion perfect communication is defined. Next, two measures quality are specified. It shown that maximization these leads production. Based on this principle, adaptation rules for development constructed. The resulting stochastic algorithm validated in computational experiments. Empirical analysis indicates mild degree stochasticity instrumental...
In many problems of interest, performance can be evaluated using tests, such as examples in concept learning, test points function approximation, and opponents game-playing. Evaluation on all tests is often infeasible. Identification an accurate evaluation or fitness a difficult problem itself, approximations are likely to introduce human biases into the search process. Coevolution evolves set used for evaluation, but has so far led inaccurate evaluation. We show that any learners, Complete...
Disaster preparedness and response, including for landslides, increasingly involves local knowledge. Incorporating contextual, dynamic experience-based knowledge leads to greater awareness of the interconnectedness geological, natural, social processes. Still, technical literature on urban landslide risk is mainly based geological natural dynamics and, a lesser extent, physical infrastructure. Moreover, although recognized as important in principle all aspects disaster management, it not...
Various multi--objective evolutionary algorithms (MOEAs) have obtained promising results on various numerical optimization problems. The combination with gradient--based local search operators has however been limited to only a few studies. In the single--objective case it is known that additional use of gradient information can be beneficial. this paper we provide an analytical parametric description set all non--dominated (i.e. most promising) directions in which solution moved such its...
Abstract Fluvial systems in which peat formation occurs are typified by autogenic processes such as river meandering, crevasse splaying and channel avulsion. Nevertheless, cannot satisfactorily explain the repetitive nature lateral continuity of many coal seams (compacted peats). The fluvial lower Palaeocene Tullock Member Fort Union Formation (Western Interior Williston Basin; Montana, USA ) contains lignite rank that traceable over distances several kilometres. This sequence is used to...
Worldwide, flood risk is on the rise.Simultaneously, UNDRR fears that humanity's perception broken.Low-intensity, high-frequency extensive floods are cumulatively most damaging phenomenon.However, in context of widely under-researched.Indonesia nexus low-risk perceptions, and poverty.Therefore, a socio-economically marginalised frequently flooded urban kampong where residents appear to exhibit low levels perception, serves as case for this study.Data was gathered through five months...
Recent advances in digital pathology have demonstrated the effectiveness of foundation models across diverse applications. In this report, we present a novel vision model based on RudolfV approach. Our was trained dataset comprising 1.2 million histopathology whole slide images, collected from two medical institutions: Mayo Clinic and Charit\'e - Universt\"atsmedizin Berlin. Comprehensive evaluations show that our achieves state-of-the-art performance twenty-one public benchmark datasets,...
Pathology Foundation Models (FMs) hold great promise for healthcare. Before they can be used in clinical practice, it is essential to ensure are robust variations between medical centers. We measure whether pathology FMs focus on biological features like tissue and cancer type, or the well known confounding center signatures introduced by staining procedure other differences. introduce Robustness Index. This novel robustness metric reflects what degree dominate features. Ten current publicly...
Coevolution can be used to adaptively choose the tests for evaluating candidate solutions. A long-standing question is how this dynamic setup may organized yield reliable search methods. Reliability only considered in connection with a particular solution concept specifying what constitutes solution. Recently, monotonic coevolution algorithms have been proposed several concepts. Here, we introduce new algorithm that guarantees monotonicity of maximizing expected utility The method, called...
Coevolution has already produced promising results, but its dynamic evaluation can lead to a variety of problems that prevent most algorithms from progressing monotonically. An important open question therefore is how progress towards chosen solution concept be achieved. A general for coevolution obtained by viewing opponents or tests as objectives. In this setup known Pareto-coevolution, the desired Pareto-optimal set. We present an archive guarantees monotonicity concept. The algorithm...
This paper describes a novel algorithm called CON-MODP for computing Pareto optimal policies deterministic multi-objective sequential decision problems. is value iteration based dynamic programming that only computes stationary policies. We observe guaranteeing convergence to the unique set of policies, needs perform policy evaluation step on particular are inconsistent in single state being expanded. prove converges functions and infinite horizon discounted Markov processes. Experiments...
Many applications in modern technology, such as self-cleaning surfaces and digital microfluidics, require control over individual fluid droplets on flat surfaces. Existing techniques may suffer from side effects resulting high electric fields temperatures. Here, we introduce a markedly different method, termed "mechanowetting," that is based the surface tension-controlled droplet motion deforming The method demonstrated by transporting using transverse waves horizontal (vertically) inclined...
Recently, gradient techniques for solving numerical multi-objective optimization problems have appeared in the literature. Although promising results already been obtained when combined with evolutionary algorithms (MOEAs), an important question remains: what is best way to integrate use of cycle a MOEA. In this paper, we present adaptive resource-allocation scheme that uses three addition variation operator During optimization, effectivity monitored and available computational resources are...
Competent Genetic Algorithms can efficiently address problems in which the linkage between variables is limited to a small order k. Problems with higher dependencies only be addressed if further problem properties exist that exploited. An important class of for this occurs hierarchical problems. Hierarchical contain all (k=n) while being solvable polynomial time.An open question so far what precise must possess efficiently. We study by investigating several features and determining their...
Coevolution has often been based on averaged outcomes, resulting in unstable evaluation. Several theoretical approaches have used archives to provide stable However, the number of tests required by some these can be prohibitive practical applications. Recent work shown existence a set underlying objectives which compress evaluation information into potentially small dimensions. We consider whether approximated online, and for coevolution algorithm. The Dimension Extracting Coevolutionary...
Convolution offers adaptive methods for the selection of tests used to evaluate individuals, but resulting evaluation can be unstable. Recently, general archive-based coevolution have become available which monotonic progress guaranteed. The size these archives may grow indefinitely however, thus limiting their application potential. Here, we investigate how an archive Pareto-coevolution limited while maintaining reliability. LAyered Pareto-Coevolution Archive (LAPCA) is presented, and...