Phd topics -
Multiscale simulation software for industrial metal forming processes50000403v52448834
|
|
Promoter: Prof. dr. ir. Dirk Roose More info: jobsite
More info: https://icts.kuleuven.be/apps/jobsite/vacatures/id/52448834/lang/en
The PhD project fits into the KU Leuven Knowledge Platform M2Form on 'Software for Multi-scale Simulations of Material Forming Processes', a collaborative effort of the Departments of Computer Science and of Metallurgy and Materials Engineering. The Knowledge Platform aims at creating and providing expertise on multi-scale computational models to industry, in the form of a software solution. The research group Scientific Computing (Dept. of Computer Science) focuses on the development of numerical methods, algorithms and software for large-scale simulation in science and engineering. Important topics are numerical methods for multi-scale simulation and high performance computing. ProjectThe KU Leuven Knowledge Platform M2Form has developed a hierarchical multi-scale model that allows for adaptive, local and on-demand updating of constitutive material models in metal forming operations. The modelling involves predicting the evolution of crystallographic texture and the associated plastic anisotropy, which is done by employing physics-based crystal plasticity codes. Such models are needed by industry to improve virtual designs of new materials, products and processes. We will pursue further model developments, software engineering and experimental validations to increase the functionality of the multiscale modelling tools. The developments will be partly driven by the needs of material producing and forming companies. To address computational feasibility of the model, the software will increasingly use high performance computing techniques and efficient numerical algorithms. We also intend to generalise and extend this multiscale modelling framework by coupling models of other physical phenomena that are relevant in industrial cases. The final goal of the KP is to deliver a software solution that will couple the multi-scale materials models with industrially relevant finite element analysis software packages. The aim of the PhD project is to further develop hierarchical multiscale modelling software for the simulation of metal forming processes. Within the existing software framework, new numerical multiscale methods will be developed.  ProfileYou should have an MSc degree in one of the following majors: applied mathematics, computer science,materials science, engineering or physics. He or she should have a background in one or more of the following areas: high performance scientific computing, computational mechanics. You should be interested in working in a multi-disciplinary research team. You will contribute to the optimization of the software for parallel execution on clusters of multicore processors. You will be able to make use of the HPC research facilities of the KU Leuven and the Flemish Supercomputer Centre (VSC). Offer We offer a fax-free PhD grant to work in an interdisciplinary environment. Interested?For more information please contact Prof. dr. ir. Dirk Roose, tel.: +3216327546, mail: dirk.roose@cs.kuleuven.be. You can apply for this job no later than June 15, 2013 via the Click here to apply to this project |
Simulations of the properties and morphology of polymers at interfaces50000403v52424897
|
|
Promoter: Prof. dr. ir. Dirk Roose More info: jobsite
More info: https://icts.kuleuven.be/apps/jobsite/vacatures/id/52424897/lang/en
The PhD project fits into the KU Leuven Knowledge Platform M2Form on 'Software for Multi-scale Simulations of Material Forming Processes', a collaborative effort of the Departments of Computer Science and of Metallurgy and Materials Engineering. The Knowledge Platform aims at creating and providing expertise on multi-scale computational models to industry, in the form of a software solution. The research group Scientific Computing (Dept. of Computer Science) focuses on the development of numerical methods, algorithms and software for large-scale simulation in science and engineering. Important topics are numerical methods for multi-scale simulation and high performance computing. ProjectThe KU Leuven Knowledge Platform M2Form has developed a hierarchical multi-scale model that allows for adaptive, local and on-demand updating of constitutive material models in metal forming operations. The modelling involves predicting the evolution of crystallographic texture and the associated plastic anisotropy, which is done by employing physics-based crystal plasticity codes. Such models are needed by industry to improve virtual designs of new materials, products and processes. We will pursue further model developments, software engineering and experimental validations to increase the functionality of the multiscale modelling tools. The developments will be partly driven by the needs of material producing and forming companies. To address computational feasibility of the model, the software will increasingly use high performance computing techniques and efficient numerical algorithms. We also intend to generalise and extend this multiscale modelling framework by coupling models of other physical phenomena that are relevant in industrial cases. The final goal of the KP is to deliver a software solution that will couple the multi-scale materials models with industrially relevant finite element analysis software packages. The aim of the PhD project is to further develop hierarchical multiscale modelling software for the simulation of metal forming processes. Within the existing software framework, new numerical multiscale methods will be developed.  ProfileYou should have an MSc degree in one of the following majors: applied mathematics, computer science,materials science, engineering or physics. He or she should have a background in one or more of the following areas: high performance scientific computing, computational mechanics. You should be interested in working in a multi-disciplinary research team. You will contribute to the optimization of the software for parallel execution on clusters of multicore processors. You will be able to make use of the HPC research facilities of the KU Leuven and the Flemish Supercomputer Centre (VSC). Offer We offer a fax-free PhD grant to work in an interdisciplinary environment. Interested?For more information please contact Prof. dr. ir. Dirk Roose, tel.: +3216327546, mail: dirk.roose@cs.kuleuven.be. You can apply for this job no later than June 15, 2013 via the The research group Scientific Computing focuses on the development of numerical methods, algorithms and software for large-scale simulation in science and engineering. Important topics are: - numerical methods for multi-scale simulation - numerical methods for non-linear dynamical systems - high performance computing Project A PhD position is available in the Scientific Computing research group of the department of Computer Science, Faculty of Engineering of the KU Leuven to participate in the project 'Simulations of morphology and properties of polymers at interfaces'. The position is part of the interuniversity collaboration program H-INT-S,sponsored by the Strategic Initiative Materials (SIM) of the Flemish Government, focusing on computational methods for molecular simulation of interfacial properties in polymer based composites. The PhD program usually takes four years. You will be able to make use in your research of the Tier 2 HPC research facilities of the KU Leuven and of the Tier 1research facilities of the Flemish Supercomputer Centre (VSC).Profile We welcome the application of highly qualified graduates with a Master's degree or comparable qualification in Applied Mathematics, Computational Sciences, Physics, Chemistry, Materials Science. A strong interest for multidisciplinary research and programming experience are required. Good knowledge of the English language, both spoken and written, is essential.  OfferWe offer a tax-free PhD research grant to work in a interdisciplinary environment.  Interested?For more information please contact Prof. Dirk Roose, tel.: +3216327546, mail: dirk.roose@cs.kuleuven.be or Prof. Giovanni Samaey, tel.: +3216327558, mail: giovanni.samaey@cs.kuleuven.be. You can apply for this job no later than June 07, 2013 via the Click here to apply to this project |
Design patterns for large-scale multi-agent systems (Brazil - Science Without Borders)50000403v7181
|
Promoter: Tom Holvoet details
Description: Design patterns for large-scale multi-agent systems This scholarship is part of the program: "Brazil - Science Without Borders" - please check eligibility before applying! Multi-agent systems are systems that are composed of multiple autonomous entities that cooperate in order to achieve common objectives. Research and developments in this domain focus on the individual behaviour of agents and on cooperation mechanisms. From a software engineering point of view, multi-agent systems provide a wide spectrum of architectural tactics. Today, however, engineering complex distributed applications using these tactics is hard and mostly follows an ad hoc approach. The objective of this PhD project is to define design patterns and a pattern language for multi-agent systems. Rather than aiming to cover all possible aspects of multi-agent systems, the pattern language will focus on multi-agent systems research that is devoted to coordination of large-scale systems. The coordination mechanism called 'delegate MAS', as developed within our research team, will serve as the starting point for this research. A formal underpinning of the pattern language will allow unambiguous specification and will foster reuse in various applications. The pattern language will start from existing work and experience of the DistriNet research group on coordination and delegate MAS, as well as on new applications under study. A prototypical example of a large-scale coordination application is the well-known pickup-and-delivery problem. Studied mainly as an optimisation problem, the nature of PDP problems today are such that on-line decentrlized control application disserves attention: the problems are often huge in scale, and are subject diverse sources of dynamism and uncertainty. This PhD project fits in the expertise and strategic objectives of the MAS-related research of DistriNet, with a focus on a software architectural of MAS. DistriNet is involved in several projects where the integration of multi-agent systems and software engineering plays a crucial role. Key words: patterns, multi-agent systems, optimization, distributed software Latest application date: 2013-05-31 Financing: Brazil - Science Without Borders Type of Position: scholarship Duration of the Project : 4 years Research group: Apply to Click here to apply to this project |
Collaborative distributed computing (Brazil - Science Without Borders)50000403v7182
|
|
Promoter: Tom Holvoetdetails
Description: Collaborative distributed computing This scholarship is part of the program: "Brazil - Science Without Borders" - please check eligibility before applying! Many problems that are traditionally studied as combinatorial optimization problems are more and more being investigated as distributed coordination problems – examples are the vehicle routing problem (VRP) and the pickup-and-delivery problem (PDP). Real-world instances of these problems (esp. in the area of logistics, such as inland shipping and multi-modal transport) are under increasing performance pressure. Operations are submitted to more demanding circumstances such as extremely large scales (large numbers of transportation vehicles, tasks pickup/dropoff points) and dynamic problem updates (unpredictable service time). Combinatorial optimization techniques are limited due to computational complexity. Collaborative distributed approaches aim to cope with this complexity by combining advanced coordination mechanisms with specialized local decision making strategies. This project aims for a scientific study and evaluation of an approach called 'delegate MAS'. Delegate MAS is a biologically-inspired coordination mechanism: the coordination of autonomous entities (e.g. vehicles, road infrastructure elements) occurs through ant-like, light weight agents that crawl the environment on behalf of their autonomous entity, and retrieve and disseminate information. Initial experiments show the usefulness of the approach in manufacturing control and traffic control, yet a critical scientific evaluation and consequent revision of delegate MAS is the next core challenge in this research topic. The evaluation as well as the revision will include the study and possibly the adoption of existing combinatorial optimisation techniques. The research will be guided by a focus on the case of large-scale and distributed PDPs. References: - Tom Holvoet, Paul Valckenaers, Beliefs, desires and intentions through the environment, Proceedings of the fifth international joint conference on Autonomous agents and multiagent systems, pages 1052-1054, Hakodate, Japan, May 8-12, 2006 - Danny Weyns, Tom Holvoet, Alexander Helleboogh, Anticipatory vehicle routing using delegate multi-agent systems, Intelligent Transportation Systems Conference, 2007. ITSC 2007. IEEE, pages 87-93, Seattle, USA, Sept. 30 2007-Oct. 3 2007 - Tom Holvoet, Danny Weyns, Paul Valckenaers, Patterns of Delegate MAS, Proceedings of the third international conference on Self-Adaptive and Self-Organizing Systems, 2009. SASO 2009. IEEE, to be published, San Francisco, USA, Sept. 14-18 2009 Key words: patterns, multi-agent systems, optimization, distributed software Latest application date: 2013-05-31 Financing: Brazil - Science Without Borders Type of Position: scholarship Duration of the Project : 4 years Research group: Apply to Click here to apply to this project |
