A5036114620- Advanced Optimization Algorithms Research
- Complexity and Algorithms in Graphs
- Advanced Graph Theory Research
- Optimization and Packing Problems
- VLSI and FPGA Design Techniques
- Computational Geometry and Mesh Generation
- Optimization and Variational Analysis
- Optimization and Search Problems
- Vehicle Routing Optimization Methods
- Sparse and Compressive Sensing Techniques
- Advanced Manufacturing and Logistics Optimization
- Matrix Theory and Algorithms
- Scheduling and Optimization Algorithms
- graph theory and CDMA systems
- Graph Labeling and Dimension Problems
- Numerical methods in inverse problems
- Machine Learning and Algorithms
- Statistical and numerical algorithms
- Limits and Structures in Graph Theory
- Psychological Well-being and Life Satisfaction
- Advanced Numerical Analysis Techniques
- Operations Management Techniques
- Manufacturing Process and Optimization
- Income, Poverty, and Inequality
- Constraint Satisfaction and Optimization
University of Klagenfurt
2013-2023
Graz University of Technology
1992-2022
University of Twente
2022
Eindhoven University of Technology
2022
Institute for Interdisciplinary Studies of Austrian Universities
1999-2015
Gesellschaft Fur Mathematik Und Datenverarbeitung
1996
Center for Discrete Mathematics and Theoretical Computer Science
1991
University of Waterloo
1988
We propose a new interior-point-based method to minimize linear function of matrix variable subject equality and inequality constraints over the set positive semidefinite matrices. show that approach is very efficient for graph bisection problems such as max–cut. Other applications include max–min eigenvalue relaxations stable problem.
A central drawback of primal-dual interior point methods for semidefinite programs is their lack ability to exploit problem structure in cost and coefficient matrices. This restricts applicability problems small dimension. Typically, relaxations arising combinatorial applications have sparse well-structured matrices huge order. We present a method that allows us compute acceptable approximations the optimal solution large within reasonable time. Semidefinite programming with constant trace...
We introduce a new class of algorithms for solving linear semidefinite programming (SDP) problems. Our approach is based on classical tools from convex optimization such as quadratic regularization and augmented Lagrangian techniques. study the theoretical properties we show that practical implementations behave very well some instances SDP having large number constraints. also “boundary point method” Povh, Rendl, Wiegele [Computing, 78 (2006), pp. 277–286] an instance this class.
New lower bounds for the quadratic assignment problem QAP are presented. These based on orthogonal relaxation of QAP. The additional improvement is obtained by making efficient use a tractable representation matrices having constant row and column sums. new bound easy to implement often provides high quality under an acceptable computational effort.
We study the problem of finding minimum bisection a graph into two parts prescribed sizes. formulate lower bounds on by relaxing node- and edge- incidence vectors cuts. prove that both relaxations provide same bound. The main fact we is duality between relaxed preserves very natural cardinality constraints present an analogous result also for max-cut problem, show relation edge relaxation some other optimality criteria studied before. Finally, briefly mention possible applications practical...