Skip to main content
españolES
European Commission logo print header
Contenido archivado el 2024-04-16

Algorithms and Complexity

Objetivo

Algorithms represent all the computational techniques and mechanisms used in computer science. Efficient algorithms are crucial for the effective performance of any information processing system. ALCOM aims to establish a network of researchers cooperating in the field of algorithm design and analysis, and to carry out research in several key areas of computational complexity theory. The goal of the Action was to strengthen and advance fundamental research in the design of analysis of computer algorithmsin the European Community.
The primary goal of ALCOM was to obtain good algorithms and data structures for practical use, and to assess their complexity by discrete mathematical or analytic techniques. The secondary goal was to understand the inherent complexity of problems and methods so that questions of optimality (or near-optimality) can be considered.
A network of researchers cooperating in the field of algorithm design and analysis has been establihed. Coordinated research took place in the following areas:
data structures (new techniques for maintaining sets of geometric objects for efficient query answering, algorithms for computing obstacle free shortest paths, advances in persistent datastructures, on line algorithms for maintaining components in dynamic graphs, and a new release of the algorithms library LEDA);
computational geometry (new algorithms for computing Voronoi diagrams and Delaunay triangulations, output sensitive algorithms for hidden surface removal for polytopes, advances in scene analysis, and randomized incremental constructions);
graph algorithms (advances in graph layout and graph drawing problems, efficient on line algorithms for finding shortest paths in planar graphs, advances in the analysis of on line algorithms for hypergraphs, and new algorithms for deciding k-connectivity);
probabilistic methods average case analysis (performance of unification algorithms and of search problems on tree structured data, analysis of algorithms for random graphs, probabilistic analysis of a geometric lattice reduction algorithm, and a new release of the LUO system for automatic analysis);
complexity theory (new bounds for arithmetic circuits, results for computational learning and strategy determination, complexity results for local search algorithms, advances in the study of polynominal complexity classes and the complexity of algebraic algorithms) parallel algorithms (advances in the study of P-complete and numerical control (NC) problems, processor efficient algorithms, efficient expected case parallel algorithms and algorithms for optimal gridrouting);
distributed algorithms (analysis of adaptive routing algorithms, new techniques for routing with compact tables, new algorithms for atomic registers, results for distributed datastructuring, and design methodologies for some classes of distri buted control algorithms and protocols).
APPROACH AND METHODS
Research in ALCOM focused on the development of general tools and techniques that can be employed in a wide variety of computer applications. Efficient algorithms and data structures were developed for selected problems in a number of key areas (eg computational geometry, graphs and networks, and combinatorial optimisation), and their complexity assessed by discrete mathematical, analytic and experimental techniques.
Probabilistic methods were also explored to analyse the performance of algorithms and to design algorithms with good expected runtimes. The research extends to fundamental problems in parallel computing and distributed processing, and addresses suitable models of processor or computer interconnection networks and communication protocols, with due attention to invariant-based design methods.
PROGRESS AND RESULTS - STATUS AS OF OCTOBER 1991
The basic research network has been realised. The Action's news-bulletin (Algorithms Review) will be upgraded to an international journal after the end of the Action. Results delivered in Year 2 include:
-Data structures: new techniques for maintaining sets of geometric objects for efficient query answering, algorithms for computing obstacle-free shortest paths, advances in persistent datastructures, efficient on-line algorithms for maintaining component s in dynamic graphs, and a new release of the algorithms library LEDA.
-Computational geometry: new algorithms for computing Voronoi diagrams and Delaunay triangulations, output-sensitive algorithms for hidden-surface removal for polytopes, advances in scene analysis, and randomised incremental constructions.
-Graph algorithms: advances in graph-layout and graph drawing problems, efficient on-line algorithms for finding shortest paths in planar graphs, advances in the analysis of on-line algorithms for hypergraphs, and new algorithms for deciding k-connectivi ty.
-Probabilistic methods and average-case analysis: performance of unification algorithms and of search problems on tree-structured data, analysis of (parallel) algorithms for random graphs, probabilistic analysis of a geometric lattice reduction algorithm , and a new release of the LUO system for automatic analysis.
-Complexity theory: new bounds for arithmetic circuits, results for computational learning and strategy determination, complexity results for local search algorithms, advances in the study of polynominal complexity classes and the complexity of algebraic algorithms.
-Parallel algorithms: advances in the study of P-complete and NC-problems (viz. in graph theory and optimisation), processor-efficient algorithms (eg for computational geometry), efficient expected-case parallel algorithms (eg for graph coloring, randomi sed integer sorting, and generating random permutations), and algorithms for optimal grid-routing.
-Distributed algorithms: analysis of adaptive routing algorithms, new techniques for routing with compact tables, new algorithms for atomic registers, results for distributed datastructuring, and design methodologies for some classes of distributed contr ol algorithms (eg termination detection) and protocols.
POTENTIAL
ALCOM has proved instrumental in creating a European platform for algorithms research. Ultimately this will lead to better training of designers of highly efficient computer applications. Both the LEDA program-library and the LUO automatic analyser make advanced algorithms and analysis techniques available for wider use. Many algorithms developed in the action deal with fundamental discretised problems and can ultimately be applied in complex software systems.

Ámbito científico

Programa(s)

Tema(s)

Data not available

Convocatoria de propuestas

Data not available

Régimen de financiación

Data not available

Coordinador

UNIVERSITEIT VAN UTRECHT
Aportación de la UE
Sin datos
Dirección
PADUALAAN, 8
3584 CH UTRECHT
Países Bajos

Ver en el mapa
Coste total
Sin datos

Participantes (11)

Compartir esta página

Descargar

Última actualización: 19 Agosto 1994

Mi folleto

Permalink: https://cordis.europa.eu/project/id/3075/es

European Union, 2024