Work

Barrier avoidance algorithms

Year: 2003
Type: dissertation
Abstract: GIS (Geographical Information Systems) bring an ever increasing range of solutions to decision makers seeking ever more accurate situational awareness in the issues they are facing. In particular, a GIS can provide a broad range of visual information resulting from computations that may be complex. These require to develop and implement efficient algorithms because of the massive amount of computations or data that are induced by the geographical nature of the processed data. The Urban and Rural Planning and Study Lab of University of Liege seeks to compute accessibility profiles for so-called "slow transport modes" in the Walloon Region. In this context, a relevant issue of computation of a socio-economical index based on spatial and population data is considered. The objective of this work is to modelize a problem of computation accessibility index, then solve it. The central part consists of the developement of a consistent and efficient algorithm chain to compute optimal shortest paths in geographical zones by avoiding barriers. The integration of the involved algorithms as well as the exploitation of intermediary results are fundemental aspects of the proposed solution. Two variants of the problem are modelized: DLOSP (Discrete Line-Of-Sight Paths, that does not take barriers into account) and DGSP (Discrete Geodesic Shortest Paths, that takes barriers into account). The detection and indexation of the barrieres, the refitting of the considered search space, the optimization of visibility requests and the computation of shortest paths constitute the main steps of the proposed algorithm chain. A distributed computing setup, naturally stemming from the structure of the problem itself, is briefly presented.
Language: English
Open Access status: green
OpenAlex ID: W6989411007
Cites: 11
Cited by:
Related to: 10