isochrone issueshttps://git.uibk.ac.at/informatik/dbis/dbis-isochrone/isochrone/-/issues2017-11-08T22:55:25+01:00https://git.uibk.ac.at/informatik/dbis/dbis-isochrone/isochrone/-/issues/8Implement algorithm using GeoHex approach2017-11-08T22:55:25+01:00User expirednikolaus.krismer@uibk.ac.atImplement algorithm using GeoHex approachThe MineT(X) algorithm uses ranges that can be represented by tiles.
If we want to use an apprach that combines the circular ranges (representing reachability a bit better than tiles) and tile ranges (not overlapping and therefore easier...The MineT(X) algorithm uses ranges that can be represented by tiles.
If we want to use an apprach that combines the circular ranges (representing reachability a bit better than tiles) and tile ranges (not overlapping and therefore easier to compute; using spatial indices better) then GeoHex could be the way to go.
So just changing the shape could lead to a new algorithm MineGX :-)
I guess that the performance of such an approach is somewhere in the middle between MineRX and MineTX, but that is only a guess. Maybe it is even better than MineTX :-)
What remains unsolved at the moment is the use of adaptive GeoHex sizes, since the hexagons of different zoom levels overlap with the ones of the next/previous level (in contrast to the tile regions used by MineTX).https://git.uibk.ac.at/informatik/dbis/dbis-isochrone/isochrone/-/issues/3Update evaluation results for paper "Isochrones in Multimodal Spatial Network"2018-05-21T12:07:32+02:00User expirednikolaus.krismer@uibk.ac.atUpdate evaluation results for paper "Isochrones in Multimodal Spatial Network"Johann Gamper is working on the paper "Isochrones in Multimodal Spatial Networks" (together with M. Böhlen, M. Innerebner and of course myself :-))
We need to re-compute the experiments for evaluation, since the datasets used by Marku...Johann Gamper is working on the paper "Isochrones in Multimodal Spatial Networks" (together with M. Böhlen, M. Innerebner and of course myself :-))
We need to re-compute the experiments for evaluation, since the datasets used by Markus Innerebner had some problems with projections and the algorithms also were quite buggy.
Now that these problems have been solved (datasets were updated during this month) the evaluation could lead to some slightly different results.