Metro Map Layout Using Multicriteria Optimization
IV '04 Proceedings of the Information Visualisation, Eighth International Conference
An ILP for the metro-line crossing problem
CATS '08 Proceedings of the fourteenth symposium on Computing: the Australasian theory - Volume 77
Automatic visualisation of metro maps
Journal of Visual Languages and Computing
Minimizing intra-edge crossings in wiring diagrams and public transportation maps
GD'06 Proceedings of the 14th international conference on Graph drawing
Line crossing minimization on metro maps
GD'07 Proceedings of the 15th international conference on Graph drawing
A mixed-integer program for drawing high-quality metro maps
GD'05 Proceedings of the 13th international conference on Graph Drawing
Path-based supports for hypergraphs
IWOCA'10 Proceedings of the 21st international conference on Combinatorial algorithms
Edge routing with ordered bundles
GD'11 Proceedings of the 19th international conference on Graph Drawing
Path-based supports for hypergraphs
Journal of Discrete Algorithms
Hi-index | 0.00 |
In the metro-line crossing minimization problem, we are given a plane graph G=(V,E) and a set $\mathcal{L}$ of simple paths (or lines) that coverG, that is, every edge e∈E belongs to at least one path in $\mathcal{L}$. The problem is to draw all paths in $\mathcal{L}$ along the edges of G such that the number of crossings between paths is minimized. This crossing minimization problem arises, for example, when drawing metro maps, in which multiple transport lines share parts of their routes. We present a new line-layout algorithm with $O(|\mathcal{L}|^2\cdot |V|)$ running time that improves the best previous algorithms for two variants of the metro-line crossing minimization problem in unrestricted plane graphs. For the first variant, in which the so-called periphery condition holds and terminus side assignments are given in the input, Asquith et al. [1] gave an $O(|\mathcal{L}|^3\cdot |E|^{2.5})$-time algorithm. For the second variant, in which all lines are paths between degree-1 vertices of G, Argyriou et al. [2] gave an $O((|E|+|\mathcal{L}|^2)\cdot |E|)$-time algorithm.