coldbubbletea / on-balanced-separators-in-road-networks

inertial-flow

Work flow:

1 finish Mincut(already 2023.06.28)

2 inertial flow(Start, but not runable,Coming Soon)

We will refer to Daniel J. H.'s code as his code has many advantages by using excellent std skills:

https://daniel-j-h.github.io/post/selection-algorithms-for-partitioning/

Q: how to dual with multi-source and multi-sink.

A: It is equivalent to single-source single-sink version.

Settings

b: source_sink_rate

According to paper, we will init b*|V| source nodes and also this amount of sink nodes. According to their experiment result, seems set default value of b as 0.25 is a good idea.

unit weight

As we just care about the coordinate and relation of edges(way of link), we can use unit weight.

data:

9th DIMACS Implementation Challenge - Shortest Paths:

http://www.diag.uniroma1.it/challenge9/download.shtml

It is a widely used dataset with many road networks. Every sub-folder, named after the graph, stores all the information of a sample graph. This is not the original data; I processed it to make it adaptable for my program's input.

USAGE:

Use Graph.addNode to add Node for a graph

we have 3 parameters for this: The first is node_id, it's type is int. The Second is latitude, it's type is double. The Third is longitude, it's type is double. Sample:

g.addNode(1, 40.730610, -73.935242);

In this sample we add a node with node_id 0 and lat:40.730610, lon:-73.935242, yes, this node coordinate is New York.