6. Shortest Path Search

pgRouting was first called pgDijkstra, because it implemented only shortest path search with Dijkstra algorithm. Later other functions were added and the library was renamed.

This chapter will explain the three different shortest path algorithms and which attributes are required.

Note

If you run osm2pgrouting tool to import OpenStreetMap data, the ways table contains all attributes already to run all shortest path functions. The ways table of the pgrouting-workshop database of the previous chapter is missing several attributes instead, which are listed as Prerequisites in this chapter.

6.1. Dijkstra

Dijkstra algorithm was the first algorithm implemented in pgRouting. It doesn’t require other attributes than source and target ID, id attribute and cost. It can distinguish between directed and undirected graphs. You can specify if your network has reverse cost or not.

Prerequisites

To be able to use reverse cost you need to add an additional cost column. We can set reverse cost as length.

ALTER TABLE ways ADD COLUMN reverse_cost double precision;
UPDATE ways SET reverse_cost = length;

Function with parameters

shortest_path( sql text,
                   source_id integer,
                   target_id integer,
                   directed boolean,
                   has_reverse_cost boolean )

Note

• Source and target IDs are vertex IDs.
• Undirected graphs (“directed false”) ignore “has_reverse_cost” setting

6.1.1. Core

Each algorithm has its core function , which is the base for its wrapper functions.

SELECT * FROM shortest_path('
                SELECT gid as id,
                         source::integer,
                         target::integer,
                         length::double precision as cost
                        FROM ways',
                5700, 6733, false, false);

 vertex_id | edge_id |        cost
-----------+---------+---------------------
      5700 |    6585 |   0.175725539559539
      5701 |   18947 |   0.178145491343884
      2633 |   18948 |   0.177501253416424
       ... |     ... |                 ...
      6733 |      -1 |                   0
 (38 rows)

6.1.2. Wrapper

Wrapper WITHOUT bounding box

Wrapper functions extend the core functions with transformations, bounding box limitations, etc.. Wrappers can change the format and ordering of the result. They often set default function parameters and make the usage of pgRouting more simple.

SELECT gid, AsText(the_geom) AS the_geom
        FROM dijkstra_sp('ways', 5700, 6733);

  gid   |                              the_geom
--------+---------------------------------------------------------------
   5534 | MULTILINESTRING((-104.9993415 39.7423284, ... ,-104.9999815 39.7444843))
   5535 | MULTILINESTRING((-104.9999815 39.7444843, ... ,-105.0001355 39.7457581))
   5536 | MULTILINESTRING((-105.0001355 39.7457581,-105.0002133 39.7459024))
    ... | ...
  19914 | MULTILINESTRING((-104.9981408 39.7320938,-104.9981194 39.7305074))
(37 rows)

Note

It’s possible to show the route in QGIS. It works for shortest path queries that return a geometry column.

• Create a database connection and add the “ways” table as a background layer.
• Add another layer of the “ways” table but select Build query before adding it.
• Type "gid"  IN ( SELECT gid FROM dijkstra_sp('ways',5700,6733)) into the SQL where clause field.

SQL query can be also selected from the layer context menu.

Wrapper WITH bounding box

You can limit your search area by adding a bounding box. This will improve performance especially for large networks.

SELECT gid, AsText(the_geom) AS the_geom
        FROM dijkstra_sp_delta('ways', 5700, 6733, 0.1);

  gid   |                              the_geom
--------+---------------------------------------------------------------
   5534 | MULTILINESTRING((-104.9993415 39.7423284, ... ,-104.9999815 39.7444843))
   5535 | MULTILINESTRING((-104.9999815 39.7444843, ... ,-105.0001355 39.7457581))
   5536 | MULTILINESTRING((-105.0001355 39.7457581,-105.0002133 39.7459024))
    ... | ...
  19914 | MULTILINESTRING((-104.9981408 39.7320938,-104.9981194 39.7305074))
(37 rows)

Note

The projection of OSM data is “degree”, so we set a bounding box containing start and end vertex plus a 0.1 degree buffer for example.

6.2. A-Star

A-Star algorithm is another well-known routing algorithm. It adds geographical information to source and target of each network link. This enables the shortest path search to prefer links which are closer to the target of the search.

Prerequisites

For A-Star you need to prepare your network table and add latitute/longitude columns (x1, y1 and x2, y2) and calculate their values.

ALTER TABLE ways ADD COLUMN x1 double precision;
ALTER TABLE ways ADD COLUMN y1 double precision;
ALTER TABLE ways ADD COLUMN x2 double precision;
ALTER TABLE ways ADD COLUMN y2 double precision;

UPDATE ways SET x1 = x(ST_startpoint(the_geom));
UPDATE ways SET y1 = y(ST_startpoint(the_geom));

UPDATE ways SET x2 = x(ST_endpoint(the_geom));
UPDATE ways SET y2 = y(ST_endpoint(the_geom));

UPDATE ways SET x1 = x(ST_PointN(the_geom, 1));
UPDATE ways SET y1 = y(ST_PointN(the_geom, 1));

UPDATE ways SET x2 = x(ST_PointN(the_geom, ST_NumPoints(the_geom)));
UPDATE ways SET y2 = y(ST_PointN(the_geom, ST_NumPoints(the_geom)));

Note

endpoint() function fails for some versions of PostgreSQL (ie. 8.2.5, 8.1.9). A workaround for that problem is using the PointN() function instead:

Function with parameters

Shortest Path A-Star function is very similar to the Dijkstra function, though it prefers links that are close to the target of the search. The heuristics of this search are predefined, so you need to recompile pgRouting if you want to make changes to the heuristic function itself.

shortest_path_astar( sql text,
                   source_id integer,
                   target_id integer,
                   directed boolean,
                   has_reverse_cost boolean )

Note

• Source and target IDs are vertex IDs.
• Undirected graphs (“directed false”) ignore “has_reverse_cost” setting

6.2.1. Core

SELECT * FROM shortest_path_astar('
                SELECT gid as id,
                         source::integer,
                         target::integer,
                         length::double precision as cost,
                         x1, y1, x2, y2
                        FROM ways',
                5700, 6733, false, false);

 vertex_id | edge_id |        cost
-----------+---------+---------------------
      5700 |    6585 |   0.175725539559539
      5701 |   18947 |   0.178145491343884
      2633 |   18948 |   0.177501253416424
       ... |     ... |                 ...
      6733 |      -1 |                   0
 (38 rows)

6.2.2. Wrapper

Wrapper function WITH bounding box

Wrapper functions extend the core functions with transformations, bounding box limitations, etc..

SELECT gid, AsText(the_geom) AS the_geom
        FROM astar_sp_delta('ways', 5700, 6733, 0.1);

  gid   |                              the_geom
--------+---------------------------------------------------------------
   5534 | MULTILINESTRING((-104.9993415 39.7423284, ... ,-104.9999815 39.7444843))
   5535 | MULTILINESTRING((-104.9999815 39.7444843, ... ,-105.0001355 39.7457581))
   5536 | MULTILINESTRING((-105.0001355 39.7457581,-105.0002133 39.7459024))
    ... | ...
  19914 | MULTILINESTRING((-104.9981408 39.7320938,-104.9981194 39.7305074))
(37 rows)

Note

• There is currently no wrapper function for A-Star without bounding box, since bounding boxes are very useful to increase performance. If you don’t need a bounding box Dijkstra will be enough anyway.
• The projection of OSM data is “degree”, so we set a bounding box containing start and end vertex plus a 0.1 degree buffer for example.

6.3. Shooting-Star

Shooting-Star algorithm is the latest of pgRouting shortest path algorithms. Its speciality is that it routes from link to link, not from vertex to vertex as Dijkstra and A-Star algorithms do. This makes it possible to define relations between links for example, and it solves some other vertex-based algorithm issues like “parallel links”, which have same source and target but different costs.

Prerequisites

For Shooting-Star you need to prepare your network table and add the rule and to_cost column. Like A-Star this algorithm also has a heuristic function, which prefers links closer to the target of the search.

-- Add rule and to_cost column
ALTER TABLE ways ADD COLUMN to_cost double precision;
ALTER TABLE ways ADD COLUMN rule text;

Shooting-Star algorithm introduces two new attributes

Attribute	Description
rule	a string with a comma separated list of edge IDs, which describes a rule for turning restriction (if you came along these edges, you can pass through the current one only with the cost stated in to_cost column)
to_cost	a cost of a restricted passage (can be very high in a case of turn restriction or comparable with an edge cost in a case of traffic light)

Function with parameters

shortest_path_shooting_star( sql text,
                   source_id integer,
                   target_id integer,
                   directed boolean,
                   has_reverse_cost boolean )

Note

• Source and target IDs are link IDs.
• Undirected graphs (“directed false”) ignores “has_reverse_cost” setting

To describe turn restrictions:

 gid | source | target | cost | x1 | y1 | x2 | y2 | to_cost | rule
-----+--------+--------+------+----+----+----+----+---------+------
  12 |      3 |     10 |    2 |  4 |  3 |  4 |  5 |    1000 | 14

... means that the cost of going from edge 14 to edge 12 is 1000, and

 gid | source | target | cost | x1 | y1 | x2 | y2 | to_cost | rule
-----+--------+--------+------+----+----+----+----+---------+------
  12 |      3 |     10 |    2 |  4 |  3 |  4 |  5 |    1000 | 14, 4

... means that the cost of going from edge 14 to edge 12 through edge 4 is 1000.

If you need multiple restrictions for a given edge then you have to add multiple records for that edge each with a separate restriction.

 gid | source | target | cost | x1 | y1 | x2 | y2 | to_cost | rule
-----+--------+--------+------+----+----+----+----+---------+------
  11 |      3 |     10 |    2 |  4 |  3 |  4 |  5 |    1000 | 4
  11 |      3 |     10 |    2 |  4 |  3 |  4 |  5 |    1000 | 12

... means that the cost of going from either edge 4 or 12 to edge 11 is 1000. And then you always need to order your data by gid when you load it to a shortest path function..

6.3.1. Core

An example of a Shooting Star query may look like this:

SELECT * FROM shortest_path_shooting_star('
                SELECT gid as id,
                         source::integer,
                         target::integer,
                         length::double precision as cost,
                         x1, y1, x2, y2,
                         rule, to_cost
                        FROM ways',
                6585, 8247, false, false);

 vertex_id | edge_id |        cost
-----------+---------+---------------------
     15007 |    6585 |   0.175725539559539
     15009 |   18947 |   0.178145491343884
      9254 |   18948 |   0.177501253416424
       ... |     ... |   ...
      1161 |    8247 |   0.051155648874288
 (37 rows)

Warning

Shooting Star algorithm calculates a path from edge to edge (not from vertex to vertex). Column vertex_id contains start vertex of an edge from column edge_id.

6.3.2. Wrapper

Wrapper functions extend the core functions with transformations, bounding box limitations, etc..

SELECT gid, AsText(the_geom) AS the_geom
        FROM shootingstar_sp('ways', 6585, 8247, 0.1, 'length', true, true);

  gid   |                              the_geom
--------+---------------------------------------------------------------
   6585 | MULTILINESTRING((-104.9975345 39.7193508,-104.9975487 39.7209311))
  18947 | MULTILINESTRING((-104.9975487 39.7209311,-104.9975509 39.7225332))
  18948 | MULTILINESTRING((-104.9975509 39.7225332,-104.9975447 39.7241295))
    ... | ...
   8247 | MULTILINESTRING((-104.9978555 39.7495627,-104.9982781 39.7498884))
(37 rows)

Note

There is currently no wrapper function for Shooting-Star without bounding box, since bounding boxes are very useful to increase performance.

Warning

The projection of OSM data is “degree”, so we set a bounding box containing start and end vertex plus a 0.1 degree buffer for example.