Yen's K-Shortest Paths

Overview

Yen's algorithm finds the K shortest simple paths between a source node and a target node. Unlike Dijkstra's shortest path which returns only the single best path, Yen's algorithm returns multiple alternative paths ranked by cost, providing route diversity. This algorithm is commonly used to identify backup paths in case the primary path fails.

The algorithm was proposed by Jin Y. Yen in 1971:

• J.Y. Yen, Finding the K Shortest Loopless Paths in a Network (1971)

Concepts

K-Shortest Paths

Given a source and target, there may be many paths connecting them. Yen's algorithm finds the top K paths with the lowest cost, where each path contains no repeated nodes.

The algorithm works by:

1. Find the shortest path from source to target using Dijkstra's algorithm. This becomes the 1st shortest path.
2. For each previously found path, consider every node along it as a potential spur node (deviation point). At each spur node:
   • Temporarily remove edges that overlap with previously found paths at this spur node, to force a different route.
   • Find the shortest path from the spur node to the target using Dijkstra's algorithm. This is the spur path.
   • Combine the portion of the original path from source to the spur node (root path) with the spur path to form a candidate path.
3. Among all candidate paths, select the one with the lowest cost as the next shortest path.
4. Repeat steps 2-3 untilK paths are found or no more candidates exist.

Considerations

• The algorithm only finds simple paths (no cycles).
• All edge weights must be non-negative.

Example Graph

GQL
INSERT (A:default {_id: "A"}), (B:default {_id: "B"}),
       (C:default {_id: "C"}), (D:default {_id: "D"}),
       (E:default {_id: "E"}), (F:default {_id: "F"}),
       (G:default {_id: "G"}),
       (A)-[:default {value: 2}]->(B), (A)-[:default {value: 4}]->(F),
       (B)-[:default {value: 3}]->(C), (B)-[:default {value: 3}]->(D),
       (B)-[:default {value: 6}]->(F), (D)-[:default {value: 2}]->(E),
       (D)-[:default {value: 2}]->(F), (E)-[:default {value: 3}]->(G),
       (F)-[:default {value: 1}]->(E)

Parameters

Run Mode

Returns:

GQL
CALL algo.yens({
  source: "A",
  target: "G",
  k: 3,
  weight: "value"
}) YIELD path, cost, rank

Result:

Stream Mode

Returns the same columns as run mode, streamed for memory efficiency.

GQL
CALL algo.yens.stream({
  source: "A",
  target: "G",
  k: 3,
  weight: "value"
}) YIELD path, cost, rank
RETURN path, cost, rank

Stats Mode

Returns:

GQL
CALL algo.yens.stats({
  source: "A",
  target: "G",
  k: 3,
  weight: "value"
}) YIELD pathsFound, minCost, maxCost

Result: