Pathfinding

1
Pathfinding
2

• PathEngine is a sophisticated middleware
  tool-kit for the implementation of intelligent
  agent movement, built around an advanced
  implementation of points-of-visibility
  pathfinding on 3D ground surfaces.
• www.pathengine.com/overview

3
SDK Licensing options
4
Cheaper Options

• Breadth-First Search
• Dijkstras Algorithm
• Best-First Search
• A Search

5
Breadth-First Search

• Examines all paths of cost 1, then 2, and so on
  until it finds the goal.
• Guaranteed to find the shortest path
• Very slow

6
Dijkstras Algorithm

• Selects closest unprocessed node from the start,
  and updates its value, and that of its neighbors.
• Nodes are valued by the distance from the start.

7
Best-First Search

• Similar to Dijkstras Algorithm, but uses a
  heuristic to determine the value of node.
• Main disadvantage is that is does not account
  well for terrain that is slower.

8
A Search

• Very similar to a Breadth-First Search.
• Orders nodes to search not just by distance from
  start node, but also by a heuristic or a
  estimate cost remaining to the end point.

9
A Search

• The cost of a node is defined asf(n) g(n)
  h(n) whereg(n) is the distance
  traveled from the starth(n) is the heuristic,
  the estimated distance to the
  end node

10
A Heuristic

• The heuristic is an estimate for the remaining
  cost to the end point
• This estimate should be a close as possible to
  the actual cost, but never higher.
• For a grid of squares, the heuristic is usually
  the number of squares in between the current node
  and the end node.(note this is different from
  the Euclidean distance)

11
A Search Examples
12
A Considerations

• You must define how diagonal squares work with
  map grids. If movement is continuous, consider
  multiplying the cost of diagonal movement by v2
• Defining an upper bound cutoff can prevent the
  search from taking too long.

13
References

• Path Enginehttp//www.pathengine.com/index.php
• Smart Moves Intelligent Path-Finding
  http//www.gamasutra.com/features/19990212/sm_01.
  htm
• Ogre Articles and Resourceshttp//www.ogre3d.org/
  wiki/index.php?titleArticlesAndSourcesredirectn
  o
• Pathfinding using A-algorithmhttp//www.cs.helsi
  nki.fi/u/apipulkk/astar/index.html