Prethodno priopćenje
Performance Issues And Gains Of Caching The Pathfinding Data
Ivan Porkolab
; Visoko učilište Algebra, Zagreb, Croatia
Goran Đambić
orcid.org/0000-0001-5790-7701
; Visoko učilište Algebra, Zagreb, Croatia
Danijel Kučak
orcid.org/0000-0003-2947-2389
; Visoko učilište Algebra, Zagreb, Croatia
Sažetak
Using non-cached methods for finding the shortest path between nodes is the most common case when using pathfinding systems. That approach generates a couple of issues. Foremost, it has a significant impact on processing resources as calculations must be done over again for each iteration, even for the repeating events. That’s not a big concern if pathfinding is invoked a reasonable number of times or the nodes involved are always different, but if pathfinding occurs many times on the same nodes, then the caching of once calculated path becomes an acceptable course of action. This paper has explored one of such caching algorithms, FAST-N algorithm and compared it with standard non-cached pathfinding. Doing so, it outlined margins of justifiable use of such systems.
On a small number of pathfinding requests or simple node structure, because of increase in memory usage and rather hefty initial calculation processing requirements, it has been concluded that non-cached system makes more sense than cached one. On the other hand, when confronted with a large number of pathfinding requests and more complex node structure, caching can generate significant benefits concerning processing power and speed.
Ključne riječi
pathfinding; caching; FAST-N; C#
Hrčak ID:
216171
URI
Datum izdavanja:
28.11.2018.
Posjeta: 859 *