When we have red-black trees and AVL trees that can perform most of operations in logarithmic times, then what is the need for splay trees?

(a) no there is no special usage

(b) In real time it is estimated that 80% access is only to 20% data, hence most used ones must be easily available

(c) redblack and avl are not upto mark

(d) they are just another type of self balancing binary search trees

My question comes from Splay Tree in portion Binary Trees of Data Structures & Algorithms I

I had been asked this question in exam.

Right option is (b) In real time it is estimated that 80% access is only to 20% data, hence most used ones must be easily available

Easy explanation – May be the stats showing 80-20% may be not accurate, but in real time that is the widely spread scenario seen. If you are into this type of situation, you must choose implementing splay trees.