In this paper we propose using experiments with Balanced-Sample Binary Trees (BSBTrees) as assignments and lecture material in intermediate data structures courses (CS2/3). BSBTrees are composite data structures that have a temporarily constructed form that precedes their normal construction. We present them in the context of binary search trees. To do this we first investigate the retrieval properties of randomly generated binary search trees and show how temporary construction can improve both worst case and average case behavior. We provide a brief analysis of BSBTree performance and description of the classes that can be used for BSBTree implementation. Last we discuss the use of BSBTrees in CS2 and CS3 courses and a survey of student opinions about BSBTrees.
[1]
Conrado Martínez,et al.
Randomized binary search trees
,
1998,
JACM.
[2]
Gaston H. Gonnet,et al.
Handbook Of Algorithms And Data Structures
,
1984
.
[3]
Luc Devroye,et al.
A note on the height of binary search trees
,
1986,
JACM.
[4]
William A. Martin,et al.
Optimizing binary trees grown with a sorting algorithm
,
1972,
CACM.
[5]
DONALD MICHIE,et al.
“Memo” Functions and Machine Learning
,
1968,
Nature.
[6]
Niklaus Wirth,et al.
Algorithms + Data Structures = Programs
,
1976
.