An implementation of a Scapegoat Tree, a self-balancing binary search tree. Unlike other balanced trees such as Red-Black or AVL trees, the Scapegoat Tree does not require storing extra information (like colors or heights) in every node, making it exceptionally memory-efficient.

This project was developed for the Algorithms course (A.Y. 2020/2021) at the University of Catania.

• Self-Balancing Logic: Employs the "scapegoat" concept to rebuild unbalanced subtrees only when necessary.
• Performance: Guarantees a worst-case amortized time complexity of $O(\log n)$ for insertion and deletion operations.
• Memory Efficient: Minimal node structure without the overhead of balancing metadata.

The project is organized as follows:

• Scapegoat-Tree.cpp: Complete C++ source code of the data structure.
• Scapelgoat tree documentation.pdf: Detailed technical report including complexity analysis and performance testing.
• Scapelgoat tree documentation.zip: Source files for the documentation (LaTeX/Assets).

A Scapegoat Tree relies on a balancing parameter $\alpha$ (typically ranging from $0.5$ to $1$). A node is considered unbalanced if:

$$\text{size}(\text{child}) > \alpha \cdot \text{size}(\text{node})$$

When this condition is violated during an insertion, the algorithm traverses up the tree to find the "scapegoat" (the first unbalanced ancestor) and rebuilds that specific subtree into a perfectly balanced state.

To compile and run the project:

g++ -O3 Scapegoat-Tree.cpp -o ScapegoatTree
./ScapegoatTree