The field of bandit algorithms and decision tree optimization is rapidly evolving, with a focus on developing more efficient and effective methods for sequential decision-making and tree pruning. Recent research has explored the use of stochastic bandit models, multi-armed bandits, and reinforcement learning techniques to improve decision-making in complex environments. Additionally, there has been a push towards optimizing decision trees, with techniques such as pruning and beam search being applied to improve tree performance. Notable papers in this area include: Stochastic Bandits for Crowdsourcing and Multi-Platform Autobidding, which proposes an algorithm with improved regret bounds. Multi-Armed Bandits-Based Optimization of Decision Trees, which introduces a new pruning approach using multi-armed bandits. A Generic Complete Anytime Beam Search for Optimal Decision Tree, which presents a new beam search algorithm for decision tree optimization. Near-Optimal Regret for Efficient Stochastic Combinatorial Semi-Bandits, which achieves improved regret bounds for combinatorial semi-bandits. Regret minimization in Linear Bandits with offline data via extended D-optimal exploration, which proposes an algorithm that effectively incorporates offline data to reduce online regret. Non-Stationary Restless Multi-Armed Bandits with Provable Guarantee, which establishes a theoretical framework for non-stationary RMAB problems.