Efisiensi Algoritma Pencarian Dalam Permainan Dan Navigasi Robot

Abstract

This article reviews the literature on the efficiency of algorithmic search and navigation in the domain of computer games and robot navigation. This study focuses on heuristic optimization methods such as bidirectional A*, sampling-based techniques, and Reinforcement-based algorithms. Statistics show that bidirectional A* routes reduce the long-distance time by 40% compared to conventional A* routes. Overall, although extreme parameter optimization is required, sampling- based algorithms definitely show the probability level up to a given value point. Reinforcement algorithm styles such as Deep Deterministic Policy Gradient and Proximal Policy Optimization have also been successful. By combining CNN, RNN, and LSTM, Reinforcement algorithms that work in dense environments enable robots to adapt to the environment, responding to human movement. Improved sample-based algorithms with boundary-based exploration have also proven to be effective and can be transferred from simulation to real environments.