Video games often make use of simulation to represent part of a real-world phenomenon; be it simulating a typical crowd behavior (e.g. chaos, rioting), or particle simulation (e.g. fire, smoke) and many other uses. Games have one common characteristic, i.e. they are interactive real-time systems, meaning to say processes that run in these applications must execute within a limited time threshold for the application to be called successful. The Boids algorithm is often used in these applications for realistic simulation of flocking type of behavior of virtual crowds. However, simulation of crowds in real-time using the algorithm is computationally time-consuming, due to how the algorithm evaluates the whole crowd when searching for possible nearest neighbors for each agent in the simulation. There are several approaches to improve the performance of these flocking simulations in real-time, and in this document, we discuss some of those methods that have been applied to the Boids Algorithm. We further implement and test one of these performance optimization methods, and use benchmarking results to compare the performance of the method versus the Boids Algorithms’ brute force neighborhood gathering approach