摘要
本文结合了机器学习和虚拟现实技术针对现有的寻路算法进行改进和研究,设计了一种适合自身特点的寻路算法,旨在弥补原有算法在系统应用中的不足。本文利用机器学习算法来寻路,相比于传统的A*寻路算法解决了局部最优的问题。当然A*算法本身存在的问题,本文也做了讨论和改进,来适应于项目的需求。本文设计的算法将人工神经网络与遗传算法结合,以达到寻找最佳适应度的目的。实验表明机器学习算法可以解决寻路问题并且用遗传算法训练优化网络权值在实时运算时减少计算量。针对火灾训练系统中复杂情况,可以达到动态寻路的效果。
This paper combines Machine Learning and virtual reality technology to improve and research the existing wayfinding algorithm, and designs a search algorithm suitable for its own characteristics, aiming at making up for the shortcomings of the original algorithm in system application. In this paper, the machine learning algorithm is used to find the path. Compared to the traditional A＊ Routing Algorithm, the problem of local optimal is solved. Of course, the A＊ Algorithm itself has a problem, this article has also been discussed and improved to adapt to the needs of the project. The algorithm designed in this paper combines Artificial Neural Network with Genetic Algorithm to achieve the goal of finding the best fitness. The experiment shows that the Machine Learning Algorithm can solve the path finding problem and the Genetic Algorithm is used to train the optimal network weights to reduce the amount of calculation in real time operation. In view of the complex situation in the fire training system, the effect of dynamic path finding can be achieved.
作者
顾雅涵
孙博文
李嘉明
田卫东
刘刚
GU Ya-han;SUN Bo-wen;LI Jia-ming;TIAN Wei-dong;LIU Gang(College of computer science and technology,Harbin University of Science and Technology,Harbin 150080,China)
出处
《软件》
2018年第4期86-91,共6页
Software
关键词
人工智能
机器学习
人工神经网络
遗传算法
A＊算法
Artificial intelligence
Machine learning
Artificial neural network
Genetic algorithm
A＊Routing algorithm
