The B+-tree-based method for nearest neighbour queries in traffic simulation systems Online publication date: Wed, 24-Jun-2015
by Zhu Song; Shijie Zhou; Jiaqing Luo; Weiwei Deng
International Journal of High Performance Computing and Networking (IJHPCN), Vol. 8, No. 2, 2015
Abstract: Extensive used traffic simulation systems are helpful in planning and controlling the traffic system. In traffic simulation systems, the state of each vehicle is affected by that of nearby vehicles, called neighbours. Nearest neighbour (NN) queries, which are multi one-dimensional and highly concurrent, largely determine the performance of traffic simulation systems. Majority of existing traffic simulation systems use linked list-based methods to process NN queries. Although they are simple and effective, existing methods are neither scalable nor efficient. In this paper, we propose a B+-tree-based method to improve the efficiency of NN queries by borrowing ideas from methods used in databases. In particular, we create a linked local B+-tree, called LLB+-tree, which is a variation of original B+-tree, to maintain the index of neighbours of each vehicle. We also build a mathematical model to optimise the parameter setting of LLB+-tree according to multiple parameters for lanes and vehicles. Our theoretical analysis shows that the time complexity of the method is O(logN) under the assumption of random distribution of vehicles. Our simulation results show that LLB+-tree can outperform linked list and original B+-tree by 64.2% and 12.8%, respectively.
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of High Performance Computing and Networking (IJHPCN):
Login with your Inderscience username and password:
Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.
If you still need assistance, please email subs@inderscience.com