Study on Data Storage Particle Size Optimization of Traffic Information Database for Floating Car Systems Based on Minimum Description Length Principle

Abstract

Particle size partition plays a key role in the optimization of historical database precision and data storage space. When establishing the historical database of traffic information for floating car systems, proper size of data storage particle can optimize data precision and storage space simultaneously and gives minimum comprehensive cost. This paper proposes a data storage particle size optimization model, which tries to balance data precision as well as data storage space for floating car systems. Furthermore, the proposed data storage particle size optimization model is executed in Beijing case study. The results show that the data storage particle size is 35 min at night while 10 min in the day under the given constraints of minimum cost of data precision and storage space, which is consistent with the real traffic condition and application requirement.