Abstract

Accurate and timely forecasting of traffic volume has long been regarded as a key point in transportation, planning and management. In order to realize effective and efficient traffic forecasting, this paper investigates quantificational method of dynamic factors from the perspective of nearness. The dynamic modeling method based on quantificational dynamics (that is, quantificational disposal of dynamic factors according to nearness) is proposed and this method can significantly improve the forecast effectiveness and efficiency. Swarm simulation is adopted as a new tool with regard to the field of traffic forecasting for analysis and verification. The testing results show that the proposed method outperforms traditional ones in choosing training samples and constituting forecasting models. This work contributes to the consideration and evaluation of dynamic factors in scientific forecasting and may bring some enlightenment to relevant scientific researchers and engineers.

Key words: Traffic forecasting, dynamic factors, quantificational dynamics, nearness, swarm simulation.