The most of existing bus systems incorporate fixed routes via bus terminals and fixed number of vehicles depending on timetables. Such static bus systems cannot be adapted to the dynamic changes of demand flows such as frequency and directions of transport demands. On the other hand, in the most of existing taxi systems, the transport routines of taxis depend on the empirical knowledge of taxi drivers. However, the knowledge is not exploited effectively to improve the system performance. Therefore, we propose a new adaptive system which enables dynamic routine patterns for transport vehicles by learning the flows of transport demands. Our idea is inspired by a heuristic algorithm called "ant colony system". Each vehicle leaves clues to construct dynamic routine patterns by storing its delivery history. The clues consist of two kinds of pheromones: "pick-up pheromone" which leads the finding of new customers and "delivery pheromone" which leads the constituting of routines. Finally, we report simulation results by using two flow patterns of transport demands.
Dynamic construction of routine patterns for transport vehicles based on ant colony system
Titel japanisch
Transactions of the Information Processing Society of Japan ; 46 , 11 ; 2687-2694
2005
8 Seiten, 12 Quellen
Article (Journal)
Japanese
TEMPORARILY ALLOCATING FIX PUBLIC TRANSPORT VEHICLES AS DYNAMIC PUBLIC TRANSPORT VEHICLES
European Patent Office | 2020
|Dynamic priority for public transport vehicles
British Library Online Contents | 1997
Electric Vehicles Energy Efficient Routing Using Ant Colony Optimization
SAE Technical Papers | 2017
|Clustering-Based Modified Ant Colony Optimizer for Internet of Vehicles (CACOIOV)
BASE | 2019
|