Abstract:

A bilevel model was proposed for the bus route network optimization. The upperlevel model takes the maximization of the through passenger flow density as the target, optimizes the bus route network based on the passenger flow density as the target, optimizes the bus route network based on the passenger flow distribution over bus stops. Considering the fact that the bus passenger travel behavior changes along with the change of the bus route network, the lowerlevel model assigns the bus passenger flow distribution again in the network optimized by the upperlevel model. The result of the passenger flow assignment(new passenger flow matrix) can also affect the bus route network design. An interative approach including an improved ant colony algorithm and a labelmarking method was used to solve the bilevel model. The model and the related algorithms were tested using the bus traffic data collected in the central area of Dalian city as a case study.

Key words: engineering of communications and transportation, bilevel model, bus route network optimization, through passenger flow density, passenger flow assignment, ant colony algorithm