MULTIOBJECTIVE DECISION-MAKING FOR TRAFFIC ASSIGNMENT

Abstract

In traditional traffic-assignment problems, only a single objective is considered. Although people's living quality improves and transportation needs increase, the environmental quality is destroyed with rapid economic and traffic growth. The concept of traffic assignment, an important procedure for transportation planning, should take all needs into consideration. Based on the viewpoints of all constituencies, this paper attempts to determine optimal flow patterns using three objectives-total travel time for road users, air pollution for nonusers (such as community residents), and travel distance for government-to formulate an effective multiobjective model for traffic assignment. By using multiobjective decision making and nonlinear programming techniques, a series of noninferior solutions is generated. By combining an eigenvector weighting method with pairwise comparison, we obtain a compromise solution for the flow pattern. An application case for the Taipei network system is also presented and discussed. The results show that if other nontraffic related factors are taken into account, the multiobjective traffic-assignment approach is more reasonable and suitable than conventional approaches.