Traffic Signal Control with Partial Grade Separation for Oversaturated Conditions.

摘要

Increasing individual vehicular traffic is a major concern all around the world. This leads to oversaturation of intersections in no time. Traffic signal control under oversaturated condition is a long lasting challenge. To address this challenge thoroughly, this thesis introduces grade separation at signalized intersections. A lane-based optimization model is developed for the integrated design of interchange ramps (e.g. tunnels), lane markings (e.g. left turns, through traffic, right turns, etc.) and signal timing settings. We take into account two types of lane configurations. One is conventional surface lanes controlled by signals, and the other is grade separated lanes. This problem is formulated as a Mixed Integer Linear Program (MILP), and this can be solved using regular branch in branch out methods. The integer decision variables help in finding if the movement is on grade separated or surface lanes, and also the successor functions to govern the order of signal display. The continuous variables include the assigned lane flow, common flow multiplier, cycle length, and start and duration of green for traffic movements and lanes. The optimized signal time settings and lane configurations are then represented in Vissim simulation. Numerical examples, along with a benefit cost analysis show the great effectiveness of the proposed optimization model.