Using Decision Trees in Order to Determine Intersection Design Rules

Road planners frequently face the challenge to determine which intersection design provides the best traffic flow for a particular traffic demand. Many road design manuals provide guidelines for the design and evaluation of different intersection alternatives, however mostly refer to specialized software in which the performances of different design alternatives can be modelled. In a planning stage of the design process, such assessments are undesirable due to time and cost. There is a need for quick design rules which need limited input data. Although some of these rules exist, their usability is limited. In this paper the authors examine the possibilities to determine intersection design rules by Decision Tree (DT) methods which are trained with data generated by Highway Capacity Manual (HCM) 2010 intersection modelling. The models consider 24 intersection designs varying the main type (all-way stop controlled, two-way stop controlled, signalized and roundabout) and the number and configuration of the entering and exiting lanes. Traffic demand patterns are randomly generated for various sizes of the dataset (5,000 – 5,000,000 cases) represented by 38 (independent) demand variables. Different DT methods (Chi-squared Automatic Interaction Detector (CHAID), Classification and Regression Trees (CRT) and Quick, Unbiased, Efficient, Statistical Tree (QUEST)), options (splitting criteria, tree depth) and datasets are tested for their predictive accuracy. The DT models provide accuracy rates between 76% and 96%. The CRT methods seem the most promising, and a further analysis was made concerning the independent variable importance and the possibilities for reducing the trees complexity. An example is shown of a DT which provides straightforward design rules and an predictive accuracy of 85.5%.

• Supplemental Notes:
  • This paper was sponsored by TRB committee AFB10 Geometric Design. Alternate title: Using Decision Trees to Determine Intersection Design Rules
• Corporate Authors:

  Transportation Research Board

  500 Fifth Street, NW
  Washington, DC  United States  20001
• Authors:
  • Bezembinder, Erwin M
  • Wismans, Luc J J
  • Van Berkum, Eric C
• Conference:
  • Transportation Research Board 94th Annual Meeting
  • Location: Washington DC, United States
  • Date: 2015-1-11 to 2015-1-15
• Date: 2015

Language

• English

Media Info

• Media Type: Digital/other
• Features: Figures; References; Tables;
• Pagination: 15p
• Monograph Title: TRB 94th Annual Meeting Compendium of Papers

Subject/Index Terms

• TRT Terms: Computer models; Decision trees; Geometric design; Intersections; Traffic flow
• Subject Areas: Design; Highways; Planning and Forecasting; I20: Design and Planning of Transport Infrastructure; I72: Traffic and Transport Planning;

Filing Info

• Accession Number: 01555038
• Record Type: Publication
• Report/Paper Numbers: 15-0544
• Files: TRIS, TRB, ATRI
• Created Date: Feb 26 2015 10:03AM