New Methodology for Transportation Investment Decisions with Consideration of Project Interdependencies

Abstract

A new methodology for transportation project evaluation and selection explicitly addresses issues of networkwide impacts of a project and interdependencies of multiple projects. The methodology consists of two key elements: a multicommodity minimum-cost network model to estimate networkwide benefits of a single project or multiple interdependent projects and a hypergraph Knapsack model to identify the best subcollection of projects to yield maximized overall benefits. The proposed methodology is applied in a computational study with data on roadway network design, geometrics, travel demand, and traffic operations as well as the scope and costs of six major projects proposed for investment in the Loop area of Chicago, Illinois. The study reveals that the overall benefits with consideration of project networkwide impacts and their interdependencies are significantly lower by 38% to 64% as compared with those benefits established without consideration of project interdependencies, and the rate of increase in overall benefits follows a diminishing trend with higher budget levels; these findings suggest that an optimal level of investment could be established. The proposed methodology may be refined by developing a decision rule to identify a reasonably small number of subcollections of projects guaranteed to contain the optimal solution so that it could practically support large-scale transportation investment decisions.