Incorporated Dempster-Shafer Theory, MACONT, and e-STEP Method (DSM-eSTEP) for Multicriteria Tradeoff Analysis in Transportation Budget Allocation

Abstract

This study introduces a multicriteria tradeoff analysis method based on Dempster-Shafer (DS) theory in conjunction with the Mixed Aggregation by Comprehensive Normalization Technique (MACONT) and the e-STEP method, termed DSM-eSTEP method, to assist transportation agencies in making optimal investment decisions under alternative budget allocation scenarios. Specifically, the DS evidence theory is employed to develop belief functions from preferences of multiple transportation performance criteria revealed by transportation decision-makers. The belief functions are transformed to probability functions using the pignistic transformation technique, leading to a set of initial weights that could reflect the priorities of decision-makers on the performance criteria. To enhance the robustness of relative weights assigned, the MACONT technique and the entropy measure are employed to derive refined weights via context-dependent adjustments. Having established the refined weights of transportation performance measures, the multi-objective budget allocation formulation according to various tradeoff scenarios could be converted to a linear programming model readily solvable for optimality. An empirical study is conducted for optimal budget allocation of 233 highway bridge and pavement preservation projects proposed for a U.S. state-maintained urban Interstate highway network. Cross comparisons of budget allocation results are made between the proposed method and the widely used compromise programming (CP) method. It shows that the proposed DSM-eSTEP method could generate efficient investment outcomes and slightly outperform the CP method.

Keywords

• Dempster-Shafer theory
• e-STEP method
• multicriteria
• transportation
• budget allocation