Proactive Control for Oversaturation Mitigation on Evacuation Network: a Multi-Agent Simulation Approach

Abstract

Using a multi-agent simulation approach composing of evacuee, cell, and signal, this study aims to proactively curb the oversaturation spread in evacuation and improve the evacuation efficiency. The specific innovation lies in the en-route path choice model and oversaturation control model. A logit model incorporating fuzzy multi-attribute utility is used to describe the path choice. A three-threshold throttling strategy is proposed to assign signal parameters for oversaturated intersections. A case study demonstrates that the multi-attribute utility performs effectively in the presentation of uncertain route-choice factors under evacuation background. The throttling strategy can decrease the total evacuation time by 7.6% in the supposed scenario.