Dynamic Path Optimization and Personnel Scheduling for Building Emergency Evacuation

Abstract

Modern buildings feature complex layouts, high occupancy, and mixed functions, making indoor safety inspections critical for emergency evacuation. Traditional inspection strategies rely heavily on experience, often leading to unreasonable paths, missed checks, and delayed evacuation—especially in smoky, low-visibility environments. Existing studies lack an integrated framework that combines standardized sweep protocols, realistic constrained path modeling, and multi-story collaborative scheduling. This paper develops a fullchain emergency evacuation sweep system. For singlefloor scenarios, we model multi-responder cooperative inspection as an Open Multiple Traveling Salesman Problem with Permanent Forbidden Edges and Allowed Revisits (OMTSP-PFE-AR) and solve it using a greedy algorithm. For multi-story composite buildings, a bilevel optimization model is proposed to separate intrafloor scheduling from global personnel configuration. Experiments on a single-story office building and a threestory mixed-use building show that the proposed method achieves full inspection coverage with balanced workloads, shorter sweep times, and lower overall costs. The system adapts to both normal and heavy-smoke conditions, supports professional role matching, and improves the safety and efficiency of real-world building emergency evacuation operations.