| Summary: | This paper proposes a new network model for the building evacuation problem considering congestion levels and provides a mixed integer linear programming (MILP) model and an efficient heuristic algorithm solving the problem. Constructing an optimization model with several congestion levels, we introduce a new network called the multi-class time-expanded (MCTE) network having several exclusive arcs connecting the same tail and head nodes. The MCTE networks make both the MILP model and the heuristic algorithm reflect a realistic situation in congested networks. Considering MCTE networks makes the problem difficult to solve, which motivates us to develop an efficient heuristic algorithm. We test our heuristic algorithm using several real-world networks such as a multiplex cinema, a subway station, and a large-size complex shopping mall in addition to an artificial network for clear comparison between the proposed algorithm and the MILP approaches. The results indicate that the proposed algorithm runs fast and produces a near-optimal solution compared with those from MILP models with a commercial solver.
|
|---|
