Multi-objectives optimization of evacuation path based on improved matrix translation model and exit balance algorithm

Emergency evacuation planning requires balancing multiple objectives like minimizing time, avoiding hazards, and ensuring fairness. Traditional methods struggle to strike a balance between macroscopic efficiency and microscopic realism. This study proposes a new multi-objective optimization framework based on an improved Matrix Translation Model (MTM) and Exit Balance Algorithm (EBA): the improved MTM efficiently simulates the evacuation process and obtains individual objectives, thereby deriving group evacuation time objective $f_t$, detour objective $f_d$; crowding objective $f_c$, injury objective $f_i$ and fatality objective $f_f$. $f_d$, $f_c$, $f_i$ and $f_f$ are converted into penalty terms for $f_t$, and the improved EBA algorithm balances evacuation times across different exits to solve the multi-objective problem. This framework ensures precise statistical analysis of individual evacuation parameters while guaranteeing that each iteration moves closer to the optimal solution, enabling rapid convergence. Optimization results from a scenario with 2 floors, 42 rooms, and 1688 evacuees demonstrate that the algorithm can complete the simulation within 8–15 s, and the evacuation time reduced by 16 % while controlling detour and crowding duration in the scenario without fire, and the cumulative injury probability cut by 42 % in the fire scenario. This work bridges macroscopic efficiency and microscopic realism, offering a practical solution for dynamic evacuation planning.