摘要：Many historic wood building areas are susceptible to fire hazards. The location of fire stations shall be carefully selected to provide adequate fire protection. In this study, an innovative procedure to determine optimal fire station locations is proposed considering the unique fire development pattern from different possible fire origins and the potential loss from both daily fires and post-earthquake fires (PEF). The fire spread processes originated from different buildings are firstly simulated using a physics-based fire spread analysis; and then different optimization objectives are selected to reflect the decision makers’ attitudes towards the two fire risks. A non-dominated sorting genetic algorithm (NSGA-II) is adopted to obtain the Pareto-optimal solutions, i.e., the locations of fire stations corresponding to the minimum average burnt loss for daily fire, and the minimum probability of exceeding an acceptable fire loss from PEF, respectively. The approach is applied to an ancient town in Southwest China for illustration. The proposed method can reduce the fire loss by 5% on average compared with the traditional max covering location method, and the impacts of uncertainties in the estimate of time for fire brigade to arrive and the change of unacceptable fire loss on the optimal locations of fire stations are discussed.
