Earthquake disaster management involves determining locations in which to construct shelters and how to allocate the affected population to them. A multi-objective, hierarchical mathematical model, allied with an interleaved modified particle swarm optimization algorithm and genetic algorithm (MPSO–GA), have been developed to solve the earthquake shelter location-allocation problem. From a set of candidate shelter locations, the model first determines which of these should act as emergency shelters and then which should be used as long-term shelters, while simultaneously optimizing the allocation of a population to them. Damage caused to evacuation routes is considered in addition to the number of evacuees and shelter capacity. In terms of the model’s emergency and long-term shelter stages, the objectives are to minimize (i) total weighted evacuation time, and (ii) total shelter area used. An interleaved MPSO–GA applied to the model yielded better results than achieved using MPSO or GA in isolation. For a case study with an earthquake affecting the area of Jinzhan within Beijing’s Chaoyang district in China, results generated present government with a range of solution options. Thus, based on government preferences, choices can be made regarding the locations in which to construct shelters and how to allocate the population to them.