MORSO for Multi-Objective Fire Station Location on Urban Road Networks: The Mosul Case

Abstract

The effectiveness of urban fire response is heavily reliant on the fire station location strategy on a realistic road network. This paper presents a two-stage GIS framework for the additive fire station location problem on an urban road graph. Stage 1 relies on an economic location criterion to identify the number of new stations (N*) needed. Stage 2 employs a reinforced multi-objective particle swarm optimizer (MORSO) to explore potential locations and simultaneously optimize (i) the maximum and (ii) the average nearest station network distance, computed using multi-source Dijkstra algorithms with careful treatment of disconnected components. Feasibility constraints (such as minimum distance and equity/coverage criteria) are used to guarantee the existence of implementable solutions. A test example on Mosul, Iraq, illustrates that adding three stations to the existing nine (total 12) leads to a significant improvement in accessibility: the best Pareto solution decreases the average distance from 3,474.67 m to 2,823.09 m ($-18.8\%$) and the maximum distance from 12,595.07 m to 8,752.15 m (-30.5%), with further tail optimization (P90). Distances are also reported as travel-time proxy bands using an average speed of 35 km/h ($\approx$3.0/6.0/7.5 km for 4/8/10 minutes): coverage increases from 52.60/89.02/93.64% to 64.16/93.06/97.11% for the $\le$4/$\le$8/$\le$10-minute bands, respectively. Multi run bench marking and sensitivity analysis further support the robustness and practical usability of the proposed planning workflow. The resulting Pareto front and mapped layouts enable transparent efficiency risk trade offs for deployment.