Determining fire station locations using the reverse fixed centered K-means algorithm considering existing stations

Fires pose a significant risk, resulting in substantial loss of life and property worldwide. Therefore, combating fires is critical to mitigating losses during natural disasters. For an effective fire protection service, the appropriate location of fire stations plays a major role. This study addresses the challenge of determining the required number of fire stations and their locations for five central districts in Kayseri. The key aspect of this problem is to consider the existing fire stations (fixed centers). This study proposes a reverse fixed-centered K-means algorithm using fixed center points to address the abovementioned problem. In the proposed method, first, the points within five minutes of the fixed centers are determined and removed from the problem space. Then, dynamic centers are determined by applying the K-means algorithm to the remaining points, considering the specified number of dynamic centers. The optimal positioning process is conducted by considering the obtained dynamic and fixed centers and updating the dynamic centers in each iteration. The experimental results compare the proposed method with related approaches, demonstrating its superior performance.